1233532 - Unable to play captured audio if output sample rate is higher than 48kHz

Status: NEW

(Core :: WebRTC: Audio/Video, defect, P4)

Description

[KOLANICH]

User Agent: Mozilla/5.0 (Windows NT 6.3; rv:42.0) Gecko/20100101 Firefox/42.0
Build ID: 20151029151421

Steps to reproduce:

https://mozilla.github.io/webrtc-landing/gum_test.html
Push audio.



Actual results:

Nothing. It doesn't captures.

My OS is Windows 8.1
My sound card is Audigy 2 ZS
I tried to change capture format of the mic (bit depth and sampling rate), it haven't helped.


Expected results:

>Audio will feedback unless you use a headset!

Comment 1

[Randell Jesup [:jesup] (needinfo me)]

What drivers do you have for the Audigy 2 ZS?  (Paul, do you have one of these?)

Can you provide logs (see https://wiki.mozilla.org/Media/WebRTC/Logging) with
NSPR_LOG_MODULES=mediamanager:4,getusermedia:4,webrtc_trace:65535 
and 
WEBRTC_TRACE_FILE=nspr

Thanks

Comment 2

[KOLANICH]

Hope it is enough
http://pastebin.com/pWn8vWFU
http://pastebin.com/ZS4cLzVq

Comment 3

[Paul Adenot (:padenot)]

I don't. However, we see things like that in the log:

0[d54a680]: ERROR     ; (**:**:**:*** |    0)        VOICE:    1    99;      5960; SetExternalRecordingStatus() invalid sample rate (error=8005)

Comment 4

[Randell Jesup [:jesup] (needinfo me)]

Note: that's here:
    if ((16000 != samplingFreqHz) && (32000 != samplingFreqHz) &&
        (48000 != samplingFreqHz) && (44100 != samplingFreqHz))
    {
         shared_->SetLastError(VE_INVALID_ARGUMENT, kTraceError,
             "SetExternalRecordingStatus() invalid sample rate");

Comment 5

[Randell Jesup [:jesup] (needinfo me)]

0[c8cda30]: STREAM    ; (**:**:**:*** |    0)        VOICE:    1    99;      4328; ExternalPlayoutData(speechData10ms=0x*******, lengthSamples=960, samplingFreqHz=96000, current_delay_ms=40)

That's why.  The question is why we were using that; I suspect the mic is set to that; see see that and go "nope!"

Paul - this may give you an easy way to check (likely this is os-dependent, too).

Comment 6

[Paul Adenot (:padenot)]

We should should insert a resampler there for sure. This is Windows/WASAPI and it cannot resample for you most of the time in the current setup, and it appears that the webrtc.org code does not handle automatic resampling.

Comment 7

[Randell Jesup [:jesup] (needinfo me)]

Paul - is this common, or rare? Is this due to user-settings (non-default) for the mic?  (really rare to need/want to capture at 96K except *maybe* in a studio)

If we fix it, the best way would be to down-sample before feeding it to the webrtc.org core code.

Kolanich - is the 96K default, or did you set that? Does it work if you switch to 48000?

Comment 8

[KOLANICH]

>KOLANICH - is the 96K default, or did you set that?
I did.

>Does it work if you switch to 48000?
No

>nSamplesPerSec : 16000
>if ((16000 != samplingFreqHz)

The first log ( http://pastebin.com/pWn8vWFU ) was for 16 kHz.

Comment 9

[KOLANICH]

This was playback issue, not capture. When I set output settings to 48000 kHz, I started to hear the mic.
BTW, mic now is set to 96kHz.

>We should should insert a resampler there for sure.
In the case of playback I think the resampler is not a good solution.

Comment 10

[KOLANICH]

*48 kHz

Comment 11

[Randell Jesup [:jesup] (needinfo me)]

Hmmm.  Even in your first example, I see samplingFreqHz=96000 (input failure, line 146) and a failure on invalid rates, but maybe it's exposing multiple interfaces at different rates?

Comment 12

[KOLANICH]

Maybe. There exists a WhatUHear interface allowing a user to record the sounds produced by OS/apps/etc. Its sample rate must be equal to sample rate of output. But I'd selected the mic as a source.

Comment 13

[Paul Adenot (:padenot)]

(In reply to KOLANICH from comment #9)
> This was playback issue, not capture. When I set output settings to 48000
> kHz, I started to hear the mic.
> BTW, mic now is set to 96kHz.
> 
> >We should should insert a resampler there for sure.
> In the case of playback I think the resampler is not a good solution.

We choose the native resampling of the audio interface for playback. I was talking about input audio.

Comment 14

[Bulk Bug Changes for mreavy's org]

Mass change P3->P4 to align with new Mozilla triage process.