如何使用选定的输出设备上的本机API在Windows上播放声音（mp3 / WAV）

Question

I simply want to be able to create options for my program, so that user can pick which output device will be used to play sounds, like this one in MS Lync:

I originally created my program in Qt and I asked similar (but not identical) question here Qt5+ How to set default audio device for QMediaPlayer

I figured out that Qt is too much bugged for this and this is impossible, so I lowered my requirements and I will use native windows API as these are probably only solution here. This unfortunately requires rewrite of some parts of my program, and now I am following this guide on msdn: https://msdn.microsoft.com/en-us/library/windows/desktop/dd371455%28v=vs.85%29.aspx

I basically want to be able to do following:

• List all available output devices and display them on preferences form - I already have a working code for that using IMMDeviceEnumerator
• Let user pick a device they want to use for output of my program - I already have that part
• Create a function, let's call it PlaySound(string path) that if called with path of .wav or .mp3 file would use the preferred IMMDevice and play a file through it - this is what I need help with

Because I was using Qt so far and I have pretty much no idea of MS windows internals, I have no idea how could one take a file stored somewhere on disk and play it using windows API's especially using that selected IMMDevice which user set in their preferences. I was googling and searching through documentation, but I could only work extremely complex and weird solutions, such as https://msdn.microsoft.com/en-us/library/windows/desktop/dd316756%28v=vs.85%29.aspx

I could even find some examples where you can play mp3 file using MCI device, but that didn't really explain how to alter preferred output device, so it isn't very useful for my use.

I understand that low-level API is probably not going to offer some simple "playmyfile" function, but it would be nice to have at least some example of super-simple solution or some tutorial that would play media files using selected output device on windows so that I could use that as a starting reference. I have a working active IMMDevice , now I just need to make it possible to play mp3 / wav files through it.

NOTE: This is not some generic "how to play a sound on windows" question. I need to be able to play that sound on selected audio output device . For my program only (just like MS Lync, VLC media player or any other advanced audio program can). I don't want to change system global preferences (default device etc).

Answer 1

I managed to do that but surprisingly using windows native libraries called "DirectShow" which are primarily designed for video rendering, but can handle audio as well.

How to:

Enumerate output devices This functions iterates over all audio devices detected by OS and store them in a list.

void Options::Initialize()
{
#ifdef WIN
    HRESULT hr;
    ICreateDevEnum *pSysDevEnum = NULL;
    hr = CoCreateInstance(CLSID_SystemDeviceEnum, NULL, CLSCTX_INPROC_SERVER, IID_ICreateDevEnum, (void **)&pSysDevEnum);
    if (FAILED(hr))
        return;

    IEnumMoniker *pEnumCat = NULL;
    hr = pSysDevEnum->CreateClassEnumerator(CLSID_AudioRendererCategory, &pEnumCat, 0);
    if (hr == S_OK)
    {
        // Enumerate the monikers.
        IMoniker *pMoniker = NULL;
        ULONG cFetched;
        while (pEnumCat->Next(1, &pMoniker, &cFetched) == S_OK)
        {
            IPropertyBag *pPropBag;
            hr = pMoniker->BindToStorage(0, 0, IID_IPropertyBag, (void **)&pPropBag);
            if (SUCCEEDED(hr))
            {
                // To retrieve the filter's friendly name, do the following:
                VARIANT varName;
                VariantInit(&varName);
                hr = pPropBag->Read(L"FriendlyName", &varName, 0);
                if (SUCCEEDED(hr))
                {
                    OutputDevice device;
                    device.Name = QString((QChar*)varName.bstrVal, wcslen(varName.bstrVal));
                    Options::devices.append(device);
                }
                VariantClear(&varName);
                pPropBag->Release();
            }
            pMoniker->Release();
        }
        pEnumCat->Release();
    }
    pSysDevEnum->Release();
#endif
}

Create a filter for device that user selected Iterate over all devices once more and make a filter for that which was selected by user

HRESULT hr;
ICreateDevEnum *pSysDevEnum = NULL;
hr = CoCreateInstance(CLSID_SystemDeviceEnum, NULL, CLSCTX_INPROC_SERVER, IID_ICreateDevEnum, (void **)&pSysDevEnum);
if (FAILED(hr))
{
    Error("Failed SystemDeviceEnum");
    return;
}

IEnumMoniker *pEnumCat = NULL;
QSettings s;
hr = pSysDevEnum->CreateClassEnumerator(CLSID_AudioRendererCategory, &pEnumCat, 0);
IBaseFilter *pFilter = NULL;
if (hr == S_OK)
{
    // Enumerate the monikers.
    IMoniker *pMoniker = NULL;
    ULONG cFetched;
    int i = 0;
    while (pEnumCat->Next(1, &pMoniker, &cFetched) == S_OK)
    {
        IPropertyBag *pPropBag;
        hr = pMoniker->BindToStorage(0, 0, IID_IPropertyBag, (void **)&pPropBag);
        if (SUCCEEDED(hr))
        {
            // retrieve the filter's friendly name now
            VARIANT varName;
            VariantInit(&varName);
            hr = pPropBag->Read(L"FriendlyName", &varName, 0);
            if (SUCCEEDED(hr))
            {
                QString name = QString((QChar*)varName.bstrVal, wcslen(varName.bstrVal));
                if (s.value("d:" + name).toBool())
                {
                    hr = pMoniker->BindToObject(NULL, NULL, IID_IBaseFilter, (void**)&pFilter);
                    // now we got the filter in pFilter so we can play sound using that filter
                    PlayWin(pFilter, path);
                }
            }
            VariantClear(&varName);
            pPropBag->Release();
        }
        pMoniker->Release();
    }
    pEnumCat->Release();
}
pSysDevEnum->Release();

Play the sound using the filter for our device In this function device is pFilter from previous function

HRESULT hr = CoCreateInstance(CLSID_FilterGraph, NULL, CLSCTX_INPROC_SERVER, __uuidof(IGraphBuilder), (void **)&x->pGraph);
if (FAILED(hr))
{
    Error("ERROR - Could not create the Filter Graph Manager.");
    return;
}

hr = x->pGraph->QueryInterface(IID_IBasicAudio, (void**)&x->pOutput);

if (FAILED(hr))
{
    Error("ERROR - Could not create the IBasicAudio.");
    return;
}

x->pFlx = device;
if (device)
    x->pGraph->AddFilter(device, L"fd");
hr = x->pGraph->QueryInterface(__uuidof(IMediaControl), (void **)&x->pControl);
hr = x->pGraph->QueryInterface(__uuidof(IMediaEvent), (void **)&x->pEvent);

// Build the graph.
hr = x->pGraph->RenderFile(path, NULL);
if (SUCCEEDED(hr))
{
    // Run the graph.
    hr = x->pControl->Run();
}
else
{
    Error("Unable to play: " + QString::fromWCharArray(path));
}

This code on itself is of course not going to compile out of box, but it gives you a clue how to do this, in nutshell:

1. Retrieve list of all devices and store it somewhere, so that we can create dialog for user
2. Before we play a sound, we check which device user selected and create a filter for it
3. We apply the filter to DirectShow BasicAudio which is itself able to play any media file supported by system codecs.

Documentation on msdn: https://msdn.microsoft.com/en-us/library/windows/desktop/dd407292%28v=vs.85%29.aspx