Can WASAPI Renderer play mono audio on both channels?

Question

Can WASAPI Renderer play mono audio on both channels?

Jim Fiocca 1

I am upgrading a synthesizer application (using Visual C++ command line tools) from Wave to WASAPI for the lower latency. I'm trying to use the same WAVEFORMATEX settings but WASAPI appears to be more particular. It's basically 16-bit samples @44.1 kHz mono. I can't get shared mode to accept this setting at all. I got exclusive mode to work, but it only comes out one side. Wave played dual-mono automatically. Can I achieve this with WASAPI?

Answer 1

Jeanine Zhang-MSFT 7,036 Microsoft Vendor

@Jim Fiocca

The client calls the methods in the IAudioRenderClient interface to write rendering data to an endpoint buffer. For a shared-mode stream, the client shares the endpoint buffer with the audio engine. For an exclusive-mode stream, the client shares the endpoint buffer with the audio device.

As far as I'm concerned WASAPI Renderer can play mono audio on both channels.

I suggest you could try to use IAudioClient::IsFormatSupported method indicates whether the audio endpoint device supports a particular stream format.

Jim Fiocca 1 Reputation point

2022-09-13T00:15:43.3+00:00
JeanineZhang, thank you for your reply. IAudioClient::IsFormatSupported method returns AUDCLNT_E_UNSUPPORTED_FORMAT for the mono stream, yet IAudioClient::Initialize() succeeds and audio plays, but only on the left channel. That's for exclusive mode.

In shared mode, here's the difference between what IAudioClient::GetMixFormat() returns and what I am requesting:

GetMixFormat(): my requested, default: wFormatTag = 0x0001, 0xfffe nChannels = 1, 2 nSamplesPerSec = 44100, 44100 nAvgBytesPerSec = 88200, 352800 nBlockAlign = 2, 8 wBitsPerSample = 16, 32 cbSize = 0, 22

Also, the 'closest match' that IsFormatSupported() returns in shared mode is always the same as what GetMixFormat() returns. So it's really not a negotiation, right? I have to use the audio engine's format, which means sign-extending the 16-bit samples to 32 bits and duplicating the mono samples to the right channel. Is that correct?
Thanks!
Jeanine Zhang-MSFT 7,036 Reputation points Microsoft Vendor

2022-09-13T05:57:25.03+00:00

@Jim Fiocca

According to the Doc: Device Formats

If an application is using a device in exclusive mode, the application must use a stream format that the audio hardware explicitly supports.

An application that uses WASAPI to manage shared-mode streams can rely on the audio engine to perform only limited format conversions. The format for an application stream typically must have the same number of channels and the same sample rate as the stream format used by the device.

In my opinion，you should try to convert convert 1 channel to 2 channel by copying one channel into two.
Jeanine Zhang-MSFT 7,036 Reputation points Microsoft Vendor

2022-09-16T07:11:50.117+00:00

@Jim Fiocca

May I know if you have got any chance to check my answer? I am glad to help if you have any other question.
Jim Fiocca 1 Reputation point

2022-09-16T15:00:08.33+00:00

@Jeanine Zhang-MSFT
Thanks for your offer to continue to help.
For exclusive mode, it is working well except for the inefficient buffer copies.
For shared mode, pAudioClient->GetDevicePeriod() returns default period = 10.16 ms, minimum period = 3.00 ms; but then; no matter what value I use for hnsBufferDuration in pAudioClient->Initialize(), pAudioClient->GetBufferSize() returns a value that corresponds to 22.31 ms! And then subsequent calls to renderClient->GetBuffer() what that big of a buffer fail with AUDCLNT_E_BUFFER_TOO_LARGE.
Jeanine Zhang-MSFT 7,036 Reputation points Microsoft Vendor

2022-09-19T08:15:49.12+00:00

@Jim Fiocca

According to the Doc: IAudioRenderClient::GetBuffer method (audioclient.h)

The available space is simply the buffer size minus the amount of data in the buffer that is already queued up to be played. If the caller specifies a NumFramesRequested value that exceeds the available space in the buffer, the call fails and returns error code AUDCLNT_E_BUFFER_TOO_LARGE.

NumFramesRequested: The number of audio frames in the data packet that the caller plans to write to the requested space in the buffer.
Jim Fiocca 1 Reputation point

2022-09-20T00:23:25.473+00:00

@Jeanine Zhang-MSFT
In exclusive mode, pAudioClient->GetBufferSize() returns exactly what I ask for in corresponding hnsRequestedDuration in the pAudioClient->Initialize() call, unless it has to go through block alignment, then I get something reasonably close.
In shared mode, using the same hnsRequestedDuration in the Initialize() call, GetBufferSize() returns something HUGE! Is that how it is supposed to work, or am I doing something wrong? Am I supposed to use a subset of the giant buffer size in increments of my desired bufferSize, using pAudioClient->GetCurrentPadding() calls to determine how much space is left in the buffer?
Jeanine Zhang-MSFT 7,036 Reputation points Microsoft Vendor

2022-09-20T05:53:44.483+00:00

@Jim Fiocca

GetBufferSize() returns something HUGE!

The GetBufferSize method retrieves the size (maximum capacity) of the endpoint buffer.

AUDCLNT_E_BUFFER_TOO_LARGE: The NumFramesRequested value exceeds the available buffer space (buffer size minus padding size).

As far as I'm concerned, you should check whether the caller specifies a NumFramesRequested value exceeds the available space in the buffer.

Can WASAPI Renderer play mono audio on both channels?

1 answer

Activity