On Thu, Aug 13, 2009 at 10:52:28AM +0200, Jan Stary wrote: > First off, thanks to alex@ for all the work being put into aucat. > OpenBSD 4.5 is the first operating system where I can easily do > multitrack recording in base. > > My situation is this: aucat -l is running, and I play two different > copies of an album in two mplayers: a ripped vinyl and a CD reissue, > pausing one and the other, listening to the difference. > > In the moment I paused the first mplayer and launched the second > one, I noticed that the signal is much weaker. First, I took that > to be one of the differences in the two versions, but then, when I > decided which one sounded better and quit the other mplayer, the > signal got stronger again. > > Does aucat somehow 'divide' the signal strengths when playing multiple > inputs (even when some are paused)? Similarly to sox(1) mixing files? >
yes, by default aucat divides the dynamic range (ie the strenght) by the number of streams, which means that when a new stream is added (even if it plays silence) dynamic ranges of other streams are reduced. Similarly when a stream is terminated, its dynamic range is given to other streams. > Trying the same with more than two clients (such as, five paused mplayers > and one playing, then quiting the paused ones; doesn't need to be > mplayer, happens with any other client) seems to confirm this, > but I didn't find anything about it in aucat(1). > > What is the relation of 'aucat -v' to this? > the -v option can be used to ``reserve'' a part of the dynamic range for future streams, so that dynamic ranges of existing streams don't have to be reduced when new streams are added. reducing the volume by 18 units doubles the number of streams that fit in the output dynamic range, but adds ~6dB attenuation per stream, example: -v 127 -> 1 stream (default) -v 109 -> 2 streams -v 91 -> 4 streams -v 73 -> 8 streams ... -- Alexandre
