Process audio frames with MediaFrameReader

This article shows you how to use a MediaFrameReader with MediaCapture to get audio data from a media frame source. To learn about using a MediaFrameReader to get image data, such as from a color, infrared, or depth camera, see Process media frames with MediaFrameReader. That article provides a general overview of the frame reader usage pattern and discusses some additional features of the MediaFrameReader class, such as using MediaFrameSourceGroup to retrieve frames from multiple sources at the same time.


The features discussed in this article are only available starting with Windows 10, version 1803.


There is an Universal Windows app sample that demonstrates using MediaFrameReader to display frames from different frame sources, including color, depth, and infrared camreas. For more information, see Camera frames sample.

Setting up your project

The process for acquiring audio frames is largely the same as acquiring other types of media frames. As with any app that uses MediaCapture, you must declare that your app uses the webcam capability before attempting to access any camera device. If your app will capture from an audio device, you should also declare the microphone device capability.

Add capabilities to the app manifest

  1. In Microsoft Visual Studio, in Solution Explorer, open the designer for the application manifest by double-clicking the package.appxmanifest item.
  2. Select the Capabilities tab.
  3. Check the box for Webcam and the box for Microphone.
  4. For access to the Pictures and Videos library check the boxes for Pictures Library and the box for Videos Library.

Select frame sources and frame source groups

The first step in capturing audio frames is to initialize a MediaFrameSource representing the source of the audio data, such as a microphone or other audio capture device. To do this, you must create a new instance of the MediaCapture object. For this example, the only initialization setting for the MediaCapture is setting the StreamingCaptureMode to indicate that we want to stream audio from the capture device.

After calling MediaCapture.InitializeAsync, you can get the list of accessible media frame sources with the FrameSources property. This example uses a Linq query to select all frame sources where the MediaFrameSourceInfo describing the frame source has a MediaStreamType of Audio, indicating that the media source produces audio data.

If the query returns one or more frame sources, you can check the CurrentFormat property to see if the source supports the audio format you desire - in this example, float audio data. Check the AudioEncodingProperties to make sure the audio encoding you desire is supported by the source.

mediaCapture = new MediaCapture();
MediaCaptureInitializationSettings settings = new MediaCaptureInitializationSettings()
    StreamingCaptureMode = StreamingCaptureMode.Audio,
await mediaCapture.InitializeAsync(settings);

var audioFrameSources = mediaCapture.FrameSources.Where(x => x.Value.Info.MediaStreamType == MediaStreamType.Audio);

if (audioFrameSources.Count() == 0)
    Debug.WriteLine("No audio frame source was found.");

MediaFrameSource frameSource = audioFrameSources.FirstOrDefault().Value;

MediaFrameFormat format = frameSource.CurrentFormat;
if (format.Subtype != MediaEncodingSubtypes.Float)

if (format.AudioEncodingProperties.ChannelCount != 1
    || format.AudioEncodingProperties.SampleRate != 48000)

Create and Start the MediaFrameReader

Get a new instance of MediaFrameReader by calling MediaCapture.CreateFrameReaderAsync, passing the MediaFrameSource object you selected in the previous step. By default, audio frames are obtained in buffered mode, making it less likely that frames will be dropped, although this can still occur if you are not processing audio frames fast enough and fill up the system's alloted memory buffer.

Register a handler for the MediaFrameReader.FrameArrived event, which is raised by the system when a new frame of audio data is available. Call StartAsync to begin the acquisition of audio frames. If the frame reader fails to start, the status value returned from the call will have a value other than Success.

mediaFrameReader = await mediaCapture.CreateFrameReaderAsync(frameSource);

// Optionally set acquisition mode. Buffered is the default mode for audio.
mediaFrameReader.AcquisitionMode = MediaFrameReaderAcquisitionMode.Buffered;

mediaFrameReader.FrameArrived += MediaFrameReader_AudioFrameArrived;

var status = await mediaFrameReader.StartAsync();

if (status != MediaFrameReaderStartStatus.Success)
    Debug.WriteLine("The MediaFrameReader couldn't start.");

In the FrameArrived event handler, call TryAcquireLatestFrame on the MediaFrameReader object passed as the sender to the handler to attempt to retrieve a reference to the latest media frame. Note that this object can be null, so you should always check before using the object. The typs of media frame wrapped in the MediaFrameReference returned from TryAcquireLatestFrame depends on what type of frame source or sources you configured the frame reader to acquire. Since the frame reader in this example was set up to acquire audio frames, it gets the underlying frame using the AudioMediaFrame property.

This ProcessAudioFrame helper method in the example below shows how to get an AudioFrame which provides information such as the timestamp of the frame and whether it is discontinuous from the AudioMediaFrame object. To read or process the audio sample data, you will need to get the AudioBuffer object from the AudioMediaFrame object, create an IMemoryBufferReference, and then call the COM method IMemoryBufferByteAccess::GetBuffer to retrieve the data. See the note below the code listing for more information on accessing native buffers.

The format of the data depends on the frame source. In this example, when selecting a media frame source, we explicitly made certain that the selected frame source used a single channel of float data. The rest of the example code shows how to determine the duration and sample count for the audio data in the frame.

private void MediaFrameReader_AudioFrameArrived(MediaFrameReader sender, MediaFrameArrivedEventArgs args)
    using (MediaFrameReference reference = sender.TryAcquireLatestFrame())
        if (reference != null)
unsafe private void ProcessAudioFrame(AudioMediaFrame audioMediaFrame)

    using (AudioFrame audioFrame = audioMediaFrame.GetAudioFrame())
    using (AudioBuffer buffer = audioFrame.LockBuffer(AudioBufferAccessMode.Read))
    using (IMemoryBufferReference reference = buffer.CreateReference())
        byte* dataInBytes;
        uint capacityInBytes;
        float* dataInFloat;

        ((IMemoryBufferByteAccess)reference).GetBuffer(out dataInBytes, out capacityInBytes);
        // The requested format was float
        dataInFloat = (float*)dataInBytes;

        // Get the number of samples by multiplying the duration by sampling rate: 
        // duration [s] x sampling rate [samples/s] = # samples 

        // Duration can be gotten off the frame reference OR the audioFrame
        TimeSpan duration = audioMediaFrame.FrameReference.Duration;

        // frameDurMs is in milliseconds, while SampleRate is given per second.
        uint frameDurMs = (uint)duration.TotalMilliseconds;
        uint sampleRate = audioMediaFrame.AudioEncodingProperties.SampleRate;
        uint sampleCount = (frameDurMs * sampleRate) / 1000;



In order to do operate on the audio data, you must access a native memory buffer. To do this, you must use the IMemoryBufferByteAccess COM interface by including the code listing below. Operations on the native buffer must be performed in a method that uses the unsafe keyword. You also need to check the box to allow unsafe code in the Build tab of the Project -> Properties dialog.

unsafe interface IMemoryBufferByteAccess
    void GetBuffer(out byte* buffer, out uint capacity);

Additional information on using MediaFrameReader with audio data

You can retrieve the AudioDeviceController associated with the audio frame source by accessing the MediaFrameSource.Controller property. This object can be used to get or set the stream properties of the capture device or to control the capture level. The following example mutes the audio device so that frames continue to be acquired by the frame reader, but all samples have value of 0.

audioDeviceController.Muted = true;

You can use an AudioFrame object to pass audio data captured by a media frame source into an AudioGraph. Pass the frame into the AddFrame method of an AudioFrameInputNode. For more information on using audio graphs to capture, process, and mix audio signals, see Audio graphs.