Recorder doesn’t work for SampleFormat.S16

[krishnavsr]

Requirements

• I have searched the existing issues to ensure this bug has not already been reported.
• I have provided a clear and concise description of the bug.
• I have provided detailed steps to reproduce the bug.
• I have included a minimal reproducible example (MRE) if applicable.
• I have provided error messages and stack traces if applicable.
• I have completed the environment information.

1. Describe the Bug

When I tried to use Recorder with SampleFormat.S16 , sample rate 16KHz and channels 1

Using the SoundFlow.Samples.SimplePlayer sample and option 3. Record and playback audio

I hear back garbled audio. Based on my analysis of the code I think the source of the problem is AudioProcessingCallback always sends Span<float> for all sample formats. But for SampleFormat.S16 it should be sending Span<short>

Is there a conversion that needs to be done from Span<float> to Span<short>? It should work for the selected SampleFormat without any post processing.

2. Steps to Reproduce

Step 1:- Run SoundFlow.Samples.SimplePlayer sample and select option 3. Record and playback audio

Recording and Playback works fine

Step 2:- Change the SampleFormat.F32 to SampleFormat.S16, build and run then choose option 3

Recording and Playback does not work

3. Expected Behavior

It should work the same for different SampleFormats that are supported

4. Current Behavior

It plays back noise with SampleFormat.S16

5. Minimal Reproducible Example (MRE)

AudioProcessCallback(Span\<float\> in AudioEngine.cs

6. Error Messages and Stack Trace (if applicable)

SoundFlow Version

3c7d8d2

.NET Version

net8.0

Operating System

macOS Sonoma 14.5

Architecture

ARM64

Audio Backend Used (if known)

No response

Specific Audio Hardware (if relevant)

No response

8. Affected Components/Modules (if known)

No response

9. Impact

incorrect audio output

10. Possible Workaround (if known)

No response

11. Additional Context

No response

[LSXPrime]

First, thank you for reporting this. My apologies for the delay in responding; I wanted to investigate the matter thoroughly before getting back to you.

I have confirmed the issue and have implemented a preliminary fix. I refactored the outdated protected void ProcessAudioInput(nint input, int length) method to be identical to protected void ProcessGraph(nint output, int length). This change successfully eliminated the garbled audio.

However, I am now encountering silence on non-F32 inputs. I am continuing to investigate this further.

[krishnavsr]

Great, thanks for looking into it. I made a temporary crude fix on my side by defining AudioProcessCallbackShort as an additional delegate and using that instead. Can confirm that's working for S16 so I am not blocked by this.

Let me know if you have a branch and I can help with testing if required.

[LSXPrime]

I got it working! It turns out I had the logic backward. I was trying to convert the generated float buffer to the native format, when I actually needed to convert the native buffer to float. Can't believe I missed that! Anyway, it's fixed now. I pushed the commit, and it should be in the next release.