fix: prevent panic in AudioInfo::channel_layout() for unsupported channel counts

[noeljackson]

Summary

Fixes crash and chipmunk audio on systems with professional audio interfaces that report high channel counts (e.g., RME Digiface with 34 outputs).

The Problem

When using Cap with a multi-channel audio interface, users experienced:

1. Crashes - App panicked with "called Option::unwrap() on a None value" in channel_layout()
2. Chipmunk audio - Playback was ~4x faster than normal, making voices sound like chipmunks

Root Cause

FFmpeg only supports channel layouts up to 8 channels (7.1 surround). When a device reported more channels (like 34), several things broke:

1. channel_layout() called .unwrap() on None because there's no FFmpeg layout for 34 channels
2. Audio buffers were sized for 34 channels but only 8 channels of data were produced
3. The output audio stream expected 34 channels of interleaved samples, but received only 8 - causing playback at 34/8 ≈ 4.25x speed

The Fix

• Store the actual device channel count to correctly parse interleaved audio data
• Clamp to 8 channels max when interfacing with FFmpeg (resampling, encoding)
• Configure the output audio stream to match the clamped channel count
• Add for_ffmpeg_output() helper for consistent clamping across the codebase
• Make channel_layout() gracefully handle any channel count

Files Changed

• crates/media-info/src/lib.rs - Core AudioInfo fixes and helper method
• crates/editor/src/audio.rs - AudioResampler and buffer fixes
• crates/editor/src/playback.rs - Stream configuration fixes

Testing

Tested on Mac with RME Digiface (34 output channels) + RØDE NT-USB Mini microphone:

• ✅ No crash on playback
• ✅ Audio plays at correct speed
• ✅ Recording works correctly

Greptile Summary

This PR fixes crashes and chipmunk audio playback on systems with professional audio interfaces reporting more than 8 channels (e.g., RME Digiface with 34 outputs).

Key Changes:

• AudioInfo now stores the actual device channel count for correct audio data parsing, while clamping to 8 channels when interfacing with FFmpeg (which only supports up to 7.1 surround)
• Added for_ffmpeg_output() helper method for consistent channel clamping across the codebase
• Updated channel_layout() to gracefully handle 0-channel and excessive channel counts instead of panicking
• Modified wrap_frame_with_max_channels() to use actual channel count for input parsing while limiting output to FFmpeg-compatible layouts
• Applied clamping in all audio processing pipelines (AudioResampler, AudioPlaybackBuffer, PrerenderedAudioBuffer) and stream creation
• Updated audio output stream configuration to match clamped channel counts, preventing speed issues
• Added comprehensive tests for edge cases (0 channels, 9-64 channels)

Impact:
The fix properly separates the concern of parsing device audio data (which needs the real channel count) from FFmpeg encoding constraints (max 8 channels), eliminating both the panic and the playback speed issue.

Confidence Score: 5/5

• This PR is safe to merge with minimal risk
• The fix is well-designed with clear separation of concerns between device channel parsing and FFmpeg encoding limits. All edge cases (0 channels, excessive channels) are handled with proper clamping and fallbacks. The implementation includes comprehensive unit tests covering the edge cases. Changes are consistently applied across all audio processing paths. The approach preserves correct audio data parsing while preventing FFmpeg-related panics and playback speed issues.
• No files require special attention

Important Files Changed

Filename	Overview
crates/media-info/src/lib.rs	Core fix: stores actual device channel count, clamps to 8 for FFmpeg, adds for_ffmpeg_output() helper, prevents panics in channel_layout(), handles 0-channel and 34+ channel edge cases correctly, includes comprehensive tests
crates/editor/src/audio.rs	Applies channel clamping via for_ffmpeg_output() in AudioPlaybackBuffer, AudioResampler, and PrerenderedAudioBuffer constructors to ensure consistent FFmpeg compatibility across all audio processing pipelines
crates/editor/src/playback.rs	Updates audio stream creation to clamp output info and sync stream config channels with clamped values in both create_stream() and create_stream_prerendered(), preventing channel count mismatches during playback

Sequence Diagram

sequenceDiagram
    participant Device as Audio Device<br/>(34 channels)
    participant AudioInfo as AudioInfo
    participant Resampler as AudioResampler
    participant FFmpeg as FFmpeg<br/>(max 8 channels)
    participant Output as Audio Output Stream

    Note over Device,Output: Recording/Playback Initialization

    Device->>AudioInfo: from_stream_config(34 channels)
    activate AudioInfo
    Note over AudioInfo: Store actual channel count (34)<br/>for correct data parsing
    AudioInfo-->>Device: AudioInfo { channels: 34 }
    deactivate AudioInfo

    Note over Device,Output: Audio Processing Pipeline

    Device->>AudioInfo: wrap_frame_with_max_channels(data, 8)
    activate AudioInfo
    Note over AudioInfo: Use 34 channels to parse input<br/>Extract only first 8 channels<br/>Avoid div-by-zero with max(1)
    AudioInfo->>FFmpeg: Create frame with 8 channels
    activate FFmpeg
    FFmpeg-->>AudioInfo: Audio frame (8 channels)
    deactivate FFmpeg
    AudioInfo-->>Device: Properly formatted frame
    deactivate AudioInfo

    Device->>Resampler: new(output_info)
    activate Resampler
    Note over Resampler: Call for_ffmpeg_output()<br/>Clamp to 8 channels
    Resampler->>AudioInfo: for_ffmpeg_output()
    activate AudioInfo
    Note over AudioInfo: Return AudioInfo { channels: 8 }
    AudioInfo-->>Resampler: Clamped AudioInfo
    deactivate AudioInfo
    Resampler->>FFmpeg: Create resampler context (8 channels)
    FFmpeg-->>Resampler: Resampler ready
    deactivate Resampler

    Device->>Output: Configure stream
    activate Output
    Note over Output: Set config.channels = 8<br/>(matches clamped AudioInfo)
    Output-->>Device: Stream configured
    deactivate Output

    Note over Device,Output: Result: No panic, correct playback speed

Loading

[greptile-apps]

Additional Comments (2)

1. crates/media-info/src/lib.rs, line 162-163 (link)

   logic: unreachable!() is now reachable since new_raw allows 0 channels

   With the changes in this PR, AudioInfo::new_raw() can now create instances with 0 channels (as shown in the test at line 419). If such an instance calls wrap_frame_with_max_channels(), it will hit this unreachable!() and panic.

   Either handle this case properly or add validation to prevent 0-channel AudioInfo from calling this method.

2. crates/media-info/src/lib.rs, line 152-153 (link)

   logic: Division by zero when self.channels == 0

   If self.channels is 0, then packed_sample_size will be 0, causing a division by zero panic on line 153. Since new_raw now allows 0 channels, this is a reachable panic condition.

_{1 file reviewed, 2 comments}

_{Edit Code Review Agent Settings | Greptile}

[greptile-apps]

Greptile found no issues!

From now on, if a review finishes and we haven't found any issues, we will not post anything, but you can confirm that we reviewed your changes in the status check section.

_{This feature can be toggled off in your Code Review Settings by deselecting "Create a status check for each PR".}

[richiemcilroy]

Legend... thank you.