You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We want to play and record directly at the hw level, not with any conversions done by alsa-lib.
This is wrong and needs to be fixed, it will hide issues or mask conversions.
@keqiaozhang I think the plughw in alsabat playback is required, because the USB card cannot support all formats and sample rates, but do you think we can remove plughw from alsabat capture?
@plbossart , I agree with you, adding plughw is not a wise choice. The reason we use plughw is that the USB sound card only support 1 channel capture and for nocodec mode, we don't use USB sound card for capture, we use the passthrough pipeline directly and it supports 2 channels capture. To unify the channel count, we decided to use this workaround.
I believe there're still have some other reasons, but I can't record right now. @xiulipan , please help to add here if you still remember.
I will try to refine the test script for alsabat test next.
@plbossart The plughw here is used because we do not use pulseaudio and we do not know what fmt/channel/rate is supported for our loopback (playback/captur) PCM. So remove plughw is not a option for alsabat test.
I have submitted a PR #401 to remove plughw for alsabat test. we can use tplgreader to get the fmt/channel and rate for playback PCM, but for USB sound card(used to capture the loopback data), it only support S16_LE/1 channel, so I just specified format and channel for USB sound card.
Activity
aiChaoSONG commentedon Sep 2, 2020
@keqiaozhang I think the plughw in alsabat playback is required, because the USB card cannot support all formats and sample rates, but do you think we can remove plughw from alsabat capture?
plbossart commentedon Sep 2, 2020
sorry I don't get the point @aiChaoSONG.
What I meant was that all access to the SOF device should be with hw:
keqiaozhang commentedon Sep 3, 2020
@plbossart , I agree with you, adding plughw is not a wise choice. The reason we use plughw is that the USB sound card only support 1 channel capture and for nocodec mode, we don't use USB sound card for capture, we use the passthrough pipeline directly and it supports 2 channels capture. To unify the channel count, we decided to use this workaround.
I believe there're still have some other reasons, but I can't record right now.
@xiulipan , please help to add here if you still remember.
I will try to refine the test script for alsabat test next.
xiulipan commentedon Sep 24, 2020
@plbossart The plughw here is used because we do not use pulseaudio and we do not know what fmt/channel/rate is supported for our loopback (playback/captur) PCM. So remove plughw is not a option for alsabat test.
keqiaozhang commentedon Sep 24, 2020
I have submitted a PR #401 to remove plughw for alsabat test. we can use tplgreader to get the fmt/channel and rate for playback PCM, but for USB sound card(used to capture the loopback data), it only support S16_LE/1 channel, so I just specified format and channel for USB sound card.
xiulipan commentedon Sep 24, 2020
@keqiaozhang We also have headset capture test, you can only removed plug from one side.
xiulipan commentedon Jan 26, 2021
merged #554 @keqiaozhang @plbossart can we close this issue.
keqiaozhang commentedon Mar 29, 2021
PR#554 is merged, this issue can be closed.