I need to convert a .wav file recorded with 2 audio channels to a .wav that has only 1 channel, as well as reduce the bit depth from 32 to 16. I've been trying to use AVAudioConverter.convertToBuffer
However, the conversion is throwing an error: Error Domain=NSOSStatusErrorDomain Code=-50 "(null)"
Basically, the only thing that really needs to change is to strip the audio down to a single channel, and the bit depth. I'm getting these files from a different tool, so I can't just change the way the files are recorded.
I'm not that awesome at working with audio, and I'm a bit stumped. The code I'm working on is below - is there anything I'm missing?
let inAudioFileURL:NSURL = <url_to_wav_file>
var inAudioFile:AVAudioFile?
do
{
inAudioFile = try AVAudioFile(forReading: inAudioFileURL)
}
catch let error
{
print ("error: \(error)")
}
let inAudioFormat:AVAudioFormat = inAudioFile!.processingFormat
let inFrameCount:UInt32 = UInt32(inAudioFile!.length)
let inAudioBuffer:AVAudioPCMBuffer = AVAudioPCMBuffer(PCMFormat: inAudioFormat, frameCapacity: inFrameCount)
do
{
try inAudioFile!.readIntoBuffer(inAudioBuffer)
}
catch let error
{
print ("readError: \(error)")
}
let startFormat:AVAudioFormat = AVAudioFormat.init(settings: inAudioFile!.processingFormat.settings)
print ("startFormat: \(startFormat.settings)")
var endFormatSettings = startFormat.settings
endFormatSettings[AVLinearPCMBitDepthKey] = 16
endFormatSettings[AVNumberOfChannelsKey] = 1
endFormatSettings[AVEncoderAudioQualityKey] = AVAudioQuality.Medium.rawValue
print ("endFormatSettings: \(endFormatSettings)")
let endFormat:AVAudioFormat = AVAudioFormat.init(settings: endFormatSettings)
let outBuffer = AVAudioPCMBuffer(PCMFormat: endFormat, frameCapacity: inFrameCount)
let avConverter:AVAudioConverter = AVAudioConverter.init(fromFormat: startFormat, toFormat: endFormat)
do
{
try avConverter.convertToBuffer(outBuffer, fromBuffer: inAudioBuffer)
}
catch let error
{
print ("avconverterError: \(error)")
}
As for the output:
startFormat:
["AVSampleRateKey": 16000,
"AVLinearPCMBitDepthKey": 32,
"AVLinearPCMIsFloatKey": 1,
"AVNumberOfChannelsKey": 2,
"AVFormatIDKey": 1819304813,
"AVLinearPCMIsNonInterleaved": 0,
"AVLinearPCMIsBigEndianKey": 0]
endFormatSettings:
["AVSampleRateKey": 16000,
"AVLinearPCMBitDepthKey": 16,
"AVLinearPCMIsFloatKey": 1,
"AVNumberOfChannelsKey": 1,
"AVFormatIDKey": 1819304813,
"AVLinearPCMIsNonInterleaved": 0,
"AVLinearPCMIsBigEndianKey": 0,
"AVEncoderQualityKey": 64]
avconverterError: Error Domain=NSOSStatusErrorDomain Code=-50 "(null)"
I'm not 100% sure why this is the case, but I found a solution that got this working for me, so here's how I understand the problem. I found this solution by trying to use the alternate
convert(to:error:withInputFrom:)
method. Using this was giving me a different error:The problem was caused in the line where I setup the
AVAudioConverter
:It appears that the audio converter wants to use the same
AVAudioFormat
that the input buffer is using, instead of using a copy based on the original's settings. Once I swappedstartFormat
out forinAudioFormat
, theconvert(to:error:withInputFrom:)
error was dismissed, and things worked as expected. I was then able to go back to using the simplerconvert(to:fromBuffer:)
method, and the original error I was dealing with also went away.To recap, the line setting up the converter now looks like:
As for the lack of docs on how to use
AVAudioConverter
, I have no idea why the API reference has next to nothing. Instead, in Xcode, CMD-click onAVAudioConverter
in your code to go to it's header file. There's plenty of comments and info there. Not full sample code or anything, but it's at least something.