|Principles of Sound Design (Part 2)
Written by Simon Cann - © 2007, Coombe Hill Publishing. Reprinted with Permission.
This article is an excerpt from the following book: How To Make A Noise - A Comprehensive Guide to Synthesizer Programming.
This article is also a continuation of Principles of Sound Design (Part 1)
Knowing the Limitations
I have made many references to real instrument sounds. This has been done because most people have an idea of a “piano sound” or a “brass sound”. However, I would not recommend using any of the techniques in this book if your goal is to capture the nuances of a real instrument.
If you are after authenticity, then I recommend you hire a studio (with an good room and an engineer), hire the instrument you want to record, and find the best session musician you can get. This is most likely to give you satisfactory results. As a second option, find a quality sample library and use samples of the real instrument—again, find a session musician as they are likely to give you the best performance. That being said, don’t be shy of using synthesizers—just remember, they are not the real thing (unless your goal is synthetic sounds).
Polyphony Limit and Note Priority
Any machine you use to create music is going to have limitations. This is true whether you are using hardware or software. There will always be a finite number of notes you can play. With hardware, this limit is usually a known number (often hardware will be described as being, say, 64 note polyphonic). With software, the system limitations are usually affected by a range of factors, such as how much system resource is used by other functions (such as the host, the effects in the synthesizer, the number of filters you are using etc).
With hardware synthesizers the limitations on voices may lead to voice stealing—when you reach the maximum polyphony, the synthesizer will cut off a previous note as a new note is triggered. With software, when the system resources are reached then your system is likely to drop the audio, freeze, or fall over. However, there are many things you can do with software to reduce the load of your CPU.
Most hosts will allow you to freeze tracks and free up resources. This may be the easiest way to prevent system overload.
With software you have the option to limit polyphony. By limiting polyphony you will have to address voice stealing rather than look at complete system failure (which is probably the less preferable option). As the voice stealing can then be performed on a per patch basis, you have the option to control the voice allocation so that key parts are not robbed.
All of the featured synthesizers allow you to set the polyphony limits on a per patch basis. With Wusikstation you can even limit voices on a per layer basis and so ensure that the more important elements of your sound are not compromised. Cameleon 5000 also allows you to limit the number of partials used, however, this may have an effect on the tone.
Simplify the Patch
With software there are many things you can do to simplify a patch and reduce its CPU load, for instance:
* Reduce the number of oscillators. Is that sixth oscillator really adding to the sound or is it just there for good luck? Equally, as we mentioned in Chapter 4: Sound Sources, it is possible to use one wave where it creates the sound of two waves in combination.
* Reduce the filters. Either reduce the number of filters used and/or reduce the slope of the filters (for instance in Z3TA+ the 36 dB filters use more system resource than the 12 dB filters).
* Cut the release time of your envelopes. A shorter release time will mean that the total number of notes still sounding at any one time may be reduced. This can be a very effective
way to reduce total system load without losing polyphony especially with very busy playing. However, it may make your patch sound awful.
* Turn off the FX units. As noted in Chapter 10: FX it can be a more efficient use of system resources to use external FX units.
Principles of Sound Design
Unfortunately there is no magic formula to creating a patch. However, I would suggest a few basic steps before you jump in:
* Understand what each knob does.
* Understand the signal flow in the synthesizer and how each element interacts.
* Have a definite aim in what you’re trying to program—don’t just fiddle with the knobs and hope something useful comes along.
Once you have these basics sorted, I suggest a two stage approach:
* Get something that is “alright”. In other words, a sound that is functional and is reasonably close to where you want to end up—you should be able to get to this point quickly.
* Undertake the detailed tweaking. This is the time to make sure the patch is exactly right for your needs—this will be the time consuming part of the programming.
When it comes to getting something that is “nearly” right, I see no reason not to use a preset, although the subsequent editing may take longer as you will have to familiarize yourself with the workings of the patch. But then again if you’re using one of the synthesizers with a more straightforward architecture or you know what you’re doing, you may be happier starting from a clean sheet.
So you know how your synthesizer works, you have a grasp of what all the knobs do, how do you get the basic sound together. For me, I then tend to think in terms of the main food groups.
Main Food Groups
Most sounds can be characterized by two key elements—their brightness and the envelope (primarily the attack of the envelope, but often you need the decay and sustain level to be right for the contrast of the attack to be clear). Once these two elements have been addressed, any tonal nuances after that may be affected by a range of factors.
If you are trying to categorise the brightness of a sound, it will generally fall within one of four descriptions: very bright, bright, dull, or very dull. You can try to subdivide further, but realistically, it is already quite difficult to try to find the line between these four categories.
If you try hard, you can find four categories of attack time:
* fast—percussive (for instance a xylophone hit or a guitar string being plucked)
* fast—but not percussive, for instance an organ
* medium, and
Again, you can try to subdivide the categories further, but it is a fairly pointless exercise.
Getting the Combination Right
Once you understand what you want in the way of brightness and envelope, getting the right sound is dependent on the right elements being drawn together. For instance, here are some combinations may want to look at.
To Get Brightness
Consider the oscillator (or oscillators) and filter combination. If they don’t work together, you will either get a shrill sound, or a thin sound, or a dull sound, but never a bright sound. For instance, if you are creating a sound based on a square wave, you may find it preferable to choose a 12 dB filter. A 24 dB or 36 dB filter may work, but to my ear they ultimately rob the square wave of much of its character, which you may want to keep if you are trying to create a bright patch.
To Get Richness
One of the key elements to a rich sound is its movement. You can achieve this with careful selection of your sound source (which will usually involve several slightly detuned oscillators and perhaps some delicate use of a sub-oscillator). Alternatively, you can just slap on loads of chorus. My preference is for the former option.
To Get Attack
Most obviously, it is important that the envelope opens quickly to get some attack in a note. Depending on the tonal quality you are after, it is often important that the note then decays quickly to the sustain level—this will give you a percussive envelope reminiscent of a plucked guitar string
However, to ensure that the attack is emphasized, another significant factor is the waveform. Some waves can sound slower than others. For instance, all other factors being equal, a square wave can sound like it has a faster attack than a sawtooth wave. In general bright sounds (especially
distorted bright sounds) are perceived as having more attack.
To Get Warmth
A warm sound is subtly different from a rich sound, often it is much less bright. Warmth is also a comparative term—if a colder sound starts a patch and is then washed over by a warm sound, the contrast can give the listener the illusion of warmth.
Any warm sound will probably have a slow attack and a slow release. There must be thickness to the sound which will often come from detuned oscillators. As noted above, one of the key elements of a warm sound is the changing tone which can be achieved with another oscillator with an even slower envelope.
To Get Punch
It is hard to get some punch into a sound without adding too much brightness. First you need a fat sound source, then you need a fast attack and finally you need a hefty filter. The filter can be modulated with the envelope so that the attack is emphasized but the brightness is controlled.
Choosing the Right Tools
When you are programming take some time to make sure you choose the right tool. It is easy to make a few glib comments about which is the right tool for any job. For instance if you want a fat “analog” sounding pad, then perhaps your first choice should be Vanguard—that will likely give you the fastest and the best results.
However, Vanguard is not the only tool you can choose—Rhino makes excellent FM sounds, but in the hands of a skilled sound designer it is equally capable of making great analog sounding patches. Which is right for you is a matter of your taste, your skill, and what is right for the particular track you are working on. However, don’t let your (or my) preconceptions about any machine’s strength let you rule out a synthesizer from any task.
More importantly, choosing the right tool is more a function of having the knowledge to control that instrument. The only way to learn how to use an instrument is to keep using it over a long time—don’t keep buying new instruments, pick one and learn every aspect of its character.
Finally, any patch has to be highly playable. The playability of the patch is one aspect that separates the average sound from the professional. Use your knowledge of how the sound will be deployed to set the modulation sources.
For instance, if the sound is going to be played by a live musician, then you might ensure that velocity has an effect and that the modulation wheel and expression pedal both offer a wide range of control. However, if the MIDI track controlling the sound is going to be programmed, then you might want to make sure that external controllers, such as track envelopes may have a significant effect, so for instance, you may add an additional filter which has little effect in the normal course but can be readily controlled by a track envelope. For more information, see How To Make A Noise - A Comprehensive Guide to Synthesizer Programming.
To stay informed about new articles, be sure to click here to sign up for the DigiFreq Music Technology Newsletter. It's free!