check-alsabat.sh: add a new option '-N' to specify the channel number… #1143
Conversation
… for playback Before the channel number of playback was hard coded to 2. since sof playback PCMs only support 2 channels. But we now we have Bluetooth offload with nocodec topology which support both 1 channel and 2 channels. So replace this hard coding to optional. Signed-off-by: Keqiao Zhang <[email protected]>
keqiaozhang
force-pushed
the
alsabat_channel_optional
branch
from
f886a03 to
cb81613
Compare
speaker has 2 channels and DMIC 4. If you want to run an alsa-bat test using speaker+dmic the number of channels MUST be different. that said I am not sure which nocodec topology this test is using, to the best of my knowledge the SSP MUST use the same number of channels in both directions. Some slots may not be used and contain zero data, but that's a separate issue. The SSP requires that both RX and TX use the same bclock, fsync, wordsize and number of channels/slots. If we are testing with two SSPs then one of the two needs to be slave. this could be fun at multiple levels. What are we trying to test really? |
I don't think we ever use this test "over the air", that would be too unreliable in very noisy lab. Even if we did or someone else does in a very quiet environment: why wouldn't using just 2 channels be enough? Or even just one in that case.
Good question; the test description does not say what it's meant for, what it supports versus not (it's also using alsabat in undocumented way, minor doc fix submitted in #1148) So far we've been using stereo USB adapters with headset jack loopback. This PR adds the following comment: |
|
Before playback channel number was hard coded to 2, but now we have a new request for Bluetooth offload tplg, it has 3 cases:
Case 1 & 2 only support mono playback/capture, that's why I replaced the hard coded channel number with parameter forms. |
|
@keqiaozhang my question is why patch below and a single option is not enough, at least for now. It would be 100% compatible with the wiki page you shared. --- a/test-case/check-alsabat.sh
+++ b/test-case/check-alsabat.sh
@@ -33,7 +33,7 @@ source "$(dirname "${BASH_SOURCE[0]}")"/../case-lib/lib.sh
OPT_NAME['p']='pcm_p' OPT_DESC['p']='pcm for playback. Example: hw:0,0'
OPT_HAS_ARG['p']=1 OPT_VAL['p']=''
-OPT_NAME['C']='channel_c' OPT_DESC['C']='channel number for capture.'
+OPT_NAME['C']='channel_c' OPT_DESC['C']='number of channels.'
OPT_HAS_ARG['C']=1 OPT_VAL['C']='1'
OPT_NAME['r']='rate' OPT_DESC['r']='sample rate'
@@ -110,9 +110,8 @@ aplay -Dplug$pcm_p -d 1 /dev/zero -q || die "Failed to play on PCM: $pcm_p"
arecord -Dplug$pcm_c -d 1 /dev/null -q || die "Failed to capture on PCM: $pcm_c"
# alsabat test
-# hardcode the channel number of playback to 2, as sof doesnot support mono wav.
-dlogc "alsabat -P$pcm_p --standalone -n $frames -r $rate -c 2 -f $format -F $frequency -k $sigmak"
-alsabat -P$pcm_p --standalone -n $frames -c 2 -r $rate -f $format -F $frequency -k $sigmak & playPID=$!
+dlogc "alsabat -P$pcm_p --standalone -n $frames -r $rate -c $channel -f $format -F $frequency -k $sigmak"
+alsabat -P$pcm_p --standalone -n $frames -c $channel -r $rate -f $format -F $frequency -k $sigmak & playPID=$!
# playback may have low latency, add one second delay to aviod recording zero at beginning.
sleep 1Even if we need a way to configure a different number of channels in the future, there should be a simple way to configure both channels the same. Keep it simple for now. |
Besides of nocodec platforms, not all our devices have stereo USB sound card connected. For most of the CAVS platforms, the USB sound card only support mono capture, that why we cannot use one option to configure both channels the same. |
|
SOFCI TEST |
| alsabat -P$pcm_p --standalone -n $frames -c 2 -r $rate -f $format -F $frequency -k $sigmak & playPID=$! | ||
| # BT offload PCMs also support mono playback. | ||
| dlogc "alsabat -P$pcm_p --standalone -n $frames -r $rate -c $channel_p -f $format -F $frequency -k $sigmak" | ||
| alsabat -P$pcm_p --standalone -n $frames -c $channel_p -r $rate -f $format -F $frequency -k $sigmak & playPID=$! |
There was a problem hiding this comment.
the change sounds fine but I can't figure out how this test works with separate commands for playback and capture. I don't have enough understanding to review this.
There was a problem hiding this comment.
We have some configuration files in framework side to specify PCM IDs and channel numbers, e.g:https://github.com/intel-innersource/drivers.audio.ci.sof-framework/blob/main/parameters/alsabat-playback.csv.
There was a problem hiding this comment.
I can't figure out how this test works with separate commands for playback and capture
Maybe because ALSA man pages are utterly out of date? I struggled with that too, see yesterday's 84dcefd
"man speaker-test" is 11 years old... speaker-test -h looks fine.
Why not use -c 1 everywhere in that case? What's the point of playing 2 channels and capturing only 1? |
Hope I understand your question correctly, our SOF doesn't support 1 channel playback except for Bluetooth offload. So we cannot set 1 channel for both playback and USB capture.
Because at that time we only have one model of USB sound card to use, and it only support mono capture. In order to enable audio quality testing, we have no other choice. I will merge this PR to move the blocker for Bluetooth test. |
… for playback
Before the channel number of playback was hard coded to 2. since sof playback PCMs only support 2 channels. But we now we have Bluetooth offload with nocodec topology which support both 1 channel and 2 channels. So replace this hard coding to optional.