On Thu, Dec 31, 2009 at 02:13:52AM +0300, Andrej Elizarov wrote: > Trying to get fully working my shiny OpenBSD 4.6-stable. > > Now i had some problem with recording. > I'm just playing around about two days and still no love. > Can't enumerate all variants that i tried. > > Try to record with: > #aucat -o file.raw > or > #aucat -e s16le -h raw -o file.raw >
could you kill aucat and then try the following: Does ``aucat -o file.raw'' produces a file with a plausible size, around 4 * 44100 * recorded seconds ? If the file size is ok, does record.errors (exposed by audioctl) stay zero (or a small integer) ? If so, does the file contain noise or just zeros? What is the result of: hexdump file.raw |head If it contains noise, then it's a mixer or wiring problem. If so, could you set to ``off'' all ``inputs.xxx.mute'' controls, set to 255 all inputs.xxx, and check that your microphone actually works? (you should hear what you're saying on the mic). If it works, try one by one all possible values of record.source until ``aucat -o file.raw'' produces a file with sound. If it works, does running ``aucat -l'' changes anything? > While playing > #aucat -e s16le -h raw -i file.raw > has no sound at all, while recording or playing. > If in testing time i had in mplayer new Scrubs episodes (or whatever) i > hear that video volume is slightly decreases while "aucat -o file.raw". > you mean volume decreases during ``aucat -i file.raw'' (not ``aucat -o file.raw''), right ? > > Changing inputs.mic.source does no effect. > All the time i have aucat -l in background, don't know can this affect > results or not. It's not supposed to change anything, but it's better to minimize the data path when searching for a bug. -- Alexandre