How to prevent audio to be sent to the default audio endpoint?

Wong, Irene 0

Hi,

I've some audio that I would like to only be sent to a particular audio endpoint (in this case, a controller output). I basically use IMMDeviceEnumerator::EnumAudioEndpoints() to get the endpoint I want and send audio to the device buffer using IAudioRenderClient::GetBuffer(). That works as I expect. But when I have another audio device (e.g. a headset) set as the default audio endpoint, the audio intended only for the controller is being output to both the controller speaker as well as the headset. Is there any way to prevent audio from going to the default audio endpoint?

To add to the issue, my 2 devices have different sample rate. So the automatic routing to both endpoints do not work as one of the endpoint will sound incorrect due to the sample rate mismatch.

Thanks.

Irene

Tong Xu - MSFT 2,116 Reputation points Microsoft Vendor

2024-05-27T04:58:32.6466667+00:00

Hi, @Wong, Irene
Have you checked my answer? What else i can do?
Tong Xu - MSFT 2,116 Reputation points Microsoft Vendor

2024-05-31T07:09:53.45+00:00

Hello, @Wong, Irene
Would you please give me some feedback on my answer?
Wong, Irene 0 Reputation points

2024-05-31T21:30:17.6433333+00:00

Sorry about the delayed response. I haven't enabled email notification so I missed your earlier comments.

Just to clarify that I've retrieved the particular device that I want (not using the device returned from IMMDeviceEnumerator::GetDefaultAudioEndpoint method). i.e. I am only sending audio data to the corresponding device buffer and I'm not sending audio to the default audio endpoint buffer. While the audio is correctly sent to the intended endpoint, but the audio is also being routed automatically to the Default Audio Device which I don't want.

Tong Xu - MSFT 2,116 Microsoft Vendor

But you also set another audio device (headphones) as the default audio endpoint. That's the problem. You should use a loop to obtain multiple endpoint devices:

HRESULT get_specific_device(LPCWSTR szLongName, IMMDevice **ppMMDevice) {
    HRESULT hr = S_OK;
 
    *ppMMDevice = NULL;
    
    // get an enumerator
    IMMDeviceEnumerator *pMMDeviceEnumerator;
 
    hr = CoCreateInstance(
        __uuidof(MMDeviceEnumerator), NULL, CLSCTX_ALL, 
        __uuidof(IMMDeviceEnumerator),
        (void**)&pMMDeviceEnumerator
    );
    if (FAILED(hr)) {
        
printf(L"CoCreateInstance(IMMDeviceEnumerator) failed: hr = 0x%08x", hr);
        return hr;
    }
    ReleaseOnExit releaseMMDeviceEnumerator(pMMDeviceEnumerator);
 
    IMMDeviceCollection *pMMDeviceCollection;
 
    // get all the active render endpoints
    hr = pMMDeviceEnumerator->EnumAudioEndpoints(
        eRender, DEVICE_STATE_ACTIVE, &pMMDeviceCollection
    );
    if (FAILED(hr)) {
        printf(L"IMMDeviceEnumerator::EnumAudioEndpoints failed: hr = 0x%08x", hr);
        return hr;
    }
    ReleaseOnExit releaseMMDeviceCollection(pMMDeviceCollection);
 
    UINT count;
    hr = pMMDeviceCollection->GetCount(&count);
    if (FAILED(hr)) {
        
    printf(L"IMMDeviceCollection::GetCount failed: hr = 0x%08x", hr);
        return hr;
    }
 
    for (UINT i = 0; i < count; i++) {
        IMMDevice *pMMDevice;
 
        // get the "n"th device
        hr = pMMDeviceCollection->Item(i, &pMMDevice);
        if (FAILED(hr)) {
            
        printf(L"IMMDeviceCollection::Item failed: hr = 0x%08x", hr);
            return hr;
        }
        ReleaseOnExit releaseMMDevice(pMMDevice);
 
        // open the property store on that device
        IPropertyStore *pPropertyStore;
        hr = pMMDevice->OpenPropertyStore(STGM_READ, &pPropertyStore);
        if (FAILED(hr)) {
            
        printf(L"IMMDevice::OpenPropertyStore failed: hr = 0x%08x", hr);
            return hr;
        }
        ReleaseOnExit releasePropertyStore(pPropertyStore);
 
        // get the long name property
        PROPVARIANT pv; PropVariantInit(&pv);
        hr = pPropertyStore->GetValue(PKEY_Device_FriendlyName, &pv);
        if (FAILED(hr)) {
            printf(L"IPropertyStore::GetValue failed: hr = 0x%08x", hr);
            return hr;
        }
        PropVariantClearOnExit clearPv(&pv);
 
        if (VT_LPWSTR != pv.vt) {
            printf(L"PKEY_Device_FriendlyName variant type is %u - expected VT_LPWSTR", pv.vt);
            return E_UNEXPECTED;
        }
 
        // is it a match?
        if (0 == _wcsicmp(pv.pwszVal, szLongName)) {
            // did we already find it?
            if (NULL == *ppMMDevice) {
                *ppMMDevice = pMMDevice;
                pMMDevice->AddRef();
            } else {
                printf(L"Found (at least) two devices named %ls", szLongName);
                return E_UNEXPECTED;
            }
        }
    }
    
    if (NULL == *ppMMDevice) {
        printf(L"Could not find a device named %ls", szLongName);
        return HRESULT_FROM_WIN32(ERROR_NOT_FOUND);
    }
 
    return S_OK;
}

About different sample rate, the format is not necessarily a format that the audio endpoint device supports. You may use mix format.
44100HZ 16bit stereo: 44100 samples per second, sample data recorded in 16-bit (2 bytes), stereo (stereo);22050HZ 8bit mono: 22050 samples per second, sample data recorded with 8 bits (1 byte), mono;

Wong, Irene 0 Reputation points

2024-06-03T20:19:55.1233333+00:00

I have obtained multiple audio endpoints devices, one for controller speaker and the other is the headphones which is the default audio endpoint.

However, the audio data that I've sent to the controller speaker via IAudioRenderClient::GetBuffer() is audible on both controller speaker and the headphones.

Note that I'm only sending zeros to the headphones' buffer at the time. So I do not expect to hear anything from the headphones. But that's not the case.
Wong, Irene 0 Reputation points

2024-06-03T20:24:48.2866667+00:00

I have obtained multiple endpoint devices, one for controller speaker and the other for headphones (which is the default endpoint).

However, the audio data that I've sent to the controller speaker buffer via IAudioRenderClient::GetBuffer() is audible on both the controller speaker and the headphones. Note that I'm sending zeros to the headphones' buffer at the time, so I expect to hear nothing from the headphones. But that's not the case.
Wong, Irene 0 Reputation points

2024-06-04T18:21:08.2266667+00:00
What I meant by sending zeros is filling the data packet pointed by ppData with 0’s. Filling the data packet with 0’s is just a test here to make listening easier for the test. Normally, I would be filling ppData with valid audio data.

Here is what I’m trying to achieve:

Play sound A only on headphones (Default Device)

Play sound B only on controller speaker

However, what I get is:

Sound A is playing on headphones

Sound B is playing on both headphones and controller speaker. But I’m filling the data packet of the controller speaker with Sound B data only.
Tong Xu - MSFT 2,116 Reputation points Microsoft Vendor

2024-06-05T02:57:44.2666667+00:00

Well. Sending zeros to ppData means writing silence. Right now, your ultimate requirement is capturing multiple audio routes to multiple devices. It has deviated from the original question, for more in-depth questions, please accept the current answer and open another thread.**
**
Taking a step back, capturing audio from a specific application and routing to a specific audio device is not possible with the standard user mode API.You need to either hook APIs or create virtual devices to accept application streams/sessions.
https://stackoverflow.com/questions/31623824/how-to-capture-audio-from-specific-application-and-route-to-specific-audio-devic?noredirect=1&lq=1
Wong, Irene 0 Reputation points

2024-06-05T17:17:12.8766667+00:00

I'm unsure what led you to thinking I'm capturing audio. I'm rendering all the sounds in my app.
Deleted

This comment has been deleted due to a violation of our Code of Conduct. The comment was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.
Tong Xu - MSFT 2,116 Reputation points Microsoft Vendor

2024-06-06T02:40:34.1466667+00:00
Also, I bring news.

Here is what I’m trying to achieve:

Play sound A only on headphones (Default Device)

Play sound B only on controller speaker

I don't think you can make it in codec.
The default endpoint has the highest level. *For example, if a user set the internal speaker as the default endpoint at any time in the past and later attached new USB headphones, the system would never move to the USB headphones without a manual selection.
*
https://learn.microsoft.com/en-us/windows-hardware/drivers/audio/default-audio-endpoint-selection#the-level-property-in-windows-10
Tong Xu - MSFT 2,116 Reputation points Microsoft Vendor

2024-06-07T06:55:26.2966667+00:00

Any other ideas shared with us? Please let me know.

3 answers

Deleted

This answer has been deleted due to a violation of our Code of Conduct. The answer was manually reported or identified through automated detection before action was taken. Please refer to our Code of Conduct for more information.

Comments have been turned off. Learn more
Tong Xu - MSFT 2,116 Reputation points Microsoft Vendor

2024-06-04T02:44:37.2066667+00:00

Sending zeros to the headphones' buffer is not advisable! It turns into autofill or makes the buffer pointer zero. Null pointer is violation. You were not send zero to ppData so here is autofill the buffer.

so I expect to hear nothing from the headphones.

You can try use IAudioClient::Stop method to stop recording the audio stream.
Or delete the audio device in driver

Because Microsoft does not provide the method to remove audio endpoints dynamically.
Please sign in to rate this answer.

0 comments No comments
Sign in to comment
Asma khalida 0 Reputation points

2024-06-04T02:49:25.3633333+00:00

To prevent audio from being sent to the default audio endpoint on a Windows system, you can redirect it to a different audio device or mute it using the following steps:

Change the Default Playback Device:

Right-click the sound icon in the system tray and select "Open Sound settings."

Under "Output," choose a different audio device from the drop-down menu.

Use Software Settings:

Many applications allow you to select a specific audio output device in their settings. Configure the application to use a non-default audio device. For more details if anyone is interested in to know about Integrity Garage door repairs huntington visit here
Please sign in to rate this answer.

0 comments No comments
Sign in to comment

Share via

How to prevent audio to be sent to the default audio endpoint?

3 answers