[Audio Bug] Microphone fails to start on second room entry – AudioUnit error 561145187 (!pri)

[SailorGa]

We encountered an audio issue when using WebRTC-SDK (version 137.7151.01) to implement video conferencing.
• On the first join to a meeting, everything works as expected;
• After leaving and rejoining the meeting, the microphone fails to start;
• Console error output:

(voice_processing_audio_unit.mm:402 Start): Failed to start audio unit. Error=561145187 (audio_device_ios.mm:335 StartRecording): StartRecording failed to start audio unit, reason 561145187

Error Analysis:

The error code 561145187 corresponds to '!pri' (kAudioUnitErr_CannotDoInCurrentContext), which means AudioUnit could not operate in the current context. Possible reasons include:
• Audio resources (e.g., AudioUnit) not released properly after first session;
• AVAudioSession was not reactivated before restarting audio;
• Audio device module may have retained static/shared state;
• Audio stack was not reset properly after background/interrupt state;

Expected Behavior:
• Microphone should work correctly on every room entry;
• Rejoining a room should not cause audio unit startup failure;

Reproduction Steps:
1. Launch the app and join a room (microphone works fine);
2. Leave the room, cleaning up PeerConnection and Factory;
3. Join the room again;
4. Observe error Failed to start audio unit and no mic input;

SDK Version & Environment:
• SDK: WebRTC-SDK 137.7151.01
• Integration: CocoaPods (source 'https://github.com/webrtc-sdk/Specs.git')
• Language: Objective-C
• Tested on: iOS 15+, real device

Suggestion / Request:

Please help check whether the SDK properly releases/cleans up the audio resources (such as AudioUnit, AVAudioSession, etc.) when leaving a room, and whether shared/static audio states persist across sessions.

[guodi1014]

I also encountered the same problem. It does exist. May I ask if you have found a solution? This issue still exists in version 137.7151.03

[guodi1014]

After testing, it was found that reconfiguring the audio each time a connection is made can solve this problem.
let audioSession = AVAudioSession.sharedInstance() do { try audioSession.setCategory(.playAndRecord, mode: .voiceChat, options: []) try audioSession.setActive(true) } catch { print("set audio session error: \(error.localizedDescription)") }

[SailorGa]

After testing, it was found that reconfiguring the audio each time a connection is made can solve this problem.经过测试，发现每次建立连接时重新配置音频可以解决这个问题。 let audioSession = AVAudioSession.sharedInstance() do { try audioSession.setCategory(.playAndRecord, mode: .voiceChat, options: []) try audioSession.setActive(true) } catch { print("set audio session error: \(error.localizedDescription)") }

Yes, ​​I can confirm this solution works for me too.​​