Synthesis and Sound Processing
Making new sounds using a computer, that is, sounds that are not based on recorded material, is called synthesis.
In a computer, sound is represented as a long series of numbers, often 44,100 numbers/per second (see text on sound in computers). To make sounds using a computer, one may use a word processing program and write the numbers:
365, 7931, 3291, 68, 8392, 192, 3819, 9231, 6421, 13613, ...
... but no one actually wants or needs to do this. Instead one uses automatic methods or mathematic formulas that allow the computer to generate the numbers for us.
Such a method must be able to generate sounds that are interesting, but not so complicated as to require an extremely long time for the computer to calculate. There are many different synthesis methods, and some of these may be tested using the programs in DSP. The best way to learn about synthesis is to make sounds oneself. Click on the help texts in the programs!
The different synthesis methods are suited for different purposes.
For example:
FM synthesis is best for making metallic clock sounds,
subtractive synthesis is best for mimicing the human voice, and
physical modelling is best for recreating guitar sounds.
One may make an entire musical piece using only a synthesis method, but the most interesting and fun approach is perhaps to use several different methods and combine the diverse sounds.
Since a sound is represented as a long series of numbers in a computer, we can make automatic methods that treat these numerical series in different ways. Thus we may change, twist and turn sounds that have been recorded using a microphone, for example.
If we add numbers from two sounds, the resulting sound is a mix between the two:
When multiplying all of the sample values with a number, we may make the sound stronger or weaker.
With more advanced sound processing methods we can add echo or resonance to a sound, and extend or alter the pitch. We may also break it up and turn it into something completely new and unrecognizable.
|
