Thank you for your patch submission. I haven't yet given it a try. Looking at the changes you made though, I would guess it attempts to make the rounding of the audio conversion less audible?
Yes.
I noticed there is a buffer which is allocated of size 48000, which I assume is related to the sampling rate. Do you think this will also work fine with 44100 for example?
rand_table not buffer, this contain precalculated dither value (we can calculate this in realtime but it cost some speed). Size of rand_table should be big enough to prevent audible low freq modulation. 44100 / 48000 = 0,91875 Hz 48000 / 48000 = 1 Hz 96000 / 48000 = 2 Hz 192000 / 48000 = 4 Hz
I'm no DSP expert, so if someone has any insight/pointers to information on this topic I'd be interested.
http://en.wikipedia.org/wiki/Dither http://audio.rightmark.org/lukin/dither/dither.htm _______________________________________________ fluid-dev mailing list fluid-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/fluid-dev
