Recording Vocals
Here are some tips and suggestions for recording, editing and mixing vocals, from the perspective of a producer/engineer. Although my focus is rock music, some of the following suggestions should be useful for other types of music but which may require different approaches.
In almost all articles about any type of audio recording, you'll hear "there are no rules — whatever sounds good, is good." The same is true for recording vocals; experimentation is important, and as each song and singer require a different technique, trying out ideas is even more important. I do believe that there is still one rule though: a great performance trumps perfect pitch and timing, and also is far more important than the vocal sound itself. I'd rather hear a vocal that is totally believable and involves me in the meaning and emotion of the song than a vocal edited into technical brilliance with cool effects and perfect sonics, but is stiff, uninvolving and robotic.
The Recording Space
Where and how you set up your recording area for vocals will have as much, if not more, effect than your choice of equipment. Make sure that the microphone is not in the center of the room; this is where standing waves will be most prominent. The best position is usually about a third along the longest dimension, but roughly equidistant from the side walls.
Use acoustic baffles and sound absorbing materials to create a good space within your recording room. Remember that you want to have sound absorption behind the singer, and also (although not as important if you're using a cardiod patterned mic) behind the mic. Pick an area of your room that is well away from reflective surfaces.
If you're recording at home, a good, inexpensive, way to create free-standing acoustic screens is to construct a frame using 1" PVC piping, usually used for garden sprinklers. This stuff is available from any big box hardware store, and is very cheap and easy to work with. Buy several lengths of piping, some T-joints to make stands, some 90 degree bend pieces and glue. Sketch out your plans, measure accurately and make them sufficiently tall - 7' is usually a good height, and you really only care about the top 4'. Once you've double-checked your measurements, start cutting the pipes. Assemble the entire frame 'dry' first to make sure that all your pieces work together, and once you have made sure that your measurements are accurate, start gluing. Now use these frames as hangers for quilts, duvets, blankets or sleeping bags thrown over the top. If you're more ambitious, use furniture movers' packing blankets which are also very cheap. You should easily be able to construct several complete screens for less than $25 each.
If you want to get even more ambitious, construct wood frames and Owens Corning 703 semi-rigid panels, covered with an open weave fabric such as burlap. A box of six 48" x 24" x 2" panels can be found for around $70. You can get fancy by making one side of the screen absorbent, and use peg board on the other side to be more reflective.
Microphone Technique and Position
Good singers will move their position relative to the microphone depending on the changing vocals levels during the song. Too much movement, and varying room reflections will change the sound, resulting in a less consistent, and potentially more remote sound. However, even experienced singers forget mic position when they get caught up in the emotion of the performance. Part of the producer/engineer's job is to make the singer relaxed and comfortable, and able to concentrate on the performance. Nagging singers about anything is usually a bad idea, and can be annoying. Mention microphone position a couple of times, but be supportive and gentle. With some more inexperienced singers, it can help to make a mark on the floor with tape, and suggest to the singer that that's their mark. This doesn't help much with singers who 'lean' into the mic, but it's a start.
Arrange the mic position so that the singer sings "up" into the mic capsule. 2 to 4 inches above the mouth should work fine; the idea is that the singer should raise their chin an inch or two. Start off with the singer's mouth about 5" - 6" from the microphone (maybe a little closer with a dynamic mic). Every singer will be different, so a few passes checking out the level changes between loud and soft passages in the song will be needed. Remember that with most cardioid mics, the closer the singer is to the mic, more bass will be accentuated because of the proximity effect.
Always use a good pop shield between the singer and the microphone to prevent 'b' and p' plosives. Although DIY stocking constructions can work ok, a decent pop shield is not expensive and worth having. Metal mesh shields are best, with fabric mesh a close second. Don't use foam wind covers: they're ineffective for pop suppression and attenuate higher frequencies. Position the shield so that it's about 3"-4" from the mic — this also should be changed if the singer tends to get too close to the mic for quieter passages — use the position of the shield to make the singer keep his/her distance from the mic.
With some voices and microphones, sibilance can be a problem. Sibilance is the bright 'hissing' noise made by consonants such as "s," "t" "c" and "f." To reduce sibilance with directional mics is to rotate it off-axis by a few degrees. This position reduces the high-frequency sensitivity of most mics a touch, and can help reduce problem sibilance. Around 15 degrees is enough. An off-axis placement also has the advantage of reducing popping. It is always better to spend time getting the vocals sounding good, with minimal pops and sibilance rather than try editing or de-essers later.
If you're recording a singer with a naturally 'boxy' sound, or perhaps a singer with an obviously nasal tone, an omni or figure-8 patterned mic can help, although you may need to position the singer a little closer to the mic so that the direct sound of the voice is more prominent than the room effect.
For singers with wildly fluctuating levels, try using two mics, one more sensitive than other. Space them a few inches apart, angled in slightly, and have the vocalist sing into the space between the mics. Or use a similar mic (if available) and switch in the -10 or -20 dB pad on one of them. Record on separate tracks, and switch them at the editing/mixing stage later if needed.
Most mics and preamps have a bass roll-off filter, often at a set turnover frequency between 80 - 150 Hz. Use the most appropriate setting for the vocal, but use it.
A common question is "Why do I see vocal mics positioned upside down?" Actually, there are two reasons: the first dates back
60+ years: one of the most used and famous microphones for vocals (along with the AKG C12) is the Neumann U47 tube mic, introduced just after WWII. This mic had the M7 PVC capsule, which had a tendency to dry out over time, and this deterioration was accelerated by heat. So it made sense to hang U47s upside down, as heat rises, so eliminating cooking the capsule over time. Another reason is that it's easier to get singers to sing 'up' a little; singers are less likely to mess with a vocal mic height when it's suspended on a boom upside down than a mic on a stand right in front of them.
Once the mics were mounted upside down, another advantage became obvious; there was a benefit for the singer by getting the mic, stand, boom and cables out of the way, offering the singer a better view of a music stand for holding lyrics, hanging headphones and storing other items.
Scheduling
There used to be a tradition of leaving vocals to the end of recording projects, after most overdubs and extra sweetening had been completed. Not a great idea: singers can't be expected to sit around for days or weeks, and then pump out all their vocals in two or three days.
The number of hours you can expect to record vocals is limited, especially with heavier rock and metal. So it makes sense to scheduled vocal recording over as many days as possible, breaking up sessions into styles so that more melodic, less stressful, vocals are recorded first which serve as a warm up to attempt more aggressive parts. Depending on the length of your sessions, include a few days off too, particularly after a sessions where the singer's voice has been really stressed.
Some singers prefer to attempt capturing a continuous performance in longer takes, punching in corrective phrases or parts. If you're recording rock vocals with dynamics ranging from a whisper to a roar, try the dual microphone suggestion I mentioned above.
Microphone Choice
Experienced singers who have recorded often may know which microphones have worked well for them in the past. Ask, if you know that the singer has this depth of experience.
If your singer doesn't have a great deal of experience, then start by choosing a microphone with an inverse character to the singer's natural sound.
Every voice is different, and therefore there is no 'perfect' vocal microphone. Most advice on recording vocals starts by telling you to use a large diaphragm condenser (LDC) microphone. Although often true, 'often' isn't the same as 'always.' If you have a singer (particularly female singers) with a thin, high, hard vocal sound, better results can be achieved using a dynamic microphone, or even a ribbon. An Electrovoice RE20 can be a great vocal mic, as can the Shure SM7B, and the lowly SM58 can sound great, paired with a good preamp. Cheap LDC microphones often have a noticeable harshness at higher frequencies and therefore are exactly the wrong choice for this type of singer. The market is crowded with sub $1,000 LDC microphones, many of which have this unnatural harshness, although I tried an inexpensive Audix CX212 recently that was rather good. Among the many other reasonably priced, but premium, LDC microphone brands, Lauten Audio and Peluso are worth checking out.
Another variable is the material itself: having found the perfect mic for the first song, perhaps a rock song with wide dynamics and a belted-out chorus, don't assume that this should be the mic used throughout all vocal sessions. An appropriate choice for an intimate ballad may require a different microphone. Usually, two different mics will cover the range of vocal delivery. Some mics don't work as well with high vocal levels, as the timbre of vocal character changes with delivery.
Buying Microphones
If you're primarily recording just one singer, don't base your purchasing decision on reviews. Try out different microphones, and pick one that best suits your singer. Microphones are tricky to buy — online stores often won't accept microphone returns unless they have a defect. Visit a good pro audio dealer with a demo room and try out several mics. Or book an hour or two at a local studio with a decent mic cabinet, and try out their options, but use the same preamp with each test, so that you're comparing apples to apples.
If you're planning to record different, unknown singers, start with at least a good dynamic mic, and a large diaphragm condenser.
If you have a limited budget, spend money on a preamp first. A good preamp can make a basic mic sound good, but a crappy preamp can make a great mic sound far less impressive. You can add to your mic collection as your budget expands, but it makes no sense to replace poor quality preamps.
Headphone Mix
The headphone mix that the singer hears is critical to achieve a great performance. A touch of reverb is often helpful as it creates a sense of space around the vocal, but ask the singer what they prefer.
The balance between the backing track and the vocal can influence the performance:
If the vocal is too loud in the headphones = flat notes — lower energy.
If the vocal is too soft in the headphones = sharp notes and a strained performance
If a singer is mostly flat, turn up the backing track in their phones a little without telling them. If the singer is mostly sharp, then turn the backing track down. This will help correct consistent, but small, pitch problems. If the singer is inconsistent; sometimes sharp, sometimes flat, then you have to do more work in helping the singer create the right performance. Remember though, the performance is the most important element: slight pitch issues are less important than a convincing, emotionally believable performance.
Take the time to set up your monitor mix so that you can easily adjust the mix so that you can immediately offer the singer changes to help; drums a bit louder to help timing issues, more keys or guitars to help with pitch, less bass or lead licks to help with focus , etc. Keep the headphone mix as simple as possible, and usually it's a good idea to pull out most of the sweetening and any parts which could be distracting, such as percussion with off rhythms.
Headphone choice is another consideration, and if your singer prefers 'one ear off' that will lead you to a different mix. Closed back headphones are needed.
Recording Chain
Try to capture the character and performance of the vocal going to the computer (or tape) and forget about trying to create it in the mix. Editing and mixing vocal tracks should be the final touches to enhance an already good vocal track.
Audio interfaces contain preamps. Most of them are just OK, but are usually bland and can easily be bettered by a good, separate preamp.
For rock songs, I always use hardware compressors as part of recording chain because this type of material often needs more dynamic control before the signal gets to the analog-to-digital converters. For different types of music, you may decide that a compressor isn't needed. Unlike the days when we recorded to tape and wanted to get the best signal-to-noise ratio, and therefore wanted to record hefty levels to tape, modern digital recording has such a huge dynamic range when recording 24-bit that this isn't strictly necessary. However, I still choose to use hardware compression, but now not as much for dynamic control but rather because I like the coloration of the sound that different compressors impart. A compressed vocal in the headphones also benefits the singer; they feel less need to compensate dynamically.
Aim for 5-6 dB of gain reduction on the signal peaks, but don't overdo compression during recording—you can't remove it later. Start with a 'not too fast' attack time, so the transients pass relatively unscathed, with medium release time. You'll need to experiment with your particular recording chain, singer and style of song.
One disadvantage of project studio recording is that there is less attention paid to the recording levels before the converters. I like riding a vocal fader during the recording, but to do this you'll need some type of outboard mixer. It then becomes simple to push/pull the vocal levels during the recording which results in a smoother recording, and less work to do at the editing/mixer stage. There are many small mixers available at reasonable cost, some of which have the audio interface and converters built-in, so that they connect to your DAW machine via FireWire or USB. If you're recording just with a DAW, be sure to set up your cue (foldback) monitoring sends so that they are sent pre-fade, so your adjustments to levels don't alter the headphone mix.
Encouragement and Criticism
The psychology of how to get the best performance out of a singer is much more important than microphone choice, preamps, compression or anything else to do with the technical aspects of recording. It's all about getting a singer to relax, and feel confident enough to give their very best performance, without feeling self-conscious. I have been a guest on many sessions where I cringed—not because of the recorded sound, but how thoughtless comments destroyed the mood of the session, and made the singer uncomfortable and doubt their performance. I have actually seen an engineer jump on to the talkback button and bark "flat, do it again" at the first delivery of an iffy note. Moments later on another pass, this same idiot impatiently interrupted again by saying "wrong again!" If it were my session, I would have fired the moron.
Be mindful that the singer is trying their hardest. Encouragement with gentle criticism pays dividends. Try to avoid comments where you're making comments containing "don't." Use positive reinforcement. Take frequent short breaks, offering the singer a chance to relax and listen back to their progress. Trying to record take after take can be confusing; it's hard for a singer to hear which parts of their performance need adjustment, so breaks, when a singer can review the takes so far, can help enormously.
Most singers prefer a relaxed environment, and don't want to feel that they are in a fishbowl. Turn down the lights, and get people out of the control room. Singers don't want the distraction of seeing band members appearing to be cracking jokes, having conversations, mixing cocktails or anything else unrelated to the vocal performance. Hearing a babble of background conversation each time the talkback button is pressed is another distraction.
Record everything. There's no such thing as a 'practice' pass, particularly with DAW recording (multiple vocal tracking to tape can get more challenging). Often the best performance is achieved in early takes so make sure you've recorded them. Set up your DAW so each pass is automatically recorded to a new track.
Make Notes
Make notes of settings used in your hardware recording chain. Microphones settings such as LF cut and pad settings, preamp settings, compressor settings , etc. All will be very useful if you decide at a later stage to record or replace additional vocal parts. Also make these notes for any hardware you use at the mixing stage.
Editing and Mixing Vocals
For the purposes of this discussion, let's talk about a very simple scenario: assume that your goal is to have two separate tracks of vocals: one for the verses, and one for the choruses. The verses are fairly constrained, but the choruses involve a much more aggressive vocal performance, where the singer's voice character (timbre) changes and the levels and vocal sound change considerably.
Comping Vocals
Once the recording stage is completed, you'll probably have multiple audio tracks of your vocals. You now will want to comp (to compile or to create a composite vocal track) vocal takes into the best comp'd track(s). You may want to keep sections of the vocals on separate tracks so that you can more easily apply processing and effects later, reducing the amount of automation and dynamic control needed, and corrective processes such as EQ which may not be needed on the entire completed vocal track. Remember that the comp should result in seamless vocal parts which do not sound disjointed. Often the individual phrases may not sound quite right when soloed, as you are auditioning out of context. Just remember to focus more on the continuity of the performance and not on all the details at this stage — we'll tackle the fine detail later.
Every DAW will have features for comping tracks, and they all work slightly differently but the result is the same; a new track is created, and parts from your vocal takes are placed on this new track, comping the best parts of all your takes. The unused parts are removed from the arrangement, but filed away in case they are needed later. See the section below though, as some out takes can be very useful for simulating double tracked vocals.
Destructive Editing
Once you have your comp'd vocal tracks, you may have residual problems to deal with. There may be noises, pops, mouth noises, breath sounds and unwanted level fluctuation between phrases.
There are two schools of thought here; some people always use automation for everything; others (like me) prefer to rely less on automation, but rather permanently editing the audio file. For me, it's easier to scoot through a file, selecting parts in a wave form editor, and eliminating noises, adding silence between phrases, reducing some breath sounds, interpolating pops and other minor fixes, and changing gain of entire phrases, than to make small adjustments manually by graphically changing automation. To me, it's just good housekeeping to have a smooth vocal track where I can use any additional compression mostly for sonic reasons rather than depending on automation and compression for too broad dynamic control.
Some dislike destructive editing as you really can't change your mind later (other than reverting to a backed up original file) but I actually like this supposed restriction; I like making decisions as early in the recoding process as possible, only leaving minor instrumental parts open to tailor later, probably because I came from the old school analogue recording days using tape, when we had to make such choices as there was much less flexibility than in the present day digital domain, and when bouncing tracks down to create a comp'd track on a tape machine was commonplace, in order to free up more tracks for record additional parts. No going back in those days!
De-Essers
De-essing isn't much of a concern for me, as I believe that recording the vocal using the right microphone, position and hardware chain prevents the problem before it becomes an issue. However, if you record many different singers, or mix material where you had no control during the recording process, then think about investing in a used dbx 902 - the best de-esser made, hardware or software. They can be found quite inexpensively. You'll need a dbx 900 series rack to house the module, but they are not expensive, and gives you somewhere to house additional 900 series units such as the excellent 903 compressor (essentially a dbx 160 in a 900 module) ; also quite cheap and a great workhorse as a single channel compressor/limiter.
EQ
Hopefully, you have already record your vocals with a LF cut setting on the microphone set at 75 or 80 kHz, or an equivalent setting on your preamp. But your work on the low end of the spectrum probably isn't done. Listen to your vocal track in context, and see how much you can roll off at the low end. Don't do this with the track soloed: you'll think that the vocal is sounding too thin. But in with the mix, you'll be surprised how much low end you can remove. This will help make the vocal sit right in the mix, and is one step in cleaning up the low end, a common problem in mixes that have too little low end definition. This concept also applies to many other instruments — get rid of frequencies that aren't the focal point of that instrument's sound and stop too many instruments competing in the same frequency band.
Taking this idea one step further, try to cut frequencies rather than boosting them. Boominess is most apparent around 200-250 Hz. Instead of boosting mid range frequencies to attempt to make the vocal more present, try notching EQ around 230 Hz with a higher Q setting (every voice will be different) and then boost the vocal level.
One of the most common problems with mixing vocals is to get them to sit well in the mix with the other instrumentation. A common mistake is just to make the vocal louder where it clashes with other instruments.
Many vocals have their most obvious frequencies in the range from about 750 Hz to 2.8 kHz. Unfortunately, guitars are in the same range and turning up the vocal will just make the vocals sit on top of the guitars, not mixed in with guitars. A better approach is to EQ the guitar so that the guitar sound is scooped out where it's clashing with the vocals, using as narrow a 'notch' as possible, the result being that the vocals now have their own space alongside the guitars but only in the competing range.
Another technique to try is compressing the guitars more when the vocal is present, using side-chain compression to affect the guitars. In my example above, use a send from the vocal track, heavily EQ's, so that the vocal being sent to the compressors side chain has a very limited frequency range, centered on the vocal's fundamental frequency. It won't sound good, but that doesn't matter as you're not hearing this vocal; its sole purpose is to control the compression of the guitars only for that frequency range.
If you're looking for a new tool for mixing, dynamic EQ is a very useful plug-in, combining an EQ and compressor with features enabling you to apply EQ only when the signal within a certain frequency range exceeds a certain loudness, as set by your threshold and filter. A good example of this is the Brainworx dynEQ.
Compression
A compressor isn't intelligent, so it doesn't know when vocal dynamics are being changed for an intentional effect or because the singer wasn't adequately controlling the dynamics of their performance, or even didn't care and got caught up in the emotional delivery of their take. Often you want vocal dynamics. Your job is to ride the vocal levels (or automate) to preserve the dynamics you want, while smoothing out those which you don't.
A compressor can make a vocal ride properly within a dense mix and it is an indispensable tool to control peaks and bring average levels up. Compressors work much faster than you can, but for slow averaging of vocal dynamics, you will do a better, and smarter, job.
Don't ask your compressor to work too hard.
The more you do, the more obvious the artifacts will be, and depending on your compressor settings, you will make the compression noticeable, which is often not desirable. If, after you have tried the recording and editing tips described above, you still find that your vocal levels are not adequately controlled, try routing the output of your vocal track to a aux, and ride the vocal level to the aux, using automation to capture your moves. Now add final compression and other effects you want to the aux, not to the original vocal track.
Serial compression:
I use this technique all the time, and I think it's the most commonly overlooked technique regarding compression. Using two compressors can be very helpful, as you're making the two compressors to perform different tasks. Try using a fast FET-type compressor (such as an 1176 or software emulation) set to a fast attack, with the threshold (or input control in the case of an 1176) set so that only the peak signals are being compressed leaving the body of the vocal untouched. Aim for a gain reduction that sounds natural but firmly controls the peaks. Now follow this with opto type compressor (an LA-2 or LA-3 type) with a slower attack and release, but this time set the threshold lower with more moderate ratio. This will level the vocal track nicely, and the two compressors together will give you a much smoother result.
Parallel Compression:
Often used on other instruments, particularly drums and guitars, a modified version of parallel compression can work great on vocals. Either duplicate those vocal parts you want to use this effect on, or send from the original vocal track to an aux channel. On the aux channel (or separate vocal FX track), add a great deal of EQ to boost presence and air, roll off the low end, and add obscene amount of compression. You now have two vocal tracks, one natural and one with heavy EQ and compression. Now set the natural vocal channel to it's appropriate mix level and bring up the FX vocal so it just peaks out under the unaffected vocal to add presence and excitement. Use the natural vocal channel for the send to reverb you might want. Set right, the vocal won't get lost in the mix, and although the unaffected vocal sounds natural there will be a presence and edge that can be a real benefit to your vocal mix.
See here for more details on compression and compressor types.
Vocal Tracking, ADT, Doubling, Thickening etc.
ADT (Artificial Double Tracking): Since the dawn of time (well, since The Beatles) double-tracking vocals is one of the most useful and commonly used effects. Double-tracked vocals are stacking the same vocal twice (sometimes more) to thicken the vocal sound. This can also work very well on backing vocals.
The best results will be obtained by doing it for real, getting the singer(s) to duplicate their part on another track. However, it's a laborious process. Getting the two vocals to be tight enough can be quite difficult, particularly if the original is more of a creative interpretation than a technically consistent performance. Where words start or end on hard consonant sounds, such "T" or "C," the doubling result can sound sloppy. Some singers are really good at it, others are not. Either way, it can be a time-consuming and demanding process.
You'd think that it would be easy to recreate the ADT effect easily using modern technology. Just adding a digital delay to the vocal at the same level as the original part, with a delay time of 40-100 ms and no feedback should do it, right? Not really.
When ADT was first created at Abbey Road, the technique involved two tape machines; one playing back the original vocal to another tape machine, which then recorded and played back the delayed vocal in to another channel on the mixing console, the delay time being changed by use of vari-speed control on the second tape machine. Because tape was being used, with two machines, variations were being introduced several times along the process. Tape machines have wow, flutter and scrape flutter, each of which introduces slight changes to the delayed vocal. Add to this the addition of harmonic distortion introduced by the delayed vocal being passed through three stages of signal electronics, and the characteristics of tape itself, all of which introduces undefinable 'analog warmth' and the end result is just not the same as slapping on a digital delay.
However, its possible to get close enough for all but the most critical ears. There are various plug-ins which claim to simulate this quite well, but they are mostly expensive. I've had some success by using VacuumSound's ADT plug-in coupled with a distortion/warmth unit like a Culture Vulture. Alternative plug-ins such as PSP's Vintage Warmer can help too. Additional, a fast compressor taming the peaks to soften hard consonants, or even an envelope shaper, can make the tracked vocal sit better with the original. Melodyne's offset by random pitch and time is another additional process worth trying.
Dual Shifter:
Another favorite. Older Eventide hardware like the h3000 had a dual shifter setting. This can be recreated by standard plug-ins. First, copy your vocal track twice. On copy 1, insert a compressor (opto type) with a low ratio and the threshold set so that the compressor is working on all but the quietest parts. Follow this with a pitch shifter, pitching the vocal down by 10 - 15 cents. Now add a digital delay, initially set to something like 10 - 50 ms. On copy 2, insert a fast FET or VCA type compressor with a fast attack, higher ratio, with the threshold set so that it's only working on the peaks. Add a pitch shifter again, this time pitching up 10 - 15 cents. Once again, add a digital delay, but this time set to a different delay than copy 1. Route the outputs of these two copies to a submix, panned L - R, where you can apply further compression if needed, together with EQ further removing low-end mud, and perhaps adding a little presence. Tweak these suggested 'starting point' settings to taste.
Whisper track:
Here's a tip for an effect which may be useful on slow, more intimate verses, especially ballads. Ask your singer to whisper, almost talking, along with the vocal track, often in a lower register, compress it, and bring this up so it's barely audible under the lead vocal.
Reamping:
Usually used for guitars and keyboard tracks, this can be very useful for really hard edged, tough vocals in hard rock or metal mixes. It can add lots of distortion and a great live sound to a vocal. Send your vocal track out to a recording room via a reamp box, or at a pinch, a passive DI box in reverse. Use this send as an input to a guitar amp, and close mic it. Also, add a room mic and compress it. Record these two new tracks, and mix them to taste under your lead vocal. As there's really nothing useful coming out of a guitar amp above about 4.5 kHz, you'll find that mixed right, it can add a lot of edge and aggression into your vocals.
Reverb
Washy vocal reverb has become less fashionable in recent years, compared to pop/rock music production of the classics from the 70's and 80's. But there are many exceptions. The current trend is to use reverb to create a little space around the vocal, but not much that it's an obviously applied reverb effect.
Unlike recording orchestral music or recordings of ensembles in folk, jazz or MOR, I find that the use of convolution reverbs are less useful in pop/rock/metal genres. Convolution reverbs use a 'fingerprint' of an actual space and mix it (or convolve) your vocal with that space, creating the impression that your singer's vocal was recorded in that space.
The downside of using convolution reverbs is that the IRs (the Impulse Response files used to convolve with the vocal) are less flexible and offer fewer options to create a suitable reverb sound. So I often prefer to use digital reverb units which use algorithms to create synthetic effects. These are purely mathematical calculations where every part of the reverb effect is freely adjustable and you can tailor exactly the vocal reverb you want. Lexicon is the best known and most used manufacturer of all digital algorithm-based reverbs since the 70's.
For a lead vocal, I'd suggest starting with a short reverb time (well under 2 secs), and a pre-delay around 40 - 60 ms and boost the early refections part of the parameters depending on the reverb device you're using. It's not possible to give hard-and-fast recommendations; you'll need to experiment. I've always been a lover of the EMT 140 fbST plate for vocals (the real one!) but some digital reverbs (the original 224XL Rich Plate that carried through the Lexicon PCM series) are all very good. In software emulations, the UAD Plate 140 is good, and the recently introduced Lexicon Native plug-ins are superb. Unlike many software emulations of hardware, reverb plug-ins work extremely well, as in reality the hardware digital reverbs are really just giant calculators, with little or no significant analogue component coloration, so their algorithms can be calculated just as well with modern computers, although the CPU requirements can be considerable; a good reason to have one or more hardware units if your budget permits.
The usual method of using reverb is to insert your reverb on an aux (containing either routing to external hardware or your reverb plug-in) and send to it from your vocal(s) tracks. Increase your return levels until you can just hear the effect, and then back off a touch. As mentioned earlier, this is merely a starting point; adjust the pre-delay, and reverb time to fit your mix, and don't worry too much about the finer digital delay parameters. You will probably need to EQ the send to your reverb, and often the return, by removing anything under at least 100 Hz and treat other areas of your reverb to removing any annoying resonances and notch out any unwanted frequency bands.
Try altering the reverb time between the different sections of your song. In rock music, start by making it shorter for verses and longer for the choruses. Generally you'll have verses and chorus parts on different tracks, so you can use multiple reverb unit instances if you're using plug-ins, as reverb plug-ins in particular are challenging to automate when altering certain parameters such as reverb times. If you're using hardware, but only one unit, you'll have to record the effect returns for parts using differing hardware settings.
Background Vocals
There are many techniques for recording, editing and mixing background vocals (B.Vox).
Start by applying the same editing and mixing techniques as suggested earlier for lead vocals. Your first task is to clean up the individual vocal parts. Most often, you're looking to create a blended, smooth sound for these vocals.
Here's an approach I like using:
Panning:
Keep your B.Vox out of the way of your lead vocal, which is almost always panned dead center. Start by panning them at 10 o'clock and 2 o'clock. I prefer to leave the extreme hard left, hard right positions for any effects returns. Having said that, I have found occasionally that panning the backing vocals hard left-right can work, returning effects to your mix with a narrower stereo spread.
Low End EQ:
Just as we removed low end from the lead vocals, you should do the same here. But this time, while listening to your B.Vox in context, move your low cut turnover frequency up until your B.Vox start to noticeably thin out. Then back off a touch. Soloed, these vocals will sound way too thin, but will work fine in the mix. It's important to make sure your B.Vox don't muddy-up low end guitar and keyboard parts, as well as clouding the space used by your bass parts.
Compression:
By now you should have applied broad level control manually to your B.Vox parts. Next step is to beat them unmercifully into a blend, where individual transients and untamed peaks are ruthlessly stamped on. Perhaps this is one area of mixing where you can forget subtlety; squash those vocal peaks! Fast attack, high ratio is the order of the day, with the threshold set so that peaks are heavily controlled.
Submix:
Route the outputs of all your B.Vox tracks to a stereo bus, maintaining pan position, and with the correct mix of your B.Vox parts. This single fader aux will control your overall B.Vox level. Leave it alone for now. You'll probably need to make multiple passes throughout your mix, automating the balance between the various B.Vox parts. Usually, I do this by grabbing automated faders in 'touch' mode and will go in to the individual tracks afterwards, smoothing out automation points and correcting any mistakes.
Once you have all your individual levels set, the next step is to apply some 'glue' to further smooth out and blend your B.Vox. Start by inserting a stereo bus compressor (hardware or software) on your aux, and set it (as a starting point) with a low ratio (2:1 or less as a start), not too fast attack, low-ish threshold, medium release and aim for gain reduction in the 3-4 dB range, adjusting the threshold until this is achieved. Continue playing through the song, adjusting the release value until it's smooth and seamless.
Almost there!
Now you want to apply some EQ, if necessary, to create a space for the lead vocal if the B.Vox parts occur at the same time as the lead vocal. If they only occur occasionally, apply the side-chain compression or dynamic EQ techniques as described earlier, but using the lead vocal as the control signal. If the B.Vox overlap the lead vocals often, then apply the EQ scooping technique described above so that the B.Vox EQ is notched in the lead vocal's range — it won't take much to make quite a difference. In any event, you should use this additional processing only if you have trouble riding the B.Vox levels in the mix.
Depending on your set up, you can send from the B.Vox aux to another aux for reverb and other effects. On some B.Vox, particularly on slower, more intimate material, a chorus effect can work well on B.Vox mixed in to taste, as well as before the reverb send. I particularly like Fluid, from Audio Damage, although there are many on the market, and if you have access to Lexicon or Eventide hardware, they're unbeatable. ADT and dual shifter effects (see above) can also work well on B.Vox, as well as many effects normally used for lead vocals. Again, experiment after you've cleaned up your B.Vox basic tracks.
Vocal Balance
Vocals not sitting well in the mix is one of the commonest questions around the audio recording forums. Hopefully, some of the above suggestions will help, but as a final note, it's important to listen to your mix, as it evolves, on a variety of playback systems. Often, too loud or too quiet vocals will be easier to judge at very low playback levels, as well as playing your mixes back on small crappy speakers, simulating the environment in which your song is most likely to be played. Also try playing back your mix and stand outside the mix room, and play back your mix on other devices, including your car and all the iGadgets. Listen to commercial mixes of the same genre and see how your vocal balance and sound stand up.
Finally, when you commit to your mix, take the time to create additional, and clearly noted, alternative mixes with the lead vocal up a little on one, and down a little on another. "A Little" depends on your particular song, but 1 or 2 dB should be plenty. When you get to the mastering stage, the additional, overall, mastering processing needed may affect your vocal balance and it can be a lifesaver to have alternative mixes available, identical other than the vocal level. Creating separations for mastering can be even more useful.
The words "mixing" and "mastering" are often incorrectly used synonymously and in this brief article I hope to offer an explanation of the main differences. They are two completely different processes, and I strongly advise you to keep them separate.
MIXING
Years ago, mixing engineers were called balance engineers, which I think is a better, more descriptive, term. The whole idea of mixing is to achieve a technically and artistically sympathetic treatment of the various musical elements that together convey the original intent of the composition. In other words, to make the song sound good.
The mixing process involves using all the tools that are available including panning, EQ, compression, limiting, delays, reverb, effects and a multitude of other tricks.
Sometimes, the mixing engineer will attempt to mix the song like a live performance by panning and creating space to give the impression that it's a band playing live and together. Other types of music are geared towards a free form sound stage that has no bearing on live shows - it's a sonic landscape that tosses out conventional performance restraints. Your music and vision will set the approach to be used.
When creating your mix you can use meticulous automation to affect each part of every track like adding momentary effects, adjust fader levels, bypass/adding effects, delays and reverbs at particular moments in the song, and changing EQ and other parameters on just parts of a track to highlight certain passages. Mixing a song can be compared to assembling a jigsaw puzzle. Other mixing approaches favor a more organic approach, particularly mixing OTB (outside the box) on an analogue console, where a 'mix performance' is made and parts are later edited together.
Often mixing will involve editing decisions: does that particular part really add to the song? If not, toss it. Do those multi-tracked guitars really work, or do they just create a muddy mess? Does the bass clash with the kick, or do they work well together? Those backing vocals - are they really adding to the earlier choruses, or would it help the song build more by bringing them in later in the song? Does that 24 bar intro maintain interest and grab and hold attention, or would it be better to trim it?
A Few Tips
Take your time. When you're tired, you will not make the best decisions. I like a three session approach, with long breaks between them (hours, if not a day). Take a song, and spend Session 1 on cleaning up the individual tracks, editing out noises, squeaks, pops, guitar amp noises, vocal noises, unwanted drum noises etc. Session 2 gets all the housekeeping done. Move tracks around, set up hardware gear (if you use hardware effects) patch everything up, assign busses to reverb, delays, submixes etc. and get rough sounds on your instrumentation. Session 3 is now confined to the main mix. The Balance.
If you're mixing the same artist's sessions, save templates of the set up you created in Session 2 above, so that your routing and tracking layout can be reused if appropriate.
Plug-ins are not a replacement for either talent or knowledge. Don't buy plug-ins thinking that somehow your recording and mixing will magically improve - learn to use the ones you already have, and only buy new software or hardware when you know what you're looking for, and are sure that you can't already achieve what you need with what you already have. Reading manuals, experimenting, and looking for on-line tutorials could save you a lot of money.
When you think you're happy with a mix, stop. Get some rest, and listen to it the following day.
Leave automation to the very end of your mix process.
Keep different versions, and name and date them. It's easy to let a mix 'run away from you' and get out of control. Sometimes an earlier mix is the best one.
Try to avoid, as much as possible, listening to solo'd instruments. Spending hours on fine-tuning a bass sound doesn't help much when you find it doesn't work with the guitars, synthesizer parts or drums.
Remember that every knob works anti-clockwise as well as clockwise. Although this seems obvious, constantly raising levels and applying EQ boosts all over the mix can quickly result is a harsh, brittle mess. Try notching out frequencies with EQ rather than boosting others. Always consider frequency ranges - too many instruments fighting for space in a particular part of a frequency spectrum will result in loss of clarity and space.
Have at least two, preferably three sets of monitors. High End (your studio's monitors), El Cheapo - equivalent to what people have in their living rooms, like their TV surround speakers or stereo systems (although these are dying out) and Crap: alarm clock radios, bad car systems, boom boxes etc. And of course, your iPod or equivalent as a fourth.
Don't worry about perceived loudness at this stage. Just create mixes that sound good, and have plenty of headroom. Advice on this varies - within a DAW -6 to -10 dbFS is a rough guide.
Listen to mixes at various levels, from loud to extremely quiet. It's amazing how an incorrect balance jumps out at you when listening very quietly. Stand up, move around the room - unless you're in a high-end studio, mixes done in a room with less than optimized acoustic treatment will sound different when you move around or stand outside the room.
Burn a CD of your mix, and transfer it to an iPod. Play the CD on other systems; as many as possible.
Get comments from people who favor your genre of music and who you trust. Not about the song, but what they perceive as the weak and strong points of your mix (can't hear the bass, vocals not loud enough, guitar licks swamp everything). Family members are often not a good choice. (Most) want to be supportive, and are unlikely to tell you that you've just spent days producing crap.
Unless you REALLY know what you're doing, do NOT add any effects, compressors, limiters to your mix bus. Some experienced engineers may mix with a compressor on the mix bus, but that's because they understand the effect and know how this affects the behavior of the entire mix. If you decide to use master bus processes, when you render your mix, make another pass and make a copy leaving off master bus processing in case you decide to take the mix to a mastering studio.
Also, ignore anyone that tells you to 'normalize' your output mix. Just don't. Ever. All you're doing is increasing the peak levels to 0 dBFS along with all other levels by the same amount, and you are increasing the noise floor, as that too is increased by the same amount as needed to bring the peaks to 0 dBFS.
Similarly, don't use any sample rate/bit rate conversion or dithering. Just don't. Leave these to the mastering stage.
Finally, DO NOT compare your mix to completed, mastered commercial CDs. They won't sound the same, and you'll make yourself crazy. Read the following mastering section.
Paraphrasing Leonardo da Vinci to this context: "A mix (art) is never finished, it's just abandoned"
MASTERING
If you're making a demo, much of what follows won't apply to you. Feel free to experiment with mastering yourself, but do it as a separate process, not part of your mix.
Read books, get advice on forums, talk to experienced engineers. One of the most frequently recommended books is "Mastering Audio: The Art and the Science" by Bob Katz.
Just learn what is involved with mastering. And following on from some earlier advice, buying a mastering 'plug-in' does not make you a mastering engineer. Again, learn how to use what you have first, and read up on Linear Phase EQ, limiting, compression, multi-band compression and loudness maximizers. Learn how to Google.
The whole point of mastering is to take a well mixed song, and add some subtle enhancements that make it sound even better. Mastering is adding sparkle and gloss to your music. You don't fix bad mixes at the mastering stage. Period.
In the same way that a ghastly recording can't be 'fixed in the mix' you can't expect that a lousy mix can be fixed by mastering. A well mixed song shouldn't require very much mastering effort at all, except to balance the various levels of the songs and to change the perceived loudness of the overall CD.
The songs should flow cohesively through the whole CD. Acoustic ballads are supposed to be quieter than a blazing rock track, and the dynamics within the song need to be allowed to live, and not get so squashed that listening fatigue sets in (see "Loudness Wars") and the whole CD is a unpleasant sonic assault on your senses. The timing between tracks is important, as are the length and style of fades and cross-fades.
A mastered song should sound good on all systems. Of course, they will sound different, but if you have a song that only sounds great when played loud on studio monitors but sounds terrible played at low level in a car or on your iPod, you probably need to fix the mix first.
Mastering is a subjective process and when you go to a mastering studio, you're really buying ears and experience.
M.E.s use specialized, often expensive, high-end equipment and monitors. M.E.s work in acoustically designed rooms and are used to listening to a wide variety of music. They know what needs to be done to your mix to get it to translate well to all listening systems. M.E.s rent you a fresh set of ears for your project and can help to correct any slight deficiencies that slipped through the mixing process, often because your room is not flat and there will probably be consistent imbalances in the frequency handling within your mixes which can be corrected.
Frankly, a M.E. doesn't have an investment in your song or you as an artist, and therefore can examine your mix solely based on what will sound good and not be influenced by what your drummer, bass player or singer might like.
A simple example: in your mix room, maybe there is a room mode, which you may not be aware of, at around 250Hz of about +2 dB. This causes you to think that your low end at around that frequency is fine, but in fact at around 250Hz your mix is 2 dB light. A good M.E. will hear this, and make the right EQ change.
Unfortunately, any element you address in the mastering effects other parts of your song whether they need it or not. This can be avoided by using mixing stems, but that is beyond the scope of this article. This illustrates another reason why your mix should be your focus - mastering is adding some fairy dust, not polishing a turd.
A Few Tips
I often hear from fledgling artists that they can't afford a M.E.
Wrong. You've invested in recording gear and instruments, endless weeks of writing, rehearsals and recording, even more hours of mixing, and you now want to toss it out there to the world?
Suppose you're selling your car. Do you just show it 'as is' or do you invest some time in an effort to fix small issues, wash, polish, vacuum, clean the windows, treat the upholstery? You want to present your product in the best possible light to your audience: a car buyer or a potential fan of your music.
Some mastering studios are expensive, but you don't have to use them - there are many mid-range facilities staffed by experienced engineers with great ears who are not that expensive. Some charge by the hour, some by the song, some by complete CD project. Relatively, all but the most expensive facilities are within the reach of most artists. Maybe budget $400 - $1,300 for a good mid-range facility for a 10 song CD - but the price varies substantially depending on your location.
Think about pre-mastering consultation. Take a couple of your most representative mixes, and buy an hour or so at a good local mastering service. Your aim here is not to actually master anything. Just turn up with your mixes, and sit with the M.E. and listen to them. Pick the M.E.'s brains - ask him/her for comments. How is the EQ? How do the spacial elements hold up? Too much reverb on certain instruments? Is there some mud or HF 'fizz?' How's the definition? Any avoidable distortion present? Phase issues? Mono compatibility? The key question to ask: "Do you think my mix is ready for mastering, or should I try again, and on which elements should I concentrate?"
The insights you get at such a meeting might help you immensely. You might decide to mix your songs again, but now you have some data to go on. You might choose to make some more changes in improving the acoustics of your mix room. Or the M.E. might tell you that it all sounds really good, and he can do a great job in mastering for you.
I usually suggest avoiding online mastering services, although some are very good, and you can certainly use them to compare different facilities later. I believe that the benefits in attending mastering sessions outweighs the slightly increased cost. You will develop a relationship with your M.E. and you'll pick up lots of tips on how to make your mixes better. Listen and be gracious - you might decide that this particular M.E. isn't right for your project, but you gain experience. You can then continue to audition mastering facilities until you find one with which you're comfortable. Relationships are all important in the music business, and remember that a good M.S. is a knowledgeable person who, if s/he likes you, could be another great connection to have.
To find a good mastering engineer, ask around in your area. Ask studio engineers. Ask other artists. Even if you have to drive a few hours, go there in person. Talk to M.E.s on the phone. Look around for M.E.s that have cared enough to put helpful articles on their web sites describing their services and explaining the various options which they offer.
Additional reading:
Here are a few "getting started" ideas, tips and tricks for recording and mixing electric guitars.
|
|
LESS EXPENSIVE |
MID-HIGH RANGE |
GENEROUS BUDGET |
|
MICROPHONES (Close) |
Shure SM57, Audix i5 (both good on snare drums) Cascade Fathead II |
Shure SM57, Audix i5, ElectroVoice RE20 (good on kick drums and some vocals), Sennheiser MD421 (good on drums too) Beyer M160, Royer 121, Cascade Fathead II |
Shure SM57, Audix i5, Sennheiser MD421, Royer R-121, 122 and 122V, Coles 4038 (great on drum overheads) |
|
MICROPHONES (Room) |
Cascade Fathead II, any good multi-pattern condenser mic especially Røde |
Cascade Fathead II, any good multi-pattern condenser mic. |
Neumann U87, Coles 4038 |
|
PREAMPS & COMPRESSORS |
Find a used M-Audio Tampa on eBay, Presonus Tube Pre with a Comp 16, Summit 2BA, Drawmer MX 60, Summit TD-100, GT Brick (active DIs and preamps) Focusrite Trackmaster Pro |
Great River MP-2NV, Universal Audio SOLO 110/610, Empirical Labs Mike-E, Purple Audio MC-77. Think about a 500 series rack or lunchbox if you're planning on expansion. |
Empirical Labs Mike-E, Empirical Labs Distressor, UA 1176, Purple Audio MC-77, EAR 660, UA LA-2, Thermionic Culture Vulture, Thermionic Phoenix SC, DW Fearn |
• Parametric EQ: EQ section with controls for frequency, gain and Q.
• Graphic EQ: An Equalizer with a number of slider controls set on octave or third octave frequency centers.
• Bell EQ: An EQ with a peak in its response.
• 'Q': How wide or narrow the range of frequencies that are affected by an EQ setting. A high Q curve is narrow and a low Q curve is wider.
If a specific band is centered at 1 kHz. a high Q setting will only boost or cut frequencies right around 1 kHz and not affect the signal too far on either side of the selected frequency. High Q settings are used more for surgical adjustments (sometimes called notching) such as removing a ring from a drum or mud from a guitar sound. EQ settings set with a high Q are less obvious, as fewer frequencies are affected, allowing more cut to be used.
A low Q setting affects a wider range of frequencies either side of the selected center frequency. Centering also 1 kHz with a low Q setting, the same curve looks like this:
Lower Q settings are generally used to change the character of a sound, rather than correcting a particular problem.
• EQ: Some will argue, but I believe that any boost of over 5 dB is an indication that you've done something wrong at the recording stage (unless you're deliberately creating an effect). A different mic, different guitar or amp settings or a different mic position would be a better solution.
Your ears are more capable of hearing a boost than a cut at the same frequency. Generally, a 5 dB boost is in a mid-frequency is about as obvious as a 9 dB cut, depending on the 'Q' You can use this to your advantage. To find a frequency where there's an annoying element to the tone of the instrument, first boost the frequency in the approximate area where you think the problem is. By sweeping through the frequencies with an increasingly higher Q, you'll be able to isolate the tone - it will be much more pronounced. Now cut the gain at that point so you're notching out that sound. Balance the Q and the level adjustment so that you're reducing the sound without affecting the sound in the surrounding frequencies. Remember, do this final adjustment listening to the whole mix; you'll often find that less notching is needed than you would think if you listened to the track solo'd.
• EQ: First, get rid of the low junk that shouldn't be there. Hopefully, you recorded using an HPF set at 80Hz, but you'll probably need to scoop out more than that. For lead parts, you'll probably set the HPF to a higher frequency - maybe 150 Hz. The trick is to make room for the bass, while not thinning the sound too much. A common problem is that guitars and bass clash, and the sound is indistinct and muddy, and the bass becomes swamped by the low end of your guitars. Another reason to make EQ adjustments when hearing the guitars and bass in context - solo'd won't work hear, and you won't be impressed with the guitars in isolation.
Secondly, there can be muddiness anywhere from 100 to 500 Hz. Find out where it is by setting a fairly narrow EQ dip (higher Q) and play with the inverted bell until you find the spot where you've got rid of the less useful mud and achieved some definition.
Unfortunately, the effective mid-range of guitars (around 1 – 4 kHz) occupies the same space as vocals. So getting them to coexist happily is one of the challenges you will face when mixing. As the vocals need to be prominent in most cases, you'll have to learn how to use EQ and compression, delays and reverb together with riding faders or using automation in your mixing techniques to make these two components work together.
• Bypass: It's there for a reason. Use it often, and once again, make adjustments in context, not solo'd. It's all too easy to tweak your way into a worse sound - be critical, and if your changes aren't adding to the sound rethink your approach.
• Room Mic: Your room mic can greatly increasing the 'fatness' of your guitar track. You can tweak the apparent size of the room (from studio iso booth to arena) with some judicious delay in front of the room track. Ten miliseconds of delay creates the impression of roughly 10 feet of added distance. It's fake, because it messes with the reflections ratio of what would be natural in a bigger room, but it can work.
In the mix, the room mic can really help create some space, and set right in the mix, can add a different element of 'fatness.' Experiment with a touch of a really short room reverb preceded by EQ rolling off low end and higher frequencies, unless you recorded in a large bright room, in which case the natural room reverb may be all you need, particularly when well processed. Pumping compression can be a great effect too - medium attack and short release. Mixed just behind your main guitar bus can add a great dimension.
Speaking of the room mic, one trick is to use the vocal track to 'duck' the room track - just set up a good compressor with a side chain driven from the vocal track(s) and adjust your compressor to clamp down on the room track during the vocals. This is great to get some movement and dynamics into your track, and help create a space for your vocals which are roughly in the same part of the spectrum as the guitars. Just don't set the attack and release too fast or it'll sound fake.
• Stop listening to solo'd tracks. To isolate a noise, a bum note, or adjust broad effect parameters (to get the routing set and make sure the settings are approximately right), go ahead, but don't waste your time fine-tuning guitar sounds in solo - they almost certainly won't work when in the mix. You need to make these changes when listening to your track in context with the other instruments and vocals. You'll be surprised how often a great sounding guitar track 'in the mix' sounds quite awful solo'd.
• Thickening: Create one or two extra tracks. Make copies of the guitar tracks or regions. Insert a pitch shifter plug-in on each copy, and adjust the cents slider slightly. Also EQ one track so that it is different from the original - you can get brutal with HP and LP filters.
• SubMix channels: Useful for controlling. processing and maybe automating your guitars. Instead of outputting your tracks to the mix bus, route them to a new bus and bring that bus up on new track(s). Now you can add additional processing (see the compression section below) or add the final touches to get them sounding right. Pan to taste - although I often prefer to not use extreme L-R, but instead narrow the stage slightly - your material will determine the best choice. Experiment with adding a really fast delay to one side, returning to the other, with different EQ.
• Dual Stage (Serial) Compression: Often, two compressors are better than one. You need two compressors (or plug-ins) of different types. The first is set to just control the peaks. Use a fast compressor (an FET compressor is good) with the threshold set so only peaks are being affected. Use a fast attack and a fairly aggressive ratio, depending on your tracks. Follow this with another compressor (perhaps a Opto or VCA type) where the threshold is set much lower, with a lower ratio and slower attack than the first compressor, which will affect the body of the sound and help add sustain. Depending on your track, reversing this order of compressors can work better. See here for more details on compressor types. Finally, if you're sending your guitar tracks (perhaps the rhythm guitars) to a guitar sub mix, you can apply more yet more compression on the bus, maybe with a muti-band compressor which can make it play better with other elements of your mix, especially vocals.
This technique often works well on other instruments too, drums and vocals in particular. Set right, two compressors together adding quite heavy compression can sound much more natural than one.
• Parallel Compression.
A technique often used for drums and vocals, this works really well for guitar tracks too. The idea is to have some/all of your guitar tracks routed to a sub mix. Then send the individual guitar tracks to another bus. This bus gets EQ'd and heavily compressed. Bringing the return of the parallel compression bus up so it's just below the level of the main guitar sub mix; this can add a great thick, fat sound without affecting the main attack and definition of your original guitars left unaltered on the GTR SUB (see below). Here's a simple routing which may help you set this up. In the example, Logic is used, but you can adapt it to your system easily.
There are several options which may apply to your particular tracks - whether you send pre/post fade is a consideration, and depends on whether your guitar parts are mostly static, or if you're likely to be riding these levels in a mix. Remember that if set to post fade sends, changing the send levels will affect the effect, as the compression change depending on the levels seen at the compressor's inputs and the threshold set.
That's it for now. Please add additional tricks and tips in the comments section. If this brief article encourages you to try to create your own sounds, then great. Above all, have fun, and help keep alive the concept of developing great rock guitar sounds and textures.
A compressor can be described as an automatic volume control.
By manipulating a compressor's settings the incoming signal can be changed: louder parts of the signal can be reduced in level, leaving the quieter parts of the signal unchanged. This reduces the dynamic range of the signal. The compressed audio might then be raised in level so that the quieter sounds (or part of the sound) are louder, but at the same time, the loudest parts are not too loud, as they have been controlled by the compressor.
This article attempts to explain the basics of compression, the types of compressors and some suggestions on selection, both hardware and software. My focus is mostly on hardware compressors, but the terminology and uses translate to plug-ins.
A compressor usually has several controls:
Threshold
Threshold settings adjust the level at which the compressor starts working. A lower threshold (say -30dB) setting makes the compressor affect more of the signal; a higher setting (say -5 dB) will affect less of the signal. Some compressors (UA 175 and 176 and their 1176 offspring, for example) don't have a threshold control - the input control determines the amount of applied compression).
Compression Ratio
This control determines what happens to the signal's level once it exceeds the threshold. If a ratio is set to 6:1, a signal exceeding the threshold by 6 dB will be reduced to 1 dB by the compressor. A gentler ratio of 2 dB will cause reduction of gain to 1 dB as it passes the threshold. A compression ratio of 1:1 means that there is no compression.
The higher the ratio, the more extreme the compression will be. Ratios from 1.5:1 up through 6:1 are usually considered to be fairly gentle, whereas ratios above 6:1 are considered harder. Ratios above 10:1 are often more used as an effect, rather than subtle gain control, and may be found in parallel compression uses, or really compressing drums or other instruments when extreme crunch is wanted.
Limiting
A limiter, either as an additional circuit within a compressor, or as a stand-alone device, can be thought of, in simple terms, as a compressor with an infinite compression ratio. A limiter with 'no-overs' setting is sometimes referred to as a 'brickwall' limiter.
Attack
The speed at which the compressor kicks in. This is often measured in milliseconds (but is really a convenient way to reference and compare the change over time - not a precise measurement).
The attack time affects how quickly gain reduction is applied as the signal crosses the threshold and the ratio setting is applied. A slow attack allows the transients of the sound to pass through largely uncompressed, but a faster attack will act quicker, and compress most, if not all, the transients (depending on the type of compressor used).
Imagine a compressor applied to a snare drum. If the attack is slow, the initial crack will pass through, but the sound of the drum - the 'body' of the sound - will be compressed. This can give a much thicker sound to the drum. If the attack is fast, the initial stick sound - the crack - will also be compressed.
Choosing the correct type of compressor to use on individual instruments is a vital part of using compression.
Release
The release time is when the compressor is reducing compression to the level set by the ratio once the level has dropped below the threshold, to the level it was before passing the threshold.
Some compressors are designed so that the attack and release times are not adjustable by the user but depend on sensitivity or input level controls and the content of the program material.
Knee
Some compressors have an option to adjust the speed at which the signal is compressed to the set compression ratio as it exceeds the threshold. A 'hard' knee abruptly kicks in; a 'soft' knee setting is more gradual and results in a less obvious transition from input signal to compressed output. Some compressors have their own terminology for the 'knee' function - dbx's softer knee circuit is named 'OverEasy.'
Peak .v. RMS
A compressor which affects the peaks of a sound combined with very fast changes in gain reduction is more obvious. An RMS compression will average the incoming audio as its level is compared to the threshold and often the compression effect is more controlled and the compression applied to the overall sound is smoother. RMS compression is more likely to change the 'perceived' loudness of a signal.
Make Up Gain
Enables the adjustment of output level to compensate for any attenuation caused by the compression settings.
How Compressors Work
Most compressors split the incoming audio signal into two paths (Feed Forward design). One, the actual audio signal, is sent to a variable gain amplifier (VGA); the other, called a side-chain, goes to a circuit which monitors the input signal and compares it to where the threshold is set. When the input signal exceeds the threshold, the sidechain control circuit sends a command to the VGA telling it to turn the gain down and by how much (ratio).
The sidechain is in fact there all the time, whether externally controlled or not. The modern reference to the sidechain has come to mean external control of that command sent to the gain circuitry to, for example, compress in tandem with another source, and/or to use the (often EQ'd) side-chain input to control the VGA. It's analogous to having a little engineer on speed living within your compressor furiously turning a volume knob up and down. His job was eventually outsourced to the light bulb, in the early days of recording, as he wasn't fast enough.
Some compressors use a 'Feed Back' type of design where the control and detection side-chain circuitry is tapped after the VGA. The 1176, Fairchild and API 525 are examples of this design. The famed API 2500 stereo compressor can be switched to either circuit types.
The gain circuit in hardware (now replicated in software) was based on several different types and each type has its own characteristics which affect the compressed sound and suitability for the instrument being compressed, including voltage controlled amps, diodes and photocells ('opto' - in fact was originally a regular light bulb) and others.
Compressor Types
Note: Some compressors are hybrids including more than one type of circuit design, or a design which doesn't neatly fit into one of the four main types of compression listed below.
Example: Empirical Labs Distressor and I would include also the Neve 33609 along with many others.
More important than any of this is the audio design of the amplifier section itself. Don't get caught in the marketing hype - if it sounds good, it really doesn't matter what the side-chain control circuit type is.
It's the way the audio is handled that's important, regardless of how the compressor did its job internally. Descriptions are given below just so you know the main differences. Listen to a compressor and see how it sounds without any gain reduction. Treat it like a preamp and play around, and see how it behaves as gain reduction is applied, and you change the parameters. Notice any harshness and grit as attack is set faster, the threshold and ratio are changed, and you alter the release time.
VCA (Voltage Controlled Amplifier)
VCA compressors are the most common type. The VCA is fast, and can change gain in response to many different parameters. The VCA compressor is the workhorse of compressors, and can used for almost all types of compression needs.
Examples of famous VCA-based compressors include the dbx 160, Focusrite Red, and SSL G Bus Compressor
OPTO Compressors (ELOP - Electro-Optical)
The earliest models were literally a light bulb shining on a photoreceptor (cell or transistor) which converted the dimming/brightness into a control for the VCA. Typically, the opto compressor is smooth, not as fast in responding to changes as an FET or some VCA compressors, and depending on the model, are often used for vocals and bass guitar. The slower response time is compensated for in some designs so that the attack performance can be quite good.
The most famous opto compressor is the all-tube Teletronix LA-2A, and later the UA LA-3A and LA-4 solid state optos. Examples of modern opto compressors are Tube-Tech CL 1B and Avalon.
FET (Field Effect Transistor)
Like an 'opto' compressor, but without the light source, FET compressors have the common characteristics of being very fast and suitable for a wide variety of applications. They are also considered to be less 'colorful' than opto compressors, but this is more to do with individual models.
Examples include the UA 1176, Audio & Design and the Allison Research Gain Brain. Modern FET compressors are made by Drawmer, Daking and the Purple Audio MC77
Vari-μ (Vari-mu)
Denotes the use of a type of tube that through re-biasing achieves compression. The Vari-μ often has a softer knee and very smooth characteristics which makes it highly useful for compression on a mix bus, and often for vocals and bass use.
The most famous models include the Fairchild 660 (mono) and 670 (stereo). Modern manufacturers of this type include Pendulum and Manley
What is a Leveling Amplifier?
Interchangeable term with many compressors - it's really not a separate device from a compressor - it depends more on how the compressor is set, but some are more suitable than others for the 'leveling' application. The leveling amplifier is a compressor set to a medium attack time, a slow-ish release time, and a low threshold.
The idea of a leveling amplifier is to be 'always on' so that the audio input is constantly smoothed. The perceived loudness of the audio becomes greater because the dynamic range is reduced, allowing for the overall level of the audio to be increased without the danger of clipping, thereby increasing the levels of quieter signals. In fact, the Teletronix LA-2A is often described as a Leveling Amp.
Plug-ins
There are some very good software compressors. I prefer to consider them in their own right, and their sound characteristics, strengths and weaknesses, rather than rely on the sales hype proclaiming "just like the 1176" or "exactly like an LA-2"
No, they're not. They might be very good, and sound fine, but they are not a direct equivalent. Hell, even old 1176s and LA-2s hardware of the same revision would sound different: especially tube devices as the individual components all age differently and that affects the sound. Many seasoned guitar veterans will tell you how their favorite amplifiers sound their best just before a tube blows - same goes for old tube preamps and compressors.
The following references are not intended to be a suggested or comprehensive list of plug-ins - I only list those that I either own, or have used extensively, and found that for me, they are just excellent. Most of these manufacturers have demos, so try them out.
Before I forget: don't buy a plug-in unless you've really figured out the ones you already own and learned how to set it effectively. For example, I use Logic Pro for my usual DAW. When Logic 8 was released, Apple greatly improved the standard Logic compressor and added some new circuit emulations. The original 'platinum' circuit is still there for backwards compatibility, but you probably won't want to use it, as it's not very good. For many Logic users, this improved compressor may be all you need; it's really quite good for a wide variety of material.
But if you're looking for something different, read on:
An amazing plug-in, with a fairly steep learning curve. The Solera has some features not seen in other plugs which are very effective once you've figured out the best uses for them. For example, try out what they call Hysteresis and Angel's Share.
A great workhorse. Does most of what you want a decent compressor to do, without any fuss. Colored or clean - you're choice.
Another good plug-in. Works very well and is very flexible. Not transparent: it imparts coloration but can be quite subtle. Quite CPU friendly and low latency.
Not a 'normal' compressor, it's designed to emulate the sound of a tape machine. Extremely effective. The pack contains five different plug-ins all with their own unique characteristics. No AU version available.
Again, not a conventional compressor plug-in, it does what it's name suggests. But rather than only slapping it on a mix bus, try it on a drum bus as the parallel compressor, or a guitar submix.
Kjaerhus Audio Mastering Limiter MPL-1 Pro SE
Very inexpensive, very good, very transparent unless pushed. Touted as a mastering limiter, I've found it highly effective on submixes. However, the lack of activity in their forum and apparently minimal sales effort is a bit worrying - grab this while you can.
Audio Damage's Rough Rider Pro
OK, this thing isn't clean, subtle or transparent. It's a beast.
As Audio Damage says on their web site: "Other compressors are kittens, rainbows, and gumdrops. This one is a 40-foot-tall fire-breathing lizard." It's cheap. It's outrageous on rock guitars. And I like any manufacturer who has a great sense of humor, with products which are very inexpensive, and work well. There's even a free version of Rough Rider.
There are some excellent plug-ins created by AirWindows. They are very good, no-frills, inexpensive plugs for a wide variety of uses, and are well worth checking out. There are also several free plug-ins available too.
Here's a good beginner's guide from Samplecraze:
More Reading
Sonnox Oxford Dynamics Technical Details
S.O.S. Advanced Compression Techniques Part 1
S.O.S. Advanced Compression Techniques Part 2
S.O.S. Guide to Mix Compression
RANE Dynamics Processors - Technology & Applications
I got together for some drinks with a few friends who are engineers and producers , and inevitably the 'loudness wars' subject came up. How do we deal with the demands of 'it's not loud enough' from both labels and artists?
This isn't new; I remember mastering vinyl LPs years ago, and even then there was an effort from producers and artists to 'get it loud.' There was another issue then too; more bass = less program time per side, so the mastering engineer had to juggle how much low end could cut .v. running time. There were complaints about players 'jumping the tracks' on particularly loud, bass heavy, cuts.
Now, when the audience listens to the radio, they have becomed accustomed to that 'compressed' sound. To them, it sounds like the industry standard - that's what songs are supposed to sound like. It sounds like a 'hit.' That's what labels and artists want their releases to sound like, but now they want it in their car, at home and on their iPod. Listening to an album, in order, all the way through is a thing of the past: iPods and CD players are set to 'shuffle' internet stations are playing isolated tracks, and more and more the audience is creating playlists, combining disparate songs into a new, and unanticipated, replacement for the album.
You can fight this all you want, but if you make a living in this business, you'd better make your music also sound like their notion of a 'hit.'
What I'm doing now is mixing and mastering two distinct and very different versions. One for streaming, the net, radio promo releases and then an entirely different mix and master which is solely intended for the manufacturing of CDs. This, to me, is a happy compromise. Soon, though, with the prospect of few physical CDs actually being released, this tactic won't work, as the delivery medium will be mostly lossy digital files so most music will be subjected to this lowest denominator.
Examples:
These are the stats from a well-known rock release. The numbers to pay attention to are the Peak Amplitude, Minimum RMS power, Average RMS power, Maximum RMS power and the Clipped Samples. The Average RMS power is the stat. that is going to make your song sound louder, less loud or about the same as commercial releases with which you might be comparing your song. This sucker is LOUD, but there were more than a few disparaging remarks on the net about how bad it sounded - distortion, no dynamic range and generally fatiguing to listen to.
Here's a waveform screen shot of this release, at maximum zoom. Notice how many peaks of the waveform are essentially cut off. The peaks are flattened.
Here's a rock song that I recently mixed, and before it was mastered.
Click here to play an audio sample of a section of a song before mastering.
Typically, when I mix, I try to have my outputs peak at around -6 dB. This is to allow enough room for the mastering engineer do his stuff, and I did want it louder. But not to the point where the music suffered and had the life squashed out of it. Here's the corresponding stats. after it had been mastered.
Notice that the Average RMS value after mastering was 12dB louder. This (very roughly, not taking into account any weighting) means that my mastered version, on average, was apparently 4 times louder than my mix. Here's the waveform:
Click here to play the same section as the earlier mix sample, after mastering.
So how has this changed over the years of rock music? Here are the stats. from The Beatles "Back in the U.S.S.R" which wasn't intended to be a quiet song.
and the waveform:
More Reading:
2.19.09
Even More Reading:
I was browsing the S.O.S Mastering forum today, and came across some posts and links from John Scrip at Massive Mastering near Chicago. In his blog, I read some good articles here; kudos to John for his contributions to this debate.
When the Shit Hits the Fan (Part 1 - OSX and Maintenance)
Here's a guide of the steps to take when disaster strikes. Much of what follows is general advice to do with keeping your Macintosh computer running smoothly (and what to do when it doesn't), as this relates to a Mac used as a Digital Audio Workstation (DAW).
Then, in Part 2, there are some suggestions about solving Logic Pro (my DAW of choice) startup and crashing problems.
Backup and Routine Maintenance
Of course, prevention is vital. Sooner or later, you'll run into trouble but if you have taken some preventative steps, and you have backups, a potential disaster becomes no more than a minor annoyance.
You need three backups.
1. At least two backups of your vital data, one offsite and one conveniently available which can be quickly restored. OS 10.5 (Leopard) introduced Time Machine. Use it! It's a good first step, and allows you to quickly restore files. Use an external drive for this, and USB is fine. It doesn't need to be fast, as once your initial backup is complete, hourly backups shouldn't take very long. Don't use an internal drive for Time Machine backups - if your computer dies because of a hardware problem, you don't want to extract hard disks and put them into another machine, or install them in a firewire HD enclosure.
For offsite backups, check into Carbonite, Mozy, or IDrive. If you have a .mac or mobile .me account, that can be used too, or even the great DropBox. All of these cost money, but one day you'll be glad you made the investment. You need offsite backups in case of fire, theft, flood, angry partners (business or domestic) or an incontinent pet. For offsite backups, don't bother to backup applications or the system: you already have those (hopefully) on the original media. Just concentrate on your data, including anything you have on secondary drives (Logic projects, for example). Similarly, don't backup samples, unless you created them. 3rd. party stuff is already on the CD/DVDs they came on, and any updates to these libraries can be re-downloaded on line.
2. A clone of your essential system and applications. Ideally, this should contain your latest, solid, system with your essential applications. In my case, I created a clone with SuperDuper and use their sandbox system for day-to-day work, while keeping a pristine 'safety' system on another hard drive or partition, which shares data with the original. This means that I can try out upgrades and new software. As I also beta test software; this method means that I can play around with iffy versions of software without putting my main system in jeopardy. If you have any flavor of a Mac tower computer, use one of the internal bays for your clone drive. If you use a Macbook or an iMac, use an external firewire drive. Using partitions, you could use this drive for storing samples too, although I prefer not to use this drive for anything other than the safety clone.
3. Also use external firewire drives to backup everything. You could use a SuperDuper script to run scheduled incremental backups once per day, when you're not working. This drive should be turned off, except when backing up. Finally, Drobo is a great solution if you have the budget.
Hard Disks
When you buy an external hard drive, the chances are that out of the box it won't be formatted properly to work with a Mac - it' will probably be formatted as FAT32. So before using it for real, plug it in, and run Disk Utility. Format and name the drive, choosing Macintosh OSX Extended. If it will be a system boot drive use the GUID partition map if you're using an Intel Mac. I'd suggest that journaling be off for a data drive (audio projects, samples etc) but on if it's a system volume.
Common Problems
Here's a couple of typical cries for help:
"There was a power outage but on rebooting I had somehow corrupted my drive so now when I plug it in all it says is this disk cannot be read"
Usually, when this happens, the hard drive is physically fine - what has happened is that the disk directory has become scrambled so that the OS can't find anything, including the directory itself. Disk Warrior is the tool of choice for these problems and every Mac user should own it, in my opinion. If Disk Warrior can recognize the disk, it will usually be able to construct new directories and your disk (and your data) will be fine.
"My hard drive is making an odd clicking sound"
It's in its death throes. Back it up immediately, and replace it. After it's backed up, you can reformat it, and maybe use it for some non-essential storage in another computer but usually clicking or thumping noises mean that it's about to die. Same advice if it takes a lot time to appear on the desktop, accompanied by a grinding noise.
Using a UPS (Uninterruptible Power Supply) is also a worthwhile addition to your setup. Plug your computer, and any external hardware into a UPS and you'll have some time to shut down gracefully in the event of a power outage and as a bonus usually some power conditioning. APC and Tripplite are good manufacturers of UPS systems.
Partitions
Partitioning is useful for housekeeping purposes, but worthless if you think that somehow making multiple partitions will speed up performance. HD platters are connected - if your system is looking for stuff on partition A at the same time as streaming audio files from partition B, performance will suffer. This is also why recording to your system drive is a Bad Idea - as the OS is trying to access application data and resources, the application is also trying to retrieve audio data at the same time. Keep your projects on one separate physical drive, and everything else (system and applications) on another drive. For those who use huge sample libraries, a third drive is a good idea - you can also use the sample drive for Logic's Apple loops if you use them.
Hard drives are cheap. As I write this (Jan 2010) a Seagate Barracuda 7200 rpm 1TB or 1.5 TB drives are around $100 and are perfect for backup drives and sample drives, and even audio project drives unless you record a lot of audio/playback, in which case a 10,000 rpm drive can give greater performance. I use Western Digital's 300 GB Velociraptors for audio drives.
Although they are much more reliable than they once were, hard disks will go bad - it's not a case of if, it is a case of when. Remember this, and plan accordingly.
Bad Memory
This too ranks up there with disk permissions, defragmenting and all the other near myths. I tried to count how many Macs I've owned (I got to about 30 and gave up) and I have never had a bad memory stick. I have cracked a edge connector, and I dropped one once and stepped on it, but I haven't had one that has gone bad. I'm not saying it can't happen - I'm saying it's rare. If your Mac is crashing, suspect bad software, a corrupted or damged drive or bad hardware in that order. As the old saying goes, "if you hear the sound of hooves thundering down the road, first suspect horses, not zebras." If you really think you might have bad RAM, use the hardware test disk that (probably) came with your Mac, or download Rember (memtest GUI)
System Maintenance
With later versions of OSX there really isn't much you need to do. However (perhaps because of a mild case of OCD) there are still some preventative steps I routinely take.
Alsoft's Disk Warrior. Buy it. It's one utility that will get you out of trouble more than anything else I've found. When you buy it, install it on every bootable drive you own, as it runs slowly from its CD. Many apparent hard disk problems are really disk directory glitches where files seem to be damaged, missing or corrupted. Disk Warrior creates an optimized disk directory, fixes many different types of file/directory problems and your disk runs noticeably faster. I run this once per month.
OSX Disk Utility is handy too. If you notice a problem, run it immediately, using 'verify disk' if you don't have Disk Warrior.
Permissions
There's a bunch of folklore surrounding permissions. Reading online forums, you'd think that incorrect permissions account for 99% of problems, including spontaneous combustion of power supplies, logic board failures and world hunger.
In fact, damaging permission problems are quite rare, unless you regularly access other people's hard disks or data files/folders which you use for working on projects. Problems can occur when installing new software, so if after installing something new you notice that something changed, it's worth running a permissions check. Apple's Disk Utility will do this for free. There are lots of utilities which will correct permission settings, such as Cocktail or Onyx (be careful with Onyx, and read the instructions). I prefer to use Atomic Bird's Macaroni which runs in the background and handles OSX maintenance scripts automatically on its own schedule. If you use multiple user accounts, be sure that any installations are always performed when in the account you will be logged in with for using of the software.
For those cases when you want to force permissions to allow yourself access when OSX thinks you shouldn't have them, BatChmod is invaluable, but please read the instructions first, and don't use it at all if you're unsure about what you're doing.
If all of the foregoing seems a little daunting, remember that it only takes a few hours to set up: going forward, it's all mostly automatic. Try it, and sleep well at night.
System and Software Updates
As soon as Logic 9 or Snow Leopard became available, the support forums were full of messages complaining about bugs, unexpected behavior, hardware not working, plug-ins not loading etc. What do these people expect? Never use the X.0 version of anything, unless you've followed my earlier advice about using a sandbox. In that case, it doesn't matter as you can easily revert to your existing virgin system.
Unless there's a compelling reason to upgrade, wait a few weeks and let the early adopters report back and then make your decision. I know it's hard to resist: when Logic 9 was released, it included many features that Logic users have begged for. But unsurprisingly, many of these features didn't fully work until the release of 9.0.2 and even now some features are still problematic. Snow Leopard has nothing that I need, so I don't use it (yet). In fairness, it was never intended as a feature release; it was more an optimization release, but unfortunately broke some apps and introduced some hardware compatibility. I bought Logic 9 (or Studio 2, whatever it's called) and Snow Leopard and tried it out (but on a separate drive). Some new features are very nice, others need updates. I continue to use 8.0.2 in 10.5.X which is quite solid. Eventually I'll jump over, but not yet, particularly when I found that the RAM footprint of LP 9 was at least 500MB higher that LP 8 and some of my 'heavier' current projects wouldn't run in 9.
Summarizing, don't jump on to the new versions of anything until you're sure that (a) you need it, (b) you have backups to revert to, and (c) your essential hardware and software will work with the new software and (d) you've upgraded all necessary drivers and versions to support the new major upgrades, and found out before if updates are even available.
Finally, don't install different updates at the same time. Ever.
Update OSX first. Test thoroughly. If all is well, then install major application updates. Test thoroughly. Install add-ons, one by one, testing between each one. If you install them all at once, you'll have trouble figuring out which new install caused a problem, should one surface.
Reinstallation of System Software
Theoretically, this shouldn't be necessary. However, 'shouldn't' isn't as good as 'never.' Every time some software freezes, hangs, unexpectedly quits or just plain crashes or you have to use 'force quit' there's a chance that somewhere in the depths of OSX's system files, a bit got flipped, an entry in a shared library got changed or some other tiny fragment of a file became corrupted. Over time, these minor errors add up, and in turn affect other files, which causes other corruption, and so on. The domino effect.
The solution: periodically run an 'archive and install' from your system disks, and run all the updates. It takes a few hours, but typically I do this every six months, or if I notice any instability (quite rare). Once you've done it a couple of times, it's very easy, and is a great opportunity to dump stuff you've acquired over time that you really never use. It's quite surprising how smoothly everything runs afterwards.
Start Up Issues
If your Mac won't boot, or you get kernel panics or other problems, restart holding down the option key after the chimes. From there you can choose another start up disk. Obviously turn on any bootable external drives, and attach a wired keyboard - wireless keyboards and mice won't help you as the software required to initiate wireless access loads later than the disk selection 'option' command needs.
All the OSX keyboard short cuts are here (with links to further explanation), but the following are the most useful when you have a problem:
C : Forces most Macs to boot from the CD-Rom drive
T : Target Disk Mode (FireWire) – Puts machines with built-in FireWire into target Disk mode
Mouse Button Held Down : Ejects any mounted removable media.
Shift : Boots into safe mode. This can take a while. Safe Boot forces a directory check of the hard drive. This is identical to using Disk Utility's Repair Disk or the fsck -fy terminal command.
Command-V : Boots Mac OS X into "Verbose Mode", reporting every console message generated during startup. This shows what’s going on behind the scenes with your machine on startup, and can be very useful.
Command-S : Boots Mac OS X into "Single User Mode" – helpful to fix problems with Mac OS X
Command-Option-P-R : Erases PRAM if held down immediately after startup tone. You will hear a chime when this has erased the PRAM; often it's suggested that you hold this key combination for 3 chimes to completely flush the PRAM. (rarely needed, check first)
Command-Option-O-F : Boots the machine into Open Firmware
Command-Option-Shift-Delete : Forces your Mac to startup from its internal CD-ROM drive or an external hard drive.
Disk Fragmentation
I don't defragment drives, mostly because I keep my drives with usually 30% minimum free space and OSX does a good job with keeping files under 20MB defragmented, which easily covers 95% of the files on my system drive. On my audio drives, I keep projects on them only until they're completed, and then I archive them. So the only projects on my audio drive are those with which I'm currently working. Copying files from one disk to another defrags. them anyway, so I don't concern myself with defragging drives. Also there's one important reason not to - if you use Logic 9.0.2 or earlier, and use long file names, Logic will truncate them and include numbers referring to the position of that file on the disk. If you defragment your drive, Logic can 'lose' you files, and you'll have to manually find them within the Logic project. Time-consuming, annoying and definitely not fun.
Your file names should be kept at under 32 characters including the file extension. So if your audio file is named "Mysongname-gtrsolo-verse3-take4.wav" it will be truncated by Logic to something like "Mysongname#A27596.wav" If that file is then moved (by defragmenting, for example) you're screwed. 
The next time you open that project, you may get an unpleasant dialog saying: "Mysongname-gtrsolo-verse3-take4.wav" cannot be found - options: search or skip.
Be careful!
More Help:
Resolve startup issues and perform disk maintenance with Disk Utility and fsck. A very useful article for when you don't have an alternative disk to boot from (and you have unwisely ignored all the above advice).
Guest Account
Always set up a guest account on your Mac. In case of problems, log in to this guest account first and see if the problem persists. If it's gone, you have some corrupted files in your main user account; probably preferences in the Library/Preferences. So if you have a problem with Logic launching, for example, this is the first step. Don't use this account to install updates or new software.
Keep an Old Machine
Having a second Mac is useful. When you buy a new machine sometimes it's tempting to sell or give away your old machine. Try to keep one - even an older Mac is a perfectly capable machine for internet use, running a browser and downloading updates of software. Even simple stuff like archiving and installing can be sped up while your main computer is having major restoration; you can be downloading updates, for example, or reading online support documentation.
Next: When the Shit Hits the Fan (Part 2 - Logic Pro)
When the Shit Hits the Fan (Part 2 - Logic Pro)
Logic problems generally fall into two categories:
1. Start up, crashes or instability problems.
2. Performance issues.
For (1) I'll try to list a few pointers to get Logic working.
But for (2) the performance issues vary so much, and are so dependent on an individual set up and system, a book could be written about them, and the various permutations of software, hardware and system resources make generalizations very challenging. I'll list a few online resources which might help with performance issues, but I can't cover them in great detail as each problem depends on the exact configuration of the system used, software instruments, effects used etc.
Logic Crashes on Start Up
Be methodical. Try these steps one by one to isolate the possible cause.
Before you begin, are the problems you're noticing confined to Logic only? Do all other applications open a run fine? If not, then you may have a OSX problem, not Logic. One useful experiment is to open GarageBand. GB uses many of the drivers and resources as Logic; if GarageBand works fine, with your audio interface (if you have one) that gives you useful information, and you can skip some of the suggestions below, particularly those regarding attached hardware.
Note: Before you move files or delete components which may be suggested, do so only when Logic isn't running.
First, look at the error message. Usually it's 'unexpectedly quit' but occasionally you just get a hang, and you have to force quit, and there is no error message. But maybe you'll get lucky and there's a reason displayed, although it's rare.
If you started Logic by clicking on a Logic project file (.logic or .lso) then don't.
Launch Logic from the Application folder. Does it open now? If yes, then you have a problem with that Logic project, not Logic itself. There could be many reasons - I'll list a few of them later.
If launching Logic without a project doesn't work, then:
Log in to your Guest Account. Does the problem still occur? If not, log in back to your usual user account and move these two files to the desktop:
~/Library/Preferences/com.apple.logic.pro.cs
~/Library/Preferences/com.apple.logic.pro.plist
Note: The "~" signifies that this is the user library, found within your Home folder. If there is no "~" in the front of the path, such as /Library/Application Support, it means that the library is the main system library in /Macintosh HD (or whatever you named your system drive).
Now restart Logic. Problem still there? If no, then you have corrupt preferences. Logic creates new prefs. on launch, so you can carry on, but know that you may have to import/recreate your custom key commands and/or the setup of any control surfaces you may have created.
Still not working?
Crash Logs
Crash logs are not intended for the end user. They're there for technical support, but primarily for developers, who can interpret these logs as the relate to their code. Having said that, the end user can still use them, and in some cases can indicate where a problem might be occurring in the processes being used by Logic.
First, go to Utilities folder - Console.
Under ~/Library/Logs listed in the left column you'll find a section named CrashReporter. In that list, you should find the item Logic Pro.crash.log
Interpreting Crash Logs
Host Name:
Date/Time:
OS Version:
Report Version:
Scroll down until you see:
PID:
Thread:#
Backtrace Section
The thread which last crashed will be identified. Let's say that the Thread# is 9. Scroll down to where the list reaches Thread 9 Crashed. Each line (frame) is numbered, starting at 0. Frame 0 identifies the last action in that thread before the crash; it shows events with the most recent at the top. The first few lines may identify a problem.
Here's a very simple example:
This crash log indicated that Thread 9 crashed. Looking at Thread 9, we see this:
Exception Type: EXC_BAD_ACCESS (SIGBUS)
Exception Codes: KERN_PROTECTION_FAILURE at 0x0000000000000000
Crashed Thread: 9
Scrolling down, we come to Thread 9
There are 4 columns - (1) Frame#, (2) name of binary (see below), (3) counter address and (4) name of debugging process symbol or a hex code (if the application developer stripped out these symbols before finalizing the application).
Thread 9 Crashed:
0 ??? 0000000000 0 + 0
1 ...lodyneEssentialRewireDevice 0x47b5b6ab GNThreadHandler(void*) + 103
2 libSystem.B.dylib 0x90fe2fbd _pthread_start + 345
3 libSystem.B.dylib 0x90fe2e42 thread_start + 34
We see that the crash seems to have been triggered by the MelodyneEssentialRewireDevice. Searching the crashlog for that component, we see that it's located here (just use your 'find' command and copy "lodyneEssentialRewire" into the search dialog):
/Library/Application Support/Propellerhead Software/ReWire/MelodyneEssentialReWireDevice.plugin/Contents/MacOS/MelodyneEssentialRewireDevice
Binary Images
In the lower part of crash logs, the components are listed currently in use, with paths.
Now we have two choices: update it by going to the manufacturer's site and downloading and installing the latest version, or just removing it to test if Logic now starts. When removing components, just move them to the desktop and restart Logic. If it doesn't fix the problem, or causes more problems, you can move it back to the same location.
Warning: Although crash logs can be very useful, it takes a while to learn all their capabilities, and the details are quite complex. The simple example above is a useful step, but explaining all the other clues you can get from logs takes some learning and experience. If you're interested in discovering more, start with this excellent Apple Tech Note
Please note: unless you have access to the code (you're developing an AU plug-in, for example) much of the data in a crash log will be meaningless as you can't backtrace to the code fragment that caused the exception.
Unless you're working with (or are) a developer, much of the contents will seem to be gobbledygook. Most crash reports that an end user might see do not contain the process labels that indicate each task in (almost) English: instead you get a representation of the task in code. So if you don't have access to which task is represented by which code, this isn't going to help much.
In the example above:
Exception Type: EXC_BAD_ACCESS (SIGBUS)
Exception Codes: KERN_PROTECTION_FAILURE at 0x0000000000000000
Crashed Thread: 9
You'll see that the Exception which caused the crash has two parts: the Type and the code. From Apple documentation, we know that there are four main types:
EXC_BAD_ACCESS/KERN_INVALID_ADDRESS
Caused by the thread accessing unmapped memory. Can triggered by either a data access or an instruction fetch. The thread state section helps indicate the difference.
EXC_BAD_ACCESS/KERN_PROTECTION_FAILURE
Caused by the thread trying to write to read-only memory. This is always caused by a data access (not very helpful, unless you know which data was being accessed).
EXC_BAD_INSTRUCTION
Caused by the thread executing an illegal instruction.
EXC_ARITHMETIC
This is caused by the thread doing an integer divide by zero on an Intel-based computer.
After the Exception code, the address which triggered the exception is shown, in this case, at 0x0000000000000000
Armed with this information, you can look a the Thread State: this is the block displayed after all the Backtrace listings. It might look like this (for an Intel machine running 32-bit, PPC is different - the eip: mentioned below is usually srr0 and for Intel running 64-bit code, it's rip):
Thread 9 crashed with X86 Thread State (32-bit):
eax: 0x427e13b0 ebx: 0x478a7e40 ecx: 0x02d418f4 edx: 0x427e11f0
edi: 0x02d418f0 esi: 0xb04bf000 ebp: 0xb04bef48 esp: 0xb04bef1c
ss: 0x0000001f efl: 0x00010206 eip: 0x00000000 cs: 0x00000017
ds: 0x0000001f es: 0x0000001f fs: 0x0000001f gs: 0x00000037
cr2: 0x00000000
Here, look for the value for eip: which is the program counter for the time the exception occurred. If that is the same value as the Exception address, we know from Apple's documentation that eip
is the program counter at the time that the exception occurred. If eip is equal to the exception address (as in this example), the exception was caused by fetching instructions. Typically this means there was a call to a bogus function pointer, or the process pointed to a bad address which, in turn, means that the stack became corrupted. If eip
is not equal to the exception address, the exception was caused by a
memory access instruction.
This is more gobbledygook to the casual user - without more detail from the developer the chances are none of the backtrace information or exception states are going to be very helpful, but with practice and a process of elimination this can be very helpful to see how curative actions at the user level affect the crash reports, and in turn can point to problem areas, particularly in cooperation with the developer.
If you're not interested in learning more about crashlogs (and I don't blame you) they can still be useful. Copy the entire crash log (for one crash, not all of them) and paste it into a message in one of the support forums mentioned in below. There are users on these forums that do understand crash logs and may be able to offer further advice.
Assuming the crash log didn't help, or there wasn't a log entry created, the next step is to remove all external devices, except for your keyboard and mouse. Disconnect any firewire or USB devices, and importantly, your audio interface. Now trying starting Logic. Logic should tell you that it can no longer use your interface, and is switching to built-in audio instead. 
For this experiment, that's good. Click ok. If Logic now launches, suspect your interface or its drivers.
Another method to check to see if your problem does stem from an audio device, start Logic without an audio driver. Hold down the Control key as soon as Logic is launched and say "No" when Logic asks if you want to use the audio drivers . If Logic launches successfully, check your audio drivers' compatibility, upgrade, or re-install.
When you launch Logic, does it 'hang' where the progress indicator says "Searching for audio unit plug-ins." If takes an unusual length of time, it may been that the audio unit cache file has become corrupted. Go to ~/Library/Caches and delete the com.apple.audiounits.cache file. It will be rebuilt the next time you start Logic.
Project Specific Issues
Undo History - Delete it. Go to Edit → Delete Undo History
Reorganize Memory - Options → Project Information → Reorganize Memory. Then save your project under a different name.
Prevention
Saving Project Files
It's rare that Logic files become corrupt, but it does happen. I try to avoid losing work by saving the Logic project file under a different name, saved to the same project folder. As my projects tend to span several weeks I've adopted a convention of starting with MySongName.logic. From then one, until the basic track is finished, I add a number (which corresponds to days worked, for me) but you should to modify this convention to suit the way you work. So I get three days of initial sessions shown as:
MySongName.logic
MySongName-01.logic
MySongName-02.logic
Then I may have a few days of dubbing guitar parts. I represent this as follows:
MySongName-gtr-01.logic
MySongName-gtr-02.logic
MySongName-gtr-03.logic
On to vocals:
MySongName-vox.logic
MySongName-vox-01.logic
and so on.
Logic creates a new set of backups (up to 10 in Logic 8/9) for each project file name. So each time you change the project file name, you get an extra 10 backups of that file, in the Project File Backups folder, which is within your project folder.
Don't forget to back everything up (See Part 1)
More Help - Online Forums
There's lots of help available online. For Logic-related issues, try:
which I think are the best.
For more general help, try the Apple forums, and search for the appropriate category for your particular issue. Macintouch is also a good resource for all things Mac.
When posting a question, for best results remember these tips:
Use a descriptive title for your post. Don't use "Help!!!!" or "Logic Won't Work" but rather something that describes your problem such as "Problem loading samples into EXS24" or "Logic 9 crashes on start up with OS 10.6" or "How do I use the same reverb for several tracks"
Create a profile, or list details of your current system, including OSX version, Logic version, interface used, details of your Mac including RAM and how many disks are used, and any other hardware relevant to your problem.
One of the top contributors in Apple's Logic Pro forums spent a lot of time compiling a list of links and resources covering many of the more frequently asked questions. Well worth browsing through this long thread to help you find topics relevant to your issue.
If you need more help with learning Logic, MacProVideo offers many tutorials, and are a great way to speed up the learning curve.
Logic and Time Machine
Time Machine works well, but on slower machines, or very high activity Logic sessions when recording many tracks with large sample library access, under some circumstances Time Machine can slow down CPU and disk access. Use these two tiny automator apps. to Launch, and later Quit, Logic. The Launch app. first disables Time Machine, and then launches Logic. The Quit app. quits Logic (after prompting you to save) and turns T.M. backups on once again.
Click here to download.
While we're talking about Time Machine, if you'd like to change T.M.'s backup interval to something other than the default one hour, go to:
/System/Library/LaunchDaemons/com.apple.backupd-auto.plist.
Open it in a text editor or Property List Editor, and look for this section:
Start Interval 3600
Change the 3600 number to some other time interval (it's in seconds)
More Coming Soon…
Fueled by suitable amounts of plonk, I had one of those "current state of recording" conversations last night. Our group consisted of a couple of old farts like me, and some young guns carving out their position in the business.
The conversations turned to choices and decisions. Choices of what techniques, equipment, plug-ins, outboard gear used in recording and mixing in 2009, and when decisions are made in the recording process.
Choices
In the age of 4 to 24 track rock recording, typically the choices were limited. For compression and limiting, there were 1176, LA-2, LA-3, maybe a Fairchild and some A&D's. There weren't that many of them either, so the tendency was to get it more or less right when the tracks were recorded, and to apply mix processing more sparingly because (a) it wasn't necessary on every track and (b) the number and type of units in the racks was limited. Of course, the channel strips often had dynamics and EQ, but even so the choices were not extensive.
Reverb options were even more scarce. A natural echo chamber. Maybe a couple of EMT plates, an AKG box, and perhaps a 250, (in the fancier rooms) and some actual tape delay with a 7" spool hung off the side of a 1/4" machine with vari-speed. Lexicon came along a bit later. We'd often pump something out into monitors in the live room or hallways and route it back to add room ambiance. So when using reverb, there may be a bit of chamber, some plate and maybe something else for a spot effect, and some delay. That's it. One of my young gun friends talked about a recent mix where he used (he tried to list them all) 22 (!) different types of reverb/delays. When asked why, he spoke of creating different spaces for different elements, but he also admitted that these subtleties often disappeared in the mix and and he found himself turning returns down at the final stages of the mix.
For rock recording, at least, maybe there is some value in cutting down on the amount and types of processing used, and instead considering using fewer electronically created spaces to achieve a more cohesive staging.
Decisions
Recording with a limited number of tracks made us make decisions early in the process. If you didn't have some plan in the recording session, you could run out of tracks. So you had to know what you were doing, what was left to record and allocate, and bounce down tracks accordingly. And you couldn't change your mind. If you didn't plan, you created a mess which made mixing more complex, and remember that through much of this early period, there was little or no automation and, until SSL came along, no on-the-fly processing changes.
So shakers appeared on guitar solo tracks, pianos popped in and out on vocal tracks and the inevitable extra bits of fairy dust were dropped in on whichever tracks had space. So often we had to mix in sections, and cut up the tape. Looking back at this time, it had its advantages and drawbacks. The music was king, and the options were fewer.
We were stuck with these decisions, which could make the whole process more spontaneous somehow. On the other hand, it was a pain in the ass and many of these decisions made at the time may have been wrong. But did they result in an inferior end result? Technically, maybe. Musically, maybe not.
Even now, with a high-end DAW, I find that old habits stick. I tidy stuff up. I create sub mixes of each essential group of tracks, so once the mix is ready to print, I'm often only dealing with a limited number of sub mixes which I can route out to an analog board and mix 'live,' or automate within the DAW. And I'll still bounce down sections to disk, with processing, and hide the original tracks.
As a kid, I remember buying new LPs and while they were playing, I settled down with the cover, read the liner notes, looked over the lyrics and gazed at the nice 12" x 12" artwork, which hopefully added to the overall experience. I liked gatefold covers even more. In doing this, I became familiar with the song titles, and as I was reading the cover while listening to the LP, the names stuck. Unless they were Zeppelin tracks - their titles never made any sense.
Then came CDs. The artwork was less impressive, the fonts used for lyrics and credits were often impossibly small, particularly in a dark room (some of the CD cover artists should have been beaten with a stick).
WTF. I'm supposed to use a lupe?
Then came downloads. No cover. No nothing. Song titles are replaced with "Track 5." I don't want to stare at an iPod screen. And if you stream audio around the house from a central server, you never get to see the song titles.
I thought it was just my grumpy self, but recently I've noticed that it isn't. More and more, I hear references to songs off an album such as "that verse on track 4 , or is it 5?" or "there's a song on the new XYZ record that you should hear - it's in the middle - maybe track 6."
Thinking about all the songs I have in my iTunes library from the past few years, I probably can actually name less than 5%. I have to hunt for them to figure out what I want to hear, auditioning the first few bars.
Technology isn't always a step forward.
After my favorite dog, Oscar, died last Summer, his companion, Hannah (a pitbull with homicidal tendencies towards other dogs) lapsed into mourning. Dealing with a depressed pitbull isn't fun, so I decided to get her a cat, intended as a pal, playmate and something to elevate her mood. Hannah, strangely, likes cats and doesn't murder them.
So, turning to Craigslist, a kitten was soon adopted, and named Bernie.
Bernie settled in well; a feisty character. Unfortunately, Hannah and Bernie just tolerate each other and soon declared détent. So now I'm saddled with an unemployed cat for the next twenty years. However, Bernie likes the mix room. Moving faders, flashing lights and patchbays and equipment racks are, it turns out, Disneyland for a curious cat.
He particularly enjoys walking on keyboards and control surfaces mid-mix, carefully stepping on 'write' and 'latch' buttons, knocking off USB hubs and torturing dongles like he would a mouse. He climbs acoustic treatment panels, swinging on patch cables to the point that connections are lost. He has also created Logic key commands which I never knew existed, like delete tracks and erase regions right now and, oh, while we're at it, destroy the undo history, all with one well-placed paw.
Damn cat.
© 2010 Sound Propaganda
View Comments