When calling getWaveform or getWaveformBytes on Android with medium to large audio files (e.g., a 5-minute MP3), the process fails with a native index error. This seems to be an integer overflow in the native Android/JNI implementation when calculating byte offsets.
The same file works perfectly on iOS and Web.
Log-Output: I/flutter (16357): Conversion failed: Index -2146671 out of bounds for length 15567358
(Note: The negative index -2146671 strongly suggests a 32-bit integer overflow during offset calculation.)
Environment:
Platform: Android (Tested on Emulator)
OS Version: Android 12+ (API 31+)
Library Version: [0.8.0]
Steps to Reproduce:
Load an MP3 file with a duration of approximately 4-6 minutes.
Call AudioDecoder.getWaveform(path, numberOfSamples: 1000).
The extraction fails with the mentioned IndexOutOfBoundsException.
Expected Behavior:
The waveform should be extracted successfully, regardless of the file length, by using 64-bit integers (long) for byte offset calculations in the native layer.
Workaround:
Currently, the issue can only be avoided by trimming the audio into smaller chunks (e.g., 30-second segments) and extracting them individually, which is highly inefficient.
When calling getWaveform or getWaveformBytes on Android with medium to large audio files (e.g., a 5-minute MP3), the process fails with a native index error. This seems to be an integer overflow in the native Android/JNI implementation when calculating byte offsets.
The same file works perfectly on iOS and Web.
Log-Output: I/flutter (16357): Conversion failed: Index -2146671 out of bounds for length 15567358
(Note: The negative index -2146671 strongly suggests a 32-bit integer overflow during offset calculation.)
Environment:
Platform: Android (Tested on Emulator)
OS Version: Android 12+ (API 31+)
Library Version: [0.8.0]
Steps to Reproduce:
Load an MP3 file with a duration of approximately 4-6 minutes.
Call AudioDecoder.getWaveform(path, numberOfSamples: 1000).
The extraction fails with the mentioned IndexOutOfBoundsException.
Expected Behavior:
The waveform should be extracted successfully, regardless of the file length, by using 64-bit integers (long) for byte offset calculations in the native layer.
Workaround:
Currently, the issue can only be avoided by trimming the audio into smaller chunks (e.g., 30-second segments) and extracting them individually, which is highly inefficient.