CODEWITHSUNDEEP
c#audiodirectxdirectsound
tenpn
tenpn

Reputation: 4716

Why do some wav files play in my c# directsound app but some don't?

I've got a c# application that plays simple wav files through directsound. With the test data I had, the code worked fine. However when I used real-world data, it produced a very unhelpful error on creation of the secondary buffer: "ArgumentException: Value does not fall within the expected range."

The test wavs had a 512kbps bit rate, 16bit audio sample size, and 32kHz audio sample rate. The new wavs is 1152kbps, 24bit and 48kHz respectively. How can I get directsound to cope with these larger values, or if not how can I programatically detect these values before attempting to play the file?

it's managed DirectX v9.00.1126 I'm using, and I've included some sample code below:

using DS = Microsoft.DirectX.DirectSound;  
...  
DS.Device device = new DS.Device();
device.SetCooperativeLevel(this, CooperativeLevel.Normal);  
...
BufferDescription bufferDesc = new BufferDescription();
bufferDesc.ControlEffects = false;  
...
try
{
    SecondaryBuffer sound = new SecondaryBuffer(path, bufferDesc, device);
    sound.Play(0, BufferPlayFlags.Default);
}
...

Additional info: the real-world wav files won't play in windows media player either, telling me a codec is needed to play the file, while they play fine in winamp.

Additional info 2: Comparing the bytes of the working test data and the bad real-world data, I can see that past the RIFF chunk, the bad data has a "bext" chunk, that the internet informs me is metadata associated with the broadcast audio extension, while the test data goes straight into a fmt chunk. There /is/ a fmt chunk in the bad data, so I don't know if it's badly-formed or if the loaders should be looking further for fmt data. I can see if I can get some information on this rouge bext chunk from the people supplying me the data - if they can remove it my code may still work.

Upvotes: 3

Views: 4812

Answers (2)

Vinko Vrsalovic
Vinko Vrsalovic

Reputation: 340446

In addition to the sampling issue, WAV is just a container format and the audio could be compressed in any of a myriad of audio formats (just like AVI is a container of video). So you could use a tool like GSpot to find out if your WAV is encoded in a non-standard format in and install the codec. Winamp has more codecs installed by default than WMP, which would explain the Winamp plays it and WMP doesn't.

Upvotes: 4

Mark Heath
Mark Heath

Reputation: 49522

Not all soundcards support 24 bit sample playback, and even when they do, they often have to be exclusively opened in that mode. There is a similar issue with sample rates. Your soundcard may be operating at 44.1kHz, in which case 48kHz needs to be resampled to be played.

I have written an open source .NET audio library called NAudio which will allow you to find out what sample rate and bit depth a given WAV file is. It also offers alternative ways of playing back audio (e.g. through the Wav... APIs), and the ability to resample files using the DMO resampler object.

Upvotes: 7

Related Questions