AudioGraph API FrameOutputNode

Hey there!

I was missing something in the examples for the Audio Graph API that is really important for me and I can’t figure out how to do this.

When I connect an AudioFileInputNode with an AudioFrameOutputNode, I am able to the read frame data with this code:

AudioBuffer buffer = frame.LockBuffer(AudioBufferAccessMode.Read);
IMemoryBufferReference reference = buffer.CreateReference();
((IMemoryBufferByteAccess)reference).GetBuffer(out byte* dataInBytes, out uint capacityInBytes);
float* dataInFloats = (float*)dataInBytes;

As I learned from this stackoverflow answer, the bytes that read into the buffer have to be considered as bytes of floats. But how do I get the actual byte data of a file that I opened with the graph from these floats?

For example, I read a WAV with PCM 16000Hz, 16 bit mono and the raw bytes of that wav are shorts:

0 0 | 255 255 | 254 255 | 255 255 | 255 255 | 254 255 | 253 255 | 252 255 ...

But the buffer of AudioFrameOutputNode reads the following bytes:

0 0 0 0 | 0 0 0 184 | 0 0 128 184 | 0 0 0 184 ...

which I assume are floats. But the question is how do I get the actual raw shorts from these floats? Is there some type of conversion I can use? As you can see in the example, the sample 255 255 in shorts seems to map to sample 0 0 0 184 in float. What’s the magic behind that?

[oldnewthing]

The sample does show that it’s an array of floats. If the original format wasn’t an array of floats, then there is naturally a format conversion. The nature of the format conversion depends on the format. In your specific example, the conversion was linear. Other formats can be more complicated (such as u-law). It’s really out of the scope of the documentation to show how to convert every possible format. You need to apply your understanding of the format. I guess the documentation could say “The buffer is in the form of an array of samples. Each sample is a series of IEEE single-precision floating point values in a linear range from -1.0 to +1.0, one value per channel.” It’s then up to you to understand how your format converts to that format.

[Zofware]

I have limited experience here, but every audio graph I’ve ever set up seems to pass float samples around regardless of how I configure the encoding properties. Sorry I can’t be of more help here!