Hi Bob, can you post the command you used to generate that file? It is indeed not "silent" at the beginning: >sox sound77.wav -n trim 0 9 stats DC offset -0.000002 Min level -0.007812 Max level 0.007812 Pk lev dB -42.14 RMS lev dB -48.17 RMS Pk dB -47.83 RMS Tr dB -48.64 Crest factor 2.00 Flat factor 2.17 Pk count 99.0k Bit-depth 1/8 Num samples 397k Length s 9.000 Scale max 1.000000 Window s 0.050 However, if I do a pad: >sox sound77.wav test.wav pad 10 and analyze the result, it's silent: > sox test.wav -n trim 0 9 stats DC offset 0.000000 Min level 0.000000 Max level 0.000000 Pk lev dB -inf RMS lev dB -inf RMS Pk dB -inf RMS Tr dB -inf Crest factor 1.00 Flat factor 111.97 Pk count 794k Bit-depth 0/0 Num samples 397k Length s 9.000 Scale max 1.000000 Window s 0.050 Something else seems to be wrong. Cheers, Erich On 23.09.2016 23:43, Bob S wrote: > >> On Sep 22, 2016, at 11:56 PM, Erich Eckner wrote: >> >> Hi Bob, >> >> I think the problem is related to "dither". Two suggestions: >> try disabling dither ("-D") in all but the last sox command and >> try to minimize the number of "snippets" you combine, e.g.: put >> non-overlapping sounds in one file in the first step. This way you will >> end up with less dither-noise. >> >> cheers, >> Erich > > Thanks Erich, > > I did figure the problem was just too many overlapping tracks piling up, even if they were supposedly silent. I wonder if something else is wrong, though — even in the partial sub-tracks, where a single effect has just been padded to the right length with ‘pad’, the padded part is not silent at all but actually a pretty loud static hiss for something that is supposed to be silence. I tried adding -D, but that did not change anything. Does it go before or after the input file in the sox call? Or after the output file? I’ve tried all three with the same results. > > For example, this file is a small sound effect padded with ‘silence’ for the first ten seconds: > https://dl.dropboxusercontent.com/u/4215526/sound77.wav > > I’ve got to be doing something wrong to get that kind of noise, aren’t I? > > Anyway, yesterday I rewrote the whole thing to assemble nonoverlapping tracks, like you say. I had not done that originally because I figured that over time, a big error would be introduced near the end as sounds did not line up with their correct video frame. But it seems to be OK and now I am only mixing two tracks together, so the problem is not so bad. The noise stays at the level you hear in the sample above. But it’s still pretty annoying and I’d like to get rid of it. > > Bob > ------------------------------------------------------------------------------ > _______________________________________________ > Sox-users mailing list > Sox-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/sox-users >