Audio codec properties not respected in release builds on iOS

[felixspitzer]

Before submitting a new issue

• I tested using the latest version of the library, as the bug might be already fixed.
• I tested using a supported version of react native.
• I checked for possible duplicate issues, with possible answers.

Bug summary

We are setting the number of channels for iOS and Android through the given properties. While all the config settings work as expected in debug builds, non of them are respected in release builds of the app.

const newFilename = await audioRecorderPlayer.startRecorder(tempFilename, { AudioSourceAndroid: 0, AVFormatIDKeyIOS: 'aac', AVEncodingOptionIOS: 'aac', OutputFormatAndroid: 5, AudioEncoderAndroid: 3, AudioSamplingRate: 22050, AudioEncodingBitRate: 64000, AudioChannels: 2, AVEncoderAudioQualityKeyIOS: AVEncoderAudioQualityIOSType.medium, AVSampleRateKeyIOS: 22050, AVNumberOfChannelsKeyIOS: 2, })

When I analyze the debug audio through afinfo terminal on my Mac I get the following output

`
File: /Users/felix/Downloads/sound_1763736577.8996592.m4a
File type ID: m4af
Num Tracks: 1

Data format: 2 ch, 22050 Hz, aac (0x00000000) 0 bits/channel, 0 bytes/packet, 1024 frames/packet, 0 bytes/frame
no channel layout.
estimated duration: 10.864036 sec
audio bytes: 76637
audio packets: 236
bit rate: 55940 bits per second
packet size upper bound: 451
maximum packet size: 451
audio data file offset: 32768
optimized
audio 239552 valid frames + 2112 priming + 0 remainder = 241664
format list:
[ 0] format: 2 ch, 22050 Hz, aac (0x00000000) 0 bits/channel, 0 bytes/packet, 1024 frames/packet, 0 bytes/frame
Channel layout: Stereo (L R)
`

The same settings, but using the release build audio file result in this output

`
File: /Users/felix/Downloads/sound_1763737051.74795.m4a
File type ID: m4af
Num Tracks: 1

Data format: 5 ch, 44100 Hz, aac (0x00000000) 0 bits/channel, 0 bytes/packet, 1024 frames/packet, 0 bytes/frame
Channel layout: 5.0 (C L R Ls Rs)
estimated duration: 10.354649 sec
audio bytes: 229256
audio packets: 448
bit rate: 176307 bits per second
packet size upper bound: 713
maximum packet size: 713
audio data file offset: 57344
optimized
audio 456640 valid frames + 2112 priming + 0 remainder = 458752
format list:
[ 0] format: 5 ch, 44100 Hz, aac (0x00000000) 0 bits/channel, 0 bytes/packet, 1024 frames/packet, 0 bytes/frame
Channel layout: 5.0 (C L R Ls Rs)
`

Our main issue that the audio channel count and the sampling rate are not as we need them to be for the audio to be handled correctly after recording.

Library version

0.2.9

Project type

React Native (bare workflow)

React Native version

0.81.4

Expo SDK version (if applicable)

No response

Environment info

System:
  OS: macOS 26.1
  CPU: (12) arm64 Apple M2 Max
  Memory: 556.05 MB / 32.00 GB
  Shell:
    version: "5.9"
    path: /bin/zsh
Binaries:
  Node:
    version: 22.19.0
    path: /Users/felix/.nvm/versions/node/v22.19.0/bin/node
  Yarn:
    version: 1.22.22
    path: /opt/homebrew/bin/yarn
  npm:
    version: 10.9.3
    path: /Users/felix/.nvm/versions/node/v22.19.0/bin/npm
  Watchman:
    version: 2025.03.03.00
    path: /opt/homebrew/bin/watchman
Managers:
  CocoaPods:
    version: 1.16.2
    path: /opt/homebrew/bin/pod
SDKs:
  iOS SDK:
    Platforms:
      - DriverKit 25.1
      - iOS 26.1
      - macOS 26.1
      - tvOS 26.1
      - visionOS 26.1
      - watchOS 26.1
  Android SDK: Not Found
IDEs:
  Android Studio: 2025.2 AI-252.25557.131.2521.14432022
  Xcode:
    version: 26.1.1/17B100
    path: /usr/bin/xcodebuild
Languages:
  Java:
    version: 17.0.9
    path: /usr/bin/javac
  Ruby:
    version: 3.2.2
    path: /Users/felix/.rbenv/shims/ruby
npmPackages:
  "@react-native-community/cli":
    installed: 20.0.0
    wanted: 20.0.0
  react:
    installed: 19.1.0
    wanted: 19.1.0
  react-native:
    installed: 0.81.4
    wanted: 0.81.4
  react-native-macos: Not Found
npmGlobalPackages:
  "*react-native*": Not Found
Android:
  hermesEnabled: true
  newArchEnabled: true
iOS:
  hermesEnabled: true
  newArchEnabled: true

Steps to reproduce

1. Run config above in debug build
2. Run config above in release build

Reproducible example repository

will provide one shortly

[savinov007]

Same issue for me

[hyochan]

Hi @felixspitzer, @savinov007

Thanks for reporting this issue. The iOS implementation was not respecting cross-platform audio settings (AudioChannels, AudioSamplingRate, AudioEncodingBitRate). It only checked for iOS-specific keys like AVSampleRateKeyIOS and AVNumberOfChannelsKeyIOS.

The debug vs release difference is likely caused by audio session hijacking behavior that varies between build configurations.

Fix

I've pushed a fix in #743

1. Adds AudioQuality preset support on iOS (matching Android behavior):

   • low: 22050 Hz, 1 channel, 64 kbps
   • medium: 44100 Hz, 1 channel, 128 kbps
   • high: 48000 Hz, 2 channels, 192 kbps (default)

2. Adds cross-platform settings fallback on iOS:

   • AudioSamplingRate → now respected if AVSampleRateKeyIOS is not set
   • AudioChannels → now respected if AVNumberOfChannelsKeyIOS is not set
   • AudioEncodingBitRate → now applied to AVEncoderBitRateKey

Priority Order (after fix)

1. iOS-specific settings (AVSampleRateKeyIOS, AVNumberOfChannelsKeyIOS)
2. Cross-platform settings (AudioSamplingRate, AudioChannels, AudioEncodingBitRate)
3. AudioQuality preset
4. Default (HIGH quality)

Workaround (until next release)

Your config should work now, but if you want to ensure compatibility with older versions, always use iOS-specific keys:

await audioRecorderPlayer.startRecorder(tempFilename, {
  // Use iOS-specific keys (these always worked)
  AVFormatIDKeyIOS: 'aac',
  AVSampleRateKeyIOS: 22050,
  AVNumberOfChannelsKeyIOS: 2,
  AVEncoderAudioQualityKeyIOS: AVEncoderAudioQualityIOSType.medium,

  // Android settings
  AudioSourceAndroid: 0,
  OutputFormatAndroid: 5,
  AudioEncoderAndroid: 3,
  AudioSamplingRate: 22050,
  AudioChannels: 2,
  AudioEncodingBitRate: 64000,
});

The fix will be included in the next release.

[felixspitzer]

@hyochan put I did set the iOS specific keys as mentioned in my report.

They work just fine in debug mode, but once the mode is set to release even the iOS specific keys are not respected.

I can check in the morning if your fix changes anything regarding this behavior, but I just wanted to correct the statement that iOS specific keys worked previously

[felixspitzer]

@hyochan just checked out the new version. As it was before, the iOS settings are not respected. Using your config (which is basically the config I posted) the audio info comes back as:

`
felix@Mac ~ % afinfo /Users/felix/Downloads/sound_1764228287.611422.m4a
File: /Users/felix/Downloads/sound_1764228287.611422.m4a
File type ID: m4af
Num Tracks: 1

Data format: 5 ch, 8000 Hz, aac (0x00000000) 0 bits/channel, 0 bytes/packet, 1024 frames/packet, 0 bytes/frame
Channel layout: 5.0 (C L R Ls Rs)
estimated duration: 5.880000 sec
audio bytes: 17070
audio packets: 48
bit rate: 22226 bits per second
packet size upper bound: 492
maximum packet size: 492
audio data file offset: 16384
optimized
audio 47040 valid frames + 2112 priming + 0 remainder = 49152
format list:
[ 0] format: 5 ch, 8000 Hz, aac (0x00000000) 0 bits/channel, 0 bytes/packet, 1024 frames/packet, 0 bytes/frame
Channel layout: 5.0 (C L R Ls Rs)

`

The only thing that changed is that the sample rate is now 8kHz - which doesn't make sense at all given the config
Channel count is still 5 - not 2 as defined in the config

[felixspitzer]

I can reproduce this in the example of this repo as well.
Set the config as you mentioned, build in release variant in Xcode, analyze with afinfo.

config:
2 channels, 22kHz

actual release result:
5 channels, 8kHz

[hyochan]

Hi @felixspitzer, @savinov007

Thanks for the detailed report. This is a very interesting issue - the fact that you're getting 5 channels (5.0 surround) and 8kHz sample rate in release builds is quite unusual, as iOS microphone recording cannot physically produce 5-channel audio.

Analysis

The symptoms suggest that the AVAudioRecorder settings are being completely ignored in release builds, and some unexpected audio session configuration is being applied instead.

Possible causes:

1. Swift compiler optimization issue - The AudioSet struct properties may not be correctly accessed in release builds due to optimization
2. Closure capture problem - The audioSets parameter might not be properly captured in nested closures during release optimization
3. Audio session hijacking - Another app or system service may be taking over the audio session with different settings

Debugging Request

To help diagnose this issue, could you add the following debug logging to your local copy of ios/Sound.swift?

Find the getAudioSettings function (around line 722) and replace it with this version that includes detailed logging:

private func getAudioSettings(audioSets: AudioSet?) -> [String: Any] {
    var settings: [String: Any] = [:]

    // Default to HIGH quality if not specified
    let audioQuality = audioSets?.AudioQuality ?? .high
    let defaults = Self.qualityPresets[audioQuality] ?? Self.qualityPresets[.high]!

    print("🎙️ [getAudioSettings] audioSets is \(audioSets == nil ? "nil" : "present")")
    print("🎙️ [getAudioSettings] audioQuality: \(audioQuality)")
    print("🎙️ [getAudioSettings] defaults - samplingRate: \(defaults.samplingRate), channels: \(defaults.channels), bitrate: \(defaults.bitrate)")

    // Apply default settings based on AudioQuality
    settings[AVFormatIDKey] = Int(kAudioFormatMPEG4AAC)
    settings[AVSampleRateKey] = defaults.samplingRate
    settings[AVNumberOfChannelsKey] = defaults.channels
    settings[AVEncoderBitRateKey] = defaults.bitrate
    settings[AVEncoderAudioQualityKey] = defaults.encoderQuality.rawValue

    // Apply custom settings with explicit overrides taking precedence
    if let audioSets = audioSets {
        print("🎙️ [getAudioSettings] Processing audioSets...")
        print("🎙️ [getAudioSettings] AVSampleRateKeyIOS: \(String(describing: audioSets.AVSampleRateKeyIOS))")
        print("🎙️ [getAudioSettings] AVNumberOfChannelsKeyIOS: \(String(describing: audioSets.AVNumberOfChannelsKeyIOS))")
        print("🎙️ [getAudioSettings] AudioSamplingRate: \(String(describing: audioSets.AudioSamplingRate))")
        print("🎙️ [getAudioSettings] AudioChannels: \(String(describing: audioSets.AudioChannels))")

        // iOS-specific settings take highest priority
        if let sampleRate = audioSets.AVSampleRateKeyIOS {
            print("🎙️ [getAudioSettings] Using AVSampleRateKeyIOS: \(sampleRate)")
            settings[AVSampleRateKey] = sampleRate
        } else if let audioSamplingRate = audioSets.AudioSamplingRate {
            print("🎙️ [getAudioSettings] Using AudioSamplingRate: \(audioSamplingRate)")
            settings[AVSampleRateKey] = Int(audioSamplingRate)
        }

        if let channels = audioSets.AVNumberOfChannelsKeyIOS {
            print("🎙️ [getAudioSettings] Using AVNumberOfChannelsKeyIOS: \(channels)")
            settings[AVNumberOfChannelsKey] = Int(channels)
        } else if let audioChannels = audioSets.AudioChannels {
            print("🎙️ [getAudioSettings] Using AudioChannels: \(audioChannels)")
            settings[AVNumberOfChannelsKey] = Int(audioChannels)
        }

        if let bitRate = audioSets.AudioEncodingBitRate {
            print("🎙️ [getAudioSettings] Using AudioEncodingBitRate: \(bitRate)")
            settings[AVEncoderBitRateKey] = Int(bitRate)
        }

        if let quality = audioSets.AVEncoderAudioQualityKeyIOS {
            let mappedQuality = mapToAVAudioQuality(quality)
            print("🎙️ [getAudioSettings] Using AVEncoderAudioQualityKeyIOS: \(quality) -> \(mappedQuality)")
            settings[AVEncoderAudioQualityKey] = mappedQuality
        }

        if let format = audioSets.AVFormatIDKeyIOS {
            let formatID = getAudioFormatID(from: format)
            print("🎙️ [getAudioSettings] Using AVFormatIDKeyIOS: \(format) -> \(formatID)")
            settings[AVFormatIDKey] = formatID
        }
    }

    print("🎙️ [getAudioSettings] Final settings: \(settings)")

    return settings
}

How to View Logs in Release Build

Since release builds don't show logs in Xcode console by default, you'll need to use one of these methods:

Option 1: Use Console.app (Recommended)

1. Build and run your release build on a connected device
2. Open Console.app on your Mac
3. Select your device from the sidebar
4. Filter by your app name or search for "🎙️"
5. Start recording and check the logs

Option 2: Use os_log (Alternative)

If print statements don't appear, you can replace print with os_log:

import os.log

// Then use:
os_log("🎙️ [getAudioSettings] audioSets is %{public}@", audioSets == nil ? "nil" : "present")

What We're Looking For

Please share the logs from both:

1. Debug build - to confirm settings are correctly parsed
2. Release build - to see if/where the settings diverge

Specifically, we want to know:

• Is audioSets nil in release builds?
• Are the individual properties (AVSampleRateKeyIOS, AVNumberOfChannelsKeyIOS, etc.) returning nil when they shouldn't?
• What are the "Final settings" in release vs debug?

Additional Questions

1. Are you using any other audio-related libraries in your app?
2. Does your app use any audio session configuration elsewhere?
3. What iOS version are you testing on?
4. Is this reproducible in the example app of this repo when built in release mode?

Looking forward to your findings!

[felixspitzer]

Hi

Debug Logs:

🎙️ [getAudioSettings] audioSets is present
🎙️ [getAudioSettings] audioQuality: margelo::nitro::sound::AudioQualityType(rawValue: 2)
🎙️ [getAudioSettings] defaults - samplingRate: 48000, channels: 2, bitrate: 192000
🎙️ [getAudioSettings] Processing audioSets...
🎙️ [getAudioSettings] AVSampleRateKeyIOS: Optional(22050.0)
🎙️ [getAudioSettings] AVNumberOfChannelsKeyIOS: Optional(2.0)
🎙️ [getAudioSettings] AudioSamplingRate: Optional(22050.0)
🎙️ [getAudioSettings] AudioChannels: Optional(2.0)
🎙️ [getAudioSettings] Using AVSampleRateKeyIOS: 22050.0
🎙️ [getAudioSettings] Using AVNumberOfChannelsKeyIOS: 2.0
🎙️ [getAudioSettings] Using AudioEncodingBitRate: 64000.0
🎙️ [getAudioSettings] Using AVEncoderAudioQualityKeyIOS: margelo::nitro::sound::AVEncoderAudioQualityIOSType(rawValue: 64) -> 64
🎙️ [getAudioSettings] Using AVFormatIDKeyIOS: margelo::nitro::sound::AVEncodingOption(rawValue: 2) -> 1633772320
🎙️ [getAudioSettings] Final settings: ["AVNumberOfChannelsKey": 2, "AVEncoderBitRateKey": 64000, "AVFormatIDKey": 1633772320, "AVEncoderQualityKey": 64, "AVSampleRateKey": 22050.0]

Release Logs:

Crashes when parsing AudioSampling rate to int which seems to be nil here. All the other values are not respected as well. Even the defaults are ignored.

🎙️ [getAudioSettings] audioSets is present
🎙️ [getAudioSettings] audioQuality: margelo::nitro::sound::AudioQualityType(rawValue: 0)
🎙️ [getAudioSettings] defaults - samplingRate: 22050, channels: 1, bitrate: 64000
🎙️ [getAudioSettings] Processing audioSets...
🎙️ [getAudioSettings] AVSampleRateKeyIOS: nil
🎙️ [getAudioSettings] AVNumberOfChannelsKeyIOS: Optional(5.0)
🎙️ [getAudioSettings] AudioSamplingRate: Optional(-nan(0x30070002ade90))
🎙️ [getAudioSettings] AudioChannels: nil
🎙️ [getAudioSettings] Using AudioSamplingRate: nan

All of this came from the example of this lib. Running the current main branch.

So Question 1 and 2 I will ignore since this is not running on in my project. And will answer 4 with YES - as I mentioned in previous comments.

3 - this is a iPhone 17 Simulator running iOS 26.1, same behavior on real device (iPhone 16 Pro, iOS 26.1)

Please just try running your own example with the given config and you will see the problem.

Any other questions I can answer to help resolve this @hyochan ?

[felixspitzer]

@hyochan do you want me to check this out before merging? :)