| 描述： | Noise created by individual oscillators, using audio paint, with different height images, to demonstrate what happens with too few oscillators vs. plenty. In the end, the result is not random enough to be noise. The first 2second burst is pure white noise for comparison. Then we have multiple 1second bursts from audiopaint in a sequence of different image sizes: 50,100,150,200,250,350, 500, 700, 1000, 1350, 1750, 2200, 2700, 3250, 3850, 4500, and 9999 (this corresponds to the number of oscillators). The last burst is longer, and there is 1/2 second gap of silence after the first (reference) burst and before that last (9999) burst. The images other dimension was 20. The spacing of frequencies was exponential, between 40 Hz and 18 KHz. This is not intended to be useful, just an illustration during a discussion in a forum (http://www.freesound.org/forum/samplerequests/35199/?page2post75605). As mentioned there, I realized only afterward that exponential spacing would be giving me an approximation to pink noise instead of white noise, so the reference burst at the start is not really a fair comparison. Ideally, I would go back and redo all this using linear spacing, but that's a lot of trouble. :) I did, however, change to linear to get a white approximation, but that's a different sound I'll upload separately (c.f. http://www.freesound.org/people/zimbot/sounds/242053/). I don't believe you can get true white noise without at least something being random in your synthesis method. |