|
From: robs <aq...@ya...> - 2008-02-20 09:23:16
|
--- Wojciech Knapik <web...@wo...> wrote: > As for $rand1 - I experimented with the values you suggested and this > range seems a little too broad to me. I don't know, maybe it was the > sample I used, but it sounded a little ridiculous when i got to more > than +50, or less than -50.. Could sticking with such a small range > cause significant problems (as opposed to using -200:200) ? I chose 200 as it's only one musical tone (i.e. the difference between a C note and a D note), but let me just give it a quick try... Ha, ha, ha! Yes, it does sound a bit ridiculous! 50 should be fine providing you make sure you get a few figures after the decimal point. > As for $rand2 and $rand3 - I read about the allpass effect in the sox > man page and the link provided there, but I honestly don't get any of it > ;] Any suggestions as to these ranges ? What we're doing here is splatting some inaudible phase shifts into the audible part of the signal so say freq between 1000 & 2000 Hz and bandwidth between 1 & 2 octaves should be sufficient. Again, a bunch of figures after the decimal point is needed for this to be effective against removal. > > sox rand-lector.wav 5min-lector.wav pad 10 10 repeat 9 > > > > to get your 5 minute file with lector roughly every 30 secs, though it > would be > > much better to use a fresh rand-lector.wav for each repeat---see the > soxexam > > manual page for hints on how this can be scripted. > > I could just generate multiple rand-lector files and concatenate them > into one 5min-lector - is that what you meant ? Basically, yes. The pad effect places the lector in the middle of a 30 sec window. Of course, concatenation is easy with sox, but if you want to do it in a loop it can be a little inefficient; the scripting example in soxexam shows how to do concatenation in a loop in a slightly more efficient way, but it might not be necessary in your case. > > Mixing will reduce the volume of the track, this can be constrained to just > the > > points during which the lector is active by using compand with sox -m. > > The compander proved to be no less esoteric, than the allpass filter > ;] I tried reading the wikipedia entries for both, but that didn't help > a lot, so once again any hints would be very much appreciated.. Might be more trouble than it's worth, but you could use something along the following lines: play -m input.au 5min-lector.au compand 0.2,.2 6:-70,-60,-20 -8 -90 0.1 > > I would recommend using a low bitrate for the demo songs (e.g. ogg level 1 > or > > 2) just to keep your net bandwidth down. > > I was thinking of streaming mp3 and ogg in two qualities - 128 and > 192 for mp3 and as for ogg, I haven't really decided yet - I have little > experience with this format. These mp3 qualities are roughly equivalent > to ogg level 3 and 4, aren't they ? I believe so; guess you'll have to play this one by ear ;) > > Please donate your SoX scripts back to the SoX project if you get them up > and > > running; they would make an interesting example for others to learn from. > > I'll share my scripts as soon as they're done. Should I send them to > this list ? Yes that would be great. Cheers, Rob __________________________________________________________ Sent from Yahoo! Mail - a smarter inbox http://uk.mail.yahoo.com |