A friend of mine asked me to write a little about sound—in other words, about how to improve the sound of your track. As I was writing, other questions came up too. I am answering them in the form of an FAQ...
This assumes that anyone following my advice has decent monitor speakers, or at least something close to them, as well as basic sound-engineering knowledge. And one more thing—don't expect advice like, “Press this button in that plugin, then this one, and everything will sound fucking great.” It won't. I am explaining the underlying principles.
How do you get good sound?
Use high-quality hosts, samples, and instruments. Remember that even if a sample is 24-bit, 96 kilohertz, that does not mean it is high-quality. Look at its frequency response and check for dips; check the dynamic range. In time, you will be able to determine this simply by ear.
If you are recording live instruments, use a good microphone, preamp, and audio interface. Record at the highest possible level that does not cause overload. An input limiter will protect you from level spikes.
Listen to the track carefully a day after writing it. Write down what you don't like about it. Fix things point by point, from the most difficult to the simplest. Make sure that subsequent changes do not ruin the improvements you have already made. Once it seems that the track sounds good, try listening to it on ordinary speakers from a hi-fi system or on cheap computer speakers. Compare it with commercial tracks in a similar genre. Do not work continuously for more than two hours. Ideally, switch the sound off once an hour and go to the kitchen for a smoke, literally for five minutes. If you don't do this, your ears will inevitably become fatigued and you will only ruin what you have started.
Launch Wavelab and turn on its 3D analysis for your track. Then do the same with any commercial track—The Prodigy or Portishead, for example; the genre does not matter. In fact, I can already tell you what you will see: overload at 60–500 Hz and a dip at 2,000–20,000 Hz. In any commercial track, everything is generally smoothed out and laid out evenly across the frequency spectrum and in terms of level. It is very difficult to achieve this at home, but it is possible. To do so, work more with the equalizer, and try to cut frequencies rather than boost them (it is better to cut something and turn it up); try to distribute instruments evenly across the frequency and panorama grid; and do not forget to compress the rhythm section if necessary.
Do not get carried away with quantity. Remember: good music is not music to which there is nothing left to add, but music from which there is nothing left to remove. This applies to production too. If you have doubts about whether a particular sound or effect is needed, rework it, replace it, or remove it.
Do not try to make the sound crystal-clear. If it is dirty to begin with, you will not be able to clean it up without losses unless you have specialized knowledge and equipment. If you really want to, here are a couple of tips: find out which track is dirtying the sound and try removing it, replacing it, or blending it in differently. If that does not help, then something is wrong with the mix.
Do not try to make the sound extremely loud. Do not overuse compressors, especially multiband ones—they mangle the sound very seriously.
An ideal mix is one that does not require mastering at all. Use a spectrum analyzer frequently, and also render the track and examine its progressive level, trying to keep it even throughout the entire track. The more even the track is, the less mastering—and consequently, destructive editing—it will need.
Alternate your sound checks between speakers and headphones. It is better to have two pairs of headphones: DJ headphones, such as Pioneer or the Sony MDR series, and monitoring headphones, such as the AKG K141/240. It is also a good idea to listen on a large sound system. After that, it immediately becomes clear what is what, although be careful: powerful speakers can camouflage certain flaws, such as underdeveloped high frequencies.
And finally, more naturalness! Even when you are writing some wild glitch drums, imagine that a live drummer is playing them, and set the volume, panning, and so on accordingly. The same goes for bass, keyboards, and other instruments. Do not be lazy about changing repeating parts a little; nonlinearity is almost always better.
Which compressor is best to use for mastering?
A multiband compressor, meaning one that provides compression within a selected frequency band. This is a fairly strong destructive processing method, and it MUST NOT be used as a creative effect. If the initial and final sound differ significantly, remix the original track. I do not recommend using mastering plugins such as Waves C4, Mackie Final mix, or iZotope Ozone. They ruin the sound very badly. This becomes noticeable over time. It is better to use PSP Vintage Warmer, PSP MasterComp, Waves L3, or DiscoDSP Trillme.
Where is it best to do mastering? Directly in the sequencer?
It is best done in Steinberg Wavelab. Unlike all other editors, it can switch on several effects simultaneously without applying them. This is convenient for comparing and auditioning different sections. It also has decent tools for analyzing audio files and a good MP3 encoder.
How should drums be equalized and processed in general?
Any way you like :). By ear. There can be no universal remedy here. Remove unused frequency ranges and blend loops with one another. I recommend equalizing after the bassline and main synthesizers. Sometimes it makes sense to use thickening—in other words, adding extra individual drums on the corresponding beats; most often, this means a kick. And don't forget to tune the drums!
Don't forget to use compressors and saturators. To achieve greater density, you can put distortion, overdrive, or a decimator on top and blend it into the main signal just a little—very quietly.
Programs: TC Native Compressor/Deesser and Sonic Destructor, PSP Vintage Warmer, iZotope Trash.
How do you make a reese?
As a basis, take two identical melodic waves rich in harmonics—in other words, a saw, square, spectral waves, and so on—and detune them by 25–50 cents. It makes sense to add a sine wave one octave below at zero detuning. After that, it is entirely up to your creativity. You can add other waves, or the same ones with different detuning; create a unison; modulate the pulse frequency and the amount of detuning; modulate the phase; place frequency and modulation filters on top and modulate those too (preferably with an LFO); process it with a flanger, phaser, or chorus; distort it, including with multiband distortion (different distortion on different frequencies); compress it very heavily; spread it across the stereo field; add reverb with compression on top; mix it with other reeses; create quiet wave underlayers at +3, +5, and/or +7; and so on. Then resample it into a separate file, load it into a sampler, and use it.
Programs: as the main synthesizer: Linplug Albino 3, NI Absynth, NI Reaktor, NI Massive, Korg MS-20. For processing—anything you like. I recommend sitting down with Reaktor and writing your own synthesizer with all the necessary bells and whistles. That is exactly what I did, for example; it is not as difficult as it seems.
Which microphone should you choose—a dynamic or a condenser? And which brand?
A dynamic microphone does not require phantom power, has a denser sound, is undemanding, and is suitable both for the studio and the stage. A condenser is more sensitive and requires phantom power, but provides higher recording quality. Keep in mind that it will most likely require a preamp. Do not be swayed by fashionable tubes; digital preamps with a compressor are no worse at all. As for brands: for dynamics, get a Shure SM58—you cannot go wrong. Among condensers, the OKTAVA 319 can be recommended; it is inexpensive and high-quality. For microphone processing, use something from the dbx line of equipment, depending on the specific task.
Author: Cyberworm.
