#asahi-dev · 2022-03-08 05:53
marcanit's not our job to make everything magically work on every distro#asahi-dev · 2022-03-08 05:53
marcanthis is all part of "distros have to care"#asahi-dev · 2022-03-08 05:51
marcanas far as the kernel goes, its only job is to provide a multichannel output device with one channel per speaker#asahi-dev · 2022-03-08 05:51
marcanbut we'd want to get such a plug-in upstreamed of course#asahi-dev · 2022-03-08 05:51
marcanchances are with that and existing ALSA infra we can make all this work without too much fuss#asahi-dev · 2022-03-08 05:50
marcanalso this: https://github.com/Artefact2/alsaimpulse#asahi-dev · 2022-03-08 05:48
marcanand UCM2 is used for things like jack plug/unplug and switching routing#asahi-dev · 2022-03-08 05:48
marcanwe can have e.g. card/device profiles that insert a filter to generate the streams for the individual speakers#asahi-dev · 2022-03-08 05:48
marcanlook in /usr/share/alsa for that stuff#asahi-dev · 2022-03-08 05:48
marcando you know about ALSA UCM2 and card profiles?#asahi-dev · 2022-03-08 05:39
marcanideally it should be transparent to PA/JACK/whatever you put in front, though of course nothing stops you from bypassing it and hitting the hardware directly, or writing your own custom profile#asahi-dev · 2022-03-08 05:39
marcanprobably in ALSA, done with profiles in such a way that downstream software just sees it as a stereo device#asahi-dev · 2022-03-08 05:39
marcanI'm thinking we need a way of specifying a matrix of FIR impulses or IIR/biquad chains#asahi-dev · 2022-03-08 05:38
marcanit pretty much has to be userland; EQ in the kernel would never fly#asahi-dev · 2022-03-08 05:22
marcanlong-term on Linux I expect userland will have to deal with this#asahi-dev · 2022-03-08 05:22
marcanwe could do that as a stopgap if the hardware can do it, but it definitely wouldn't be the long-term solution#asahi-dev · 2022-03-08 05:21
marcansince EQ/filtering will be involved#asahi-dev · 2022-03-08 05:21
marcanthe different speakers need different sample streams though#asahi-dev · 2022-03-08 05:21
marcan< kettenis> what I would like to do in OpenBSD is have a single stereo sample stream and be able for the amplifier codecs to pick the right channel out of that stream#asahi-dev · 2022-03-07 15:40
marcan24000.01 certainly sounds like the same reference, it's too good for two xtals :)#asahi-dev · 2022-03-07 15:39
marcanif it's the same clock then it's possible somewhere higher up in the clock tree there's just an inaccurately specified clock#asahi-dev · 2022-03-07 15:39
marcanif it's a different reference clock, that would be down to individual variance#asahi-dev · 2022-03-07 15:38
marcan47997 vs 48000 is 60ppm, which is not unheard of for xtal tolerance#asahi-dev · 2022-03-07 15:37
marcanah#asahi-dev · 2022-03-07 15:37
marcanhow much?#asahi-dev · 2022-03-07 15:37
marcanoh, huh#asahi-dev · 2022-03-07 15:36
marcanafter we worked out the NCO formula#asahi-dev · 2022-03-07 15:36
marcanhm, I thought it was accurate in the end?#asahi-dev · 2022-03-07 15:33
marcanthere's one way of figuring out if the divider is before or after the NCO; deliberately making the fractional counter take a while to converge and measuring how long it takes#asahi-dev · 2022-03-07 15:26
marcanlike div enable#asahi-dev · 2022-03-07 15:26
marcanor maybe it needs bit 8 set and also one of the others?#asahi-dev · 2022-03-07 15:26
marcanI have no idea then :)#asahi-dev · 2022-03-07 15:25
marcanah, could be bits 7:0#asahi-dev · 2022-03-07 15:25
marcanI'm not even sure if that bit is the divider any more, tbh#asahi-dev · 2022-03-07 15:23
marcanis it implemented for the NCO inputs?#asahi-dev · 2022-03-07 15:22
marcanyeah#asahi-dev · 2022-03-07 15:22
marcanon mca0 only!#asahi-dev · 2022-03-07 15:22
marcanmaybe we should just also set the dividers to 2 everywhere?#asahi-dev · 2022-03-07 15:21
marcanalso looks like on t6k only the first one is set weird (and not divided)#asahi-dev · 2022-03-07 15:21
marcan#117: 1068000000 0 0xa4/0x28e0380ac: regval: 0x89100100#asahi-dev · 2022-03-07 15:21
marcan#116: 1068000000 0 0xa4/0x28e0380a8: regval: 0x89100100#asahi-dev · 2022-03-07 15:21
marcan#114: 1068000000 0 0xa4/0x28e0380a0: regval: 0x87100100#asahi-dev · 2022-03-07 15:21
marcan#115: 1068000000 0 0xa4/0x28e0380a4: regval: 0x88100100#asahi-dev · 2022-03-07 15:21
marcan#113: 1068000000 0 0xa4/0x28e03809c: regval: 0x86100100#asahi-dev · 2022-03-07 15:21
marcan#112: 24000000 0 0xa4/0x28e038098: regval: 0x80100000#asahi-dev · 2022-03-07 15:20
marcanare you seeing that?#asahi-dev · 2022-03-07 15:20
marcanon t8103 it would seem like only mca1 is not divided#asahi-dev · 2022-03-07 15:20
marcan#117: 900000000 0 0xa0/0x23b0400e8: regval: 0x89100100#asahi-dev · 2022-03-07 15:20
marcan#116: 900000000 0 0xa0/0x23b0400e4: regval: 0x89100100#asahi-dev · 2022-03-07 15:20
marcan#114: 900000000 0 0xa0/0x23b0400dc: regval: 0x87100100