Skip to content

[BUG][CML]Arecord0,4 rt715 captured audio files output pop noise  #1969

Closed

Description

@sinahuang

Describe the bug
Arecord0,4 rt715 captured audio files have overshoot noise output at the beginning.

Captured audio files:
audio file.zip

To Reproduce
1.Boot up system successfully.
2. Arecord -Dhw:0,4 -fs16_le -c2 -r 48000 -vv -i 00.wav
3. Check captured audio files 00.wav

Reproduce rate
5/5

Expected result
Arecord0,4 rt715 should captured audio files normally, no noise.

Actual result
Arecord0,4 rt715 captured audio files have pop noise output at beginning.

Test recipe

Kernel:https://github.com/thesofproject/linux/commits/topic/sof-dev commit:3bdd89d
FW: https://github.com/thesofproject/sof/commits/master commit: thesofproject/sof@1e39cee
Tplg: Same with FW branch, sof-cml-rt711-rt1308-rt715.tplg
Platform:CML-H Laptop with codec ALC711, 2 x RT1308, RT715 in SDW mode.

3

2

dmesg1.log no error info in dmesg.

Activity

  1. added
    bugSomething isn't working
    CMLApplies to Comet Lake platform
    on Apr 1, 2020
  2. changed the title [-][BUG][CML]Arecord0,4 rt715 captured audio files have "po" noise output.[/-] [+][BUG][CML]Arecord0,4 rt715 captured audio files output "po" noise [/+] on Apr 1, 2020
  3. plbossart commented on Apr 1, 2020

    @plbossart
    Member

    what does "po" mean?

  4. changed the title [-][BUG][CML]Arecord0,4 rt715 captured audio files output "po" noise [/-] [+][BUG][CML]Arecord0,4 rt715 captured audio files output pop noise [/+] on Apr 2, 2020
  5. added
    P2Critical bugs or normal features
    on Apr 2, 2020
  6. plbossart commented on May 27, 2020

    @plbossart
    Member

    @sinahuang isn't this fixed now? can we close?

  7. sinahuang commented on May 29, 2020

    @sinahuang
    Author

    Issue still can be reproduced with sof-dev(commit: 0d22f66) +master(commit :caae629) on CML-H.

  8. plbossart commented on Jun 9, 2020

    @plbossart
    Member

    Indeed this doesn't look very good, see what I capture on my side.

    rt715

    @singalsu can you take a look? Maybe borrow @kv2019i 's SoundWire laptop to see what happens on the capture side?

  9. singalsu commented on Jun 10, 2020

    @singalsu
    Collaborator

    @plbossart Yup, it looks strange. I'll get such device and test.

  10. singalsu commented on Jun 15, 2020

    @singalsu
    Collaborator

    I have a suitable device now. The volume capture ramp does some really strange things, good that I can now replicate this:

    Screenshot from 2020-06-15 19-58-14

    (the different capture level on L & R does not matter, this was just a quick acoustical test)

  11. self-assigned this
    on Jun 18, 2020
  12. singalsu commented on Jul 2, 2020

    @singalsu
    Collaborator

    This was on hold while I debugged ASRC. I repeated this with current SOF git development repos:

    FW 7b532d3
    kernel f6be379

    The firmware PGA5.0 gain is sane. I traced volume volume channels 0 and 1 gains in every copy(), at 1 ms rate. The 400 ms ramp as defined in topology is similar for both channels without gain jumps.

    Screenshot from 2020-07-02 15-36-40

    Still the captured level envolope shows strange gain changes for both channels at 50 ms, 80ms, and for channel 0 only at 300 ms. The FW ramp envelope can be seen to last for both channels 400 ms that matches with previous plot.

    Screenshot from 2020-07-02 15-35-45

    The capture pipeline contains a 40 Hz high-pass IIR filter as it should. The low frequency down and up and down waveform is from DC signal that is emitted by the codec. I'll do a capture test with pipeline without IIR so it becomes better visisible.

  13. singalsu commented on Jul 2, 2020

    @singalsu
    Collaborator

    The capture with IIR removed from pipeline looks like this:

    Screenshot from 2020-07-02 15-59-44

    The same gain changes can be seen as above but they "ride" above the huge DC level variation in the first 500 ms of the capture. Note: This picture is with less gain than the previous with IIR.

    @mengdonglin @bardliao Do you happen to know contacts to codec driver developer? The control sequence that is applied to codec looks strange. It's not possible to mitigate any further the capture start artifacts with firmware than this.

  14. 24 remaining items

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Labels

CMLApplies to Comet Lake platformP2Critical bugs or normal featuresbugSomething isn't workingcodecCodec HW or driver restriction

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions