What’s Coming to Windows Phone in 2011?

With more that 13,000 applications in their Marketplace Microsoft came out on stage today to make some new announcements about what’s coming to Windows Phone 7. We’ll see a major upgrade on all Windows Phone devices by the end of the year and developers are getting access to a lot of new APIs enabling new application scenarios.

Here is a summary list of the new features.

  • Multitasking – in addition to faster application switching multitasking will allow applications to continue processing in the background.
  • Live Tile Functionality Enhancements
  • Sensor Library Enhancements – You’ll have enhanced access to the sensor library and
      Access to the camera and compass-! Reality augmentation is possible!
    • Sockets – This needs no explanation
    • Database
    • IE9 – with hardware acceleration
    • Silverlight+XNA – you can use Silverlight and XNA in the same application
    • Twitter in the People Hub
    • Background Transfers
    • Profile
    • Silverlight 4 Runtime

    Several new countries are being added to the Marketplace. This brings up the total count from 17 to 35. The new countries are listed below with *.

    • Australia
    • Austria
    • Belgium
    • Brazil*
    • Canada
    • Chile*
    • Columbia*
    • Czech Republic*
    • Denmark*
    • Finland*
    • France
    • Germany
    • Greece*
    • Hong Kong
    • Hungary*
    • India*
    • Ireland
    • Italy
    • Japan*
    • Mexico
    • Netherlands*
    • New Zealand
    • Norway*
    • Poland*
    • Portugal*
    • Russia*
    • Singapore
    • South Africa*
    • South Korea*
    • Spain
    • Sweden*
    • Switzerland
    • Taiwan*
    • UK
    • USA /

Streaming from the Microphone to Isolated Storage

Last week I posted a sample voice recorder on CodeProject. The application would buffer the entire recording in memory before writing it to a file. A rather astute reader asked me what would happen if the user let the recording go long enough to fill up memory. The answer to that question is the application would crash due to an exception being trhown when it fails to allocate more memory and all of the recordingwould be lost. I had already been thinking of a sime reusable solution for doing this but I also offered to the user the following code sample to handle streaming directly to IsolatedStorage.
My two goals in writing it were to keep it simple and keep it portable/reusable. As far as usage goes I can’t think of any ways to make it any easier.
   //To start a recording
   StreamingRecorder myRecorder  = new StreamingRecorder();

  //To stop a recording();
After the code has run you will have a WAVE file with a proper header ready to be consumed by a SoundEffect, MediaElement, or whatever it is that you want to do with it.

In implementing this I must say that I have a hiher appreciation for how MediaElement‘s interface is designed. The starting and stopping process are not immediate. In otherwords when you call Start() or Stop() it is not until a few moments later that the request is fully processed. Because of the asynchronous nature of these processes I’ve implemented the event RecordingStateChanged and the property RecordingState so that I would know when a state change was complete. If you are familiar with the media element class then your recognize the similarity of this pattern.
I’ll go into further details on how this works along with implemeting some other functionality (such as a Pause method) in a later post. But the code is in a working state now so I’m sharing it. 🙂
Here is the source:
public class StreamingRecorder :INotifyPropertyChanged,  IDisposable

    object SyncLock = new object();

    private Queue<MemoryStream> _availablBufferQueue;        
    private Queue<MemoryStream> _writeBufferQueue;

    private int _bufferCount;
    private byte[] _audioBuffer;

    //private int _currentRecordingBufferIndex;

    private TimeSpan _bufferDuration;
    private int _bufferSize;
    private Stream _outputStream;
    private Microphone _currentMicrophone;
    private bool _ownsStream = false;
    private long _startPosition;


    public  StreamingRecorder(TimeSpan? bufferDuration = null, int bufferCount=2)
        _bufferDuration = bufferDuration.HasValue ? bufferDuration.Value : TimeSpan.FromSeconds(0);
        _bufferCount = bufferCount;
        _currentMicrophone= Microphone.Default;   

    private MemoryStream CurrentBuffer
        get; set;

    public void Start(string fileName)
        var isoStore = System.IO.IsolatedStorage.IsolatedStorageFile.GetUserStoreForApplication();
        var targetFile = isoStore.OpenFile(fileName, FileMode.Create);
        Start(targetFile, true);

    public void Start(Stream outputStream, bool ownsStream=false)
        _outputStream = outputStream;
        _ownsStream = ownsStream;
        _startPosition = outputStream.Position;

        Size = 0;

        //Create our recording buffers
        _availablBufferQueue = new Queue<MemoryStream>();
        _writeBufferQueue = new Queue<MemoryStream>();
        _audioBuffer = new byte[_currentMicrophone.GetSampleSizeInBytes(_currentMicrophone.BufferDuration)];
        _bufferSize = _currentMicrophone.GetSampleSizeInBytes(_bufferDuration + _currentMicrophone.BufferDuration);
        for (var i = 0; i < _bufferCount; ++i)
            _availablBufferQueue.Enqueue(new MemoryStream(_bufferSize));

        CurrentBuffer = _availablBufferQueue.Dequeue();
        //Stuff a bogus wave header in the output stream as a space holder.
        //we will come back and make it valid later. For now the size is invalid.
        //I could have just as easily stuffed any set of values here as long as 
        //the size of those values equaled 0x2C
        WaveHeaderWriter.WriteHeader(CurrentBuffer, -1, 1, _currentMicrophone.SampleRate);
        Size += (int)CurrentBuffer.Position;

        //Subscribe to the Microphone's buffer ready event and start listening.
        _currentMicrophone.BufferReady += new EventHandler<EventArgs>(_currentMicrophone_BufferReady);            

    void _currentMicrophone_BufferReady(object sender, EventArgs e)
        //If the recorder is paused (not implemented) then don't add this audio chunk to
        // the output. If HasFlushed is set then the recording is actually ready to shut
        //down and we shouldn't accumulate anything more. 
        if ((CurrentState != RecordingState.Paused))
            //Append the audio chunk to our current buffer
            CurrentBuffer.Write(_audioBuffer, 0, _audioBuffer.Length);
            //Increment the size of the recording.
            Size += _audioBuffer.Length;
            //If the buffer is full or if we are shutting down then we need to submit
            //the buffer to be written to the output stream.
            if ((CurrentBuffer.Length > _bufferSize)||(CurrentState==RecordingState.Stopping))

                //If we were shutting down then set a flag so that it is known that the last audio
                //chunk has been written. 
                if (CurrentState == RecordingState.Stopping)
                    _currentMicrophone.BufferReady -= _currentMicrophone_BufferReady;
                CurrentBuffer = _availablBufferQueue.Count > 0 ? _availablBufferQueue.Dequeue() : new MemoryStream();


    // CurrentState - generated from ObservableField snippet - Joel Ivory Johnson

    private RecordingState _currentState;
    public RecordingState CurrentState
        get { return _currentState; }
            if (_currentState != value)
                _currentState = value;

    void WriteData(object a )

            while (_writeBufferQueue.Count > 0)
                var item = _writeBufferQueue.Dequeue();
                var buffer = item.GetBuffer();
                _outputStream.Write(buffer, 0,(int) item.Length);


                if (CurrentState == RecordingState.Stopping)
                    //Correct the information in the wave header. After it is
                    //written set the file pointer back to the end of the file.
                    long prePosition = _outputStream.Position;
                    _outputStream.Seek(_startPosition, SeekOrigin.Begin);
                    _outputStream.Seek(prePosition, SeekOrigin.Begin);
                    if (_ownsStream)
                    CurrentState = RecordingState.Stopped;

    void SubmitToWriteBuffer(MemoryStream target)
        //Do the writing on another thread so that processing on this thread can continue. 
        ThreadPool.QueueUserWorkItem(new WaitCallback(WriteData));

    public void Pause()
        if ((CurrentState != RecordingState.Paused) && (CurrentState != RecordingState.Recording))
            throw new Exception("you can't pause if you are not recording");
        CurrentState = RecordingState.Paused;

    public void Stop()
        CurrentState = RecordingState.Stopping;

    // Size - generated from ObservableField snippet - Joel Ivory Johnson

    private int  _size;
    public int Size
        get { return _size; }
            if (_size != value)
                _size = value;

    public long RemainingSpace
            return System.IO.IsolatedStorage.IsolatedStorageFile.GetUserStoreForApplication().AvailableFreeSpace;

    public TimeSpan RecordingDuration
            return _currentMicrophone.GetSampleDuration((int)Size);

    public TimeSpan RemainingRecordingTime
            return _currentMicrophone.GetSampleDuration((int)RemainingSpace);


    public event EventHandler<RecordingStateChangedEventArgs> RecordingStateChanged;
    protected void OnRecordingStateChanged(RecordingState newState)
            RecordingStateChanged(this, new RecordingStateChangedEventArgs(){NewState = newState});

    public event PropertyChangedEventHandler PropertyChanged;
    protected void OnPropertyChanged(string propertyName)
        if (PropertyChanged != null)
            PropertyChanged(this, new PropertyChangedEventArgs(propertyName));

    public void Dispose()