On Tue, Feb 11, 2020 at 07:15:00PM GMT, Alexandre Ratchov wrote: > On Tue, Feb 11, 2020 at 07:01:28PM +0100, Florian Obser wrote: > > I've been running the base diffs since you posted them. Firefox, > > chrome and mpv still make noise :) > > > > I'm puzzled by this: > > > > $ cat /etc/mixerctl.conf > > > > outputs.master=255,255 > > record.enable=off > > > > $ mixerctl outputs.master > > > > outputs.master=255,255 > > > > $ sndioctl > > output.level=127 > > > > I don't understand how they relate and why one goes to 255 and the > > other to 127. > > The error reporting is confusing, too: > > > > $ sndioctl output.level=128 > > > > integer overflow > > > > But no regressions to report :) > > > > Thanks, the code is base on MIDI bits, which uses the 0..127 range; > sndiod, aucat and many codecs also use the 0..127 range. Anyway, > replaced the error message by: > > $ sndioctl output.level=128 > 128: expected integer in the 0..127 range > > [...] > > I'm wondering if persents or floating points in the [0:1] range would > be less confusing and solve most "units" problems. >
Hi Alexandre, I have to say that I also find the two ranges mildly confusing, i.e. 0-255 in one place, and 0-127 in another. In terms of units, personally, I'm used to, and quite like, the granularity of 0-255. Again, not my place so others will certainly be more help here. One more point regarding the interface, though. This is the way mixerctl(1) currently behaves: $ mixerctl outputs.master outputs.master=255,255 $ mixerctl outputs.master=100 outputs.master: 255,255 -> 100,100 $ mixerctl outputs.master=300 outputs.master: 100,100 -> 255,255 Should sndioctl(1) behave the same way? Cheers, Raf