Spectral Sieve

The spectral sieve allows you to "sift" out the weak partials. The results one achieves from using this program are dependent on the properties of the original material, and may be totally unpredictable. The program is also useful in removing background noise.

How does Spectral Sieve work?
The machine makes a spectral analysis of a sound approximately 40 times a second. Each analysis removes the weak partials and the sound is then rebuilt from the sifted spectrum. This is a large calculation so it takes quite a long time to use the spectral sieve.

Values that may be set:
The graph sets values that control how many partials are to be kept. If the value is high, many of the partials are maintained and the sound is not changed significantly. If the value is low, you keep only the strongest partials and what you hear may sound completely different than the original.

Example:
This trickling sound has been sieved.
This noise was the original sound before the "sifting."

Removal of noise
As you can see, the program can also be used to remove unwanted noise. If you reduce the number of partials a little bit, some of the noise will disappear. Be prepared to experiment a little bit to find the right number of partials.