Additive synthesis… what systems for doing it are out there?
Well, i’m hoping that some people who know more than i do about this will
read this, and tell me what they think.
My main interest is in sound design. i want to synthesize sounds that human
ears haven’t heard before, and which are pleasing and interesting to the
One possible method to achieve this goal would be to simply start
experimenting with the various methods of audio synthesis, trying different
things, observing the results and learning from these observations, and then
repeating the process ad infinitum.
Another possible method, which i think is a better way to begin, is to study
and try to understand already existing sounds, using analysis tools like
SPEAR. In this way, a framework of knowledge about already existing sounds
can be built up, which will help me make more informed decisions about what
to experiment with and how, as described in the previous paragraph, and thus
produce better results sooner.
i’m familiar with the various methods of sound synthesis, and with the
various common waveforms in synthesis and their sinusoidal makeup (triangle
is odd harmonics of the fundamental with a certain decrease in amplitude
with increasing frequency, etc etc). i know what goes into and comes out of
a fourier transform, although i haven’t taken the time to learn the math
behind it (and i probably won’t unless i find myself needing this knowledge
for some some specific reason.) Since additive synthesis is the most
versatile and powerful method of audio synthesis in that it can synthesize
any waveform, this is the model that i want to ground my study and
understanding in. I recognize the other synthesis methods, the common
waveforms, and the various signal processing tools (audo compressors,
reverb, etc.) as useful shortcuts and shorthand notations to use when
thinking and working in terms of the sum of sine waves.
i know that additive synthesis models tend to be large (many sine waves).
however, even in sounds that are composed of large numbers of sine waves,
the behavior of these sine waves may often be governed by simple and easily
understandable sets of rules. Does anyone know if there is any literature
out there about this, to save me trying to reinvent the wheel if someone has
already done it?
Also, does anyone know what computer programs or hardware systems etc. are
out there that are geared toward performing additive synthesis? At the
most basic level, such a program would consist of a large number of sine
oscillators, with time-varying control over the frequency and amplitude of
each oscillator. There are probably many other useful features that could
be included, which would make implementing the simple rules mentioned in the
previous paragraph easier. It would also be good for such a program to
incorporate fourier analysis and resynthesis, so that it could serve as an
all in one analyzer and synthesizer, which would make incorporating data
from an analysis, modifying it, and incorporating part or all of it into a
new sound a streamlined, fast and easy process. Such a tool would be great
for studying sounds, experimenting and learning. It could even incorporate
curve fitting (to smooth out data and to help make sense of the data) and
pattern matching (find component basic waveforms (triangle, square etc), and
identify instances of known simplifying rules).
As far as I know, SPEAR is the closest thing to this out there, although
it’s still pretty far from it. Does anyone know if there is something like
what i described out there? I’ve been trying to build a patcher in Max to
take analysis data (SDIF) from SPEAR and allow the user to view and
manipulate it in two key ways that aren’t allowed in SPEAR:
1) allow users to view the amplitude vs. time for each sine wave in a 2D
2) allow users to quickly and easily (with one or two mouse clicks) hold one
or both of the time varying properties of any partial constant (frequency or
amplitude). this will allow the user to hear how each time varying property
contributes to the overall sound, which is really the very heart of learning
additive synthesis, i think
But it’s easier said than done to build something like this. I’m still
working on it, but it would be great to find a tool which can already do
There must be people who work professionally with additive synthesis… does
anyone know what software/hardware and techniques they use?
So, any thoughts or ideas? My mind’s completely open, and I’m of course
open to constructive criticism… if there’s something that i’m missing i’d
rather know than stumble in the dark.
Thanks, any help or info would be greatly appreciated
Quote: erik.bowen wrote on Sun, 04 February 2007 13:48
> Also, does anyone know what computer programs or hardware systems etc. are
> out there that are geared toward performing additive synthesis?
Certainly, check out SPEAR. Also, there might be some interesting things in the CNMAT Spectral Tutorials:
Is there any demo version of Diphone available? If not, do you know what the most recent version is, and where i can find a detailed description of its functionality? (the ircam page doesn’t say much about either)
The most recent version is 4.1. If you google "Diphone, ircam" you’ll
find some stuff I think.
You should have a look at Kyma X:
Additive, aggregate, real-time resynthesis, editors, etc.
It’s a soft front-end (Kyma X) with a DSP farm (Capyara 320).
It might be just what you’re looking for.
You can get good deals on secondhand systems.