如何在iOS中添加帶有PCM數據/緩沖區的可播放（如wav，wmv）header？

Question

我正在嘗試在原始 PCM 數據之上添加一個 wav header 以使其可通過 AVAudioPlayer 播放。 但是我找不到任何解決方案或源代碼來使用 Objective-C/Swift 在 iOS 上執行此操作。 雖然我找到了這個，但沒有正確答案。

但我在這里找到了一段代碼，它位於 C 中，也包含一些問題。 從該代碼生成的 wav 文件無法正常播放。

我已經在下面給出了我到目前為止編碼的代碼。

int NumChannels = AUDIO_CHANNELS_PER_FRAME;
short BitsPerSample = AUDIO_BITS_PER_CHANNEL;
int SamplingRate = AUDIO_SAMPLE_RATE;
int numOfSamples = [[NSData dataWithContentsOfFile:filePath] length];

int ByteRate = NumChannels*BitsPerSample*SamplingRate/8;
short BlockAlign = NumChannels*BitsPerSample/8;
int DataSize = NumChannels*numOfSamples*BitsPerSample/8;
int chunkSize = 16;
int totalSize = 36 + DataSize;
short audioFormat = 1;

if((fout = fopen([wavFilePath cStringUsingEncoding:1], "w")) == NULL)
{
    printf("Error opening out file ");
}

fwrite("RIFF", sizeof(char), 4,fout);
fwrite(&totalSize, sizeof(int), 1, fout);
fwrite("WAVE", sizeof(char), 4, fout);
fwrite("fmt ", sizeof(char), 3, fout);
fwrite(&chunkSize, sizeof(int),1,fout);
fwrite(&audioFormat, sizeof(short), 1, fout);
fwrite(&NumChannels, sizeof(short),1,fout);
fwrite(&SamplingRate, sizeof(int), 1, fout);
fwrite(&ByteRate, sizeof(int), 1, fout);
fwrite(&BlockAlign, sizeof(short), 1, fout);
fwrite(&BitsPerSample, sizeof(short), 1, fout);
fwrite("data", sizeof(char), 3, fout);
fwrite(&DataSize, sizeof(int), 1, fout);

文件播放速度太快，聲音失真，只播放前 10 到 20（大約）秒。 我認為，wav header 沒有正確生成（因為我能夠使用 AudioUnit/AudioQueue 播放相同的 PCM 數據/緩沖區）。 那么我的代碼中缺少什么？ 任何幫助將不勝感激。

提前致謝。

Answer 1

好的，如果它對其他人有幫助，我正在回答我自己的問題。 經過幾天的不懈努力，我終於讓它工作了。 下面是一個用 Objective-C 和 C 編寫的完整函數。它以文件路徑作為參數，其中包含直接從麥克風捕獲的 RAW PCM 數據，並返回一個文件路徑，其中包含 PCM 數據，后跟適當的 wav 頭信息。 然后您可以使用 AVAudioPlayer 或 AVPlayer 播放該文件。 這是代碼...

- (NSURL *) getAndCreatePlayableFileFromPcmData:(NSString *)filePath
{
    NSString *wavFileName = [[filePath lastPathComponent] stringByDeletingPathExtension];
    NSString *wavFileFullName = [NSString stringWithFormat:@"%@.wav",wavFileName];

    [self createFileWithName:wavFileFullName];
    NSArray *dirPaths = NSSearchPathForDirectoriesInDomains(NSDocumentDirectory, NSUserDomainMask, YES);
    NSString *docsDir = [dirPaths objectAtIndex:0];
    NSString *wavFilePath = [docsDir stringByAppendingPathComponent:wavFileFullName];

    NSLog(@"PCM file path : %@",filePath);

    FILE *fout;

    short NumChannels = AUDIO_CHANNELS_PER_FRAME;
    short BitsPerSample = AUDIO_BITS_PER_CHANNEL;
    int SamplingRate = AUDIO_SAMPLE_RATE;
    int numOfSamples = [[NSData dataWithContentsOfFile:filePath] length];

    int ByteRate = NumChannels*BitsPerSample*SamplingRate/8;
    short BlockAlign = NumChannels*BitsPerSample/8;
    int DataSize = NumChannels*numOfSamples*BitsPerSample/8;
    int chunkSize = 16;
    int totalSize = 46 + DataSize;
    short audioFormat = 1;

    if((fout = fopen([wavFilePath cStringUsingEncoding:1], "w")) == NULL)
    {
        printf("Error opening out file ");
    }

    fwrite("RIFF", sizeof(char), 4,fout);
    fwrite(&totalSize, sizeof(int), 1, fout);
    fwrite("WAVE", sizeof(char), 4, fout);
    fwrite("fmt ", sizeof(char), 4, fout);
    fwrite(&chunkSize, sizeof(int),1,fout);
    fwrite(&audioFormat, sizeof(short), 1, fout);
    fwrite(&NumChannels, sizeof(short),1,fout);
    fwrite(&SamplingRate, sizeof(int), 1, fout);
    fwrite(&ByteRate, sizeof(int), 1, fout);
    fwrite(&BlockAlign, sizeof(short), 1, fout);
    fwrite(&BitsPerSample, sizeof(short), 1, fout);
    fwrite("data", sizeof(char), 4, fout);
    fwrite(&DataSize, sizeof(int), 1, fout);

    fclose(fout);

    NSMutableData *pamdata = [NSMutableData dataWithContentsOfFile:filePath];
    NSFileHandle *handle;
    handle = [NSFileHandle fileHandleForUpdatingAtPath:wavFilePath];
    [handle seekToEndOfFile];
    [handle writeData:pamdata];
    [handle closeFile];

    return [NSURL URLWithString:wavFilePath];
}

但該功能僅適用於以下音頻設置。

// Audio settings.
#define AUDIO_SAMPLE_RATE 8000
#define AUDIO_FRAMES_PER_PACKET 1
#define AUDIO_CHANNELS_PER_FRAME 1
#define AUDIO_BITS_PER_CHANNEL 16
#define AUDIO_BYTES_PER_PACKET 2
#define AUDIO_BYTES_PER_FRAME 2

Answer 2

非常有幫助的問題和答案，非常感謝。

這個 swift 版本適用於有需要的人：

static func createWAV(from pcmFilePath: String, to wavFilePath: String) -> Bool {

    // Make sure that the path does not contain non-ascii characters
    guard let fout = fopen(wavFilePath.cString(using: .ascii), "w") else { return false }

    guard let pcmData = try? Data(contentsOf: URL(fileURLWithPath: pcmFilePath)) else { return false }

    var numChannels: CShort = 1
    let numChannelsInt: CInt = 1
    var bitsPerSample: CShort = 16
    let bitsPerSampleInt: CInt = 16
    var samplingRate: CInt = 16000
    let numOfSamples = CInt(pcmData.count)
    var byteRate = numChannelsInt * bitsPerSampleInt * samplingRate / 8
    var blockAlign = numChannelsInt * bitsPerSampleInt / 8
    var dataSize = numChannelsInt * numOfSamples * bitsPerSampleInt / 8
    var chunkSize: CInt = 16
    var totalSize = 46 + dataSize
    var audioFormat: CShort = 1

    fwrite("RIFF".cString(using: .ascii), MemoryLayout<CChar>.size, 4, fout)
    fwrite(&totalSize, MemoryLayout<CInt>.size, 1, fout)
    fwrite("WAVE".cString(using: .ascii), MemoryLayout<CChar>.size, 4, fout);
    fwrite("fmt ".cString(using: .ascii), MemoryLayout<CChar>.size, 4, fout);
    fwrite(&chunkSize, MemoryLayout<CInt>.size,1,fout);
    fwrite(&audioFormat, MemoryLayout<CShort>.size, 1, fout);
    fwrite(&numChannels, MemoryLayout<CShort>.size,1,fout);
    fwrite(&samplingRate, MemoryLayout<CInt>.size, 1, fout);
    fwrite(&byteRate, MemoryLayout<CInt>.size, 1, fout);
    fwrite(&blockAlign, MemoryLayout<CShort>.size, 1, fout);
    fwrite(&bitsPerSample, MemoryLayout<CShort>.size, 1, fout);
    fwrite("data".cString(using: .ascii), MemoryLayout<CChar>.size, 4, fout);
    fwrite(&dataSize, MemoryLayout<CInt>.size, 1, fout);
    fclose(fout);

    guard let handle = FileHandle(forUpdatingAtPath: wavFilePath) else { return false }

    handle.seekToEndOfFile()
    handle.write(pcmData)
    handle.closeFile()

    return true
}

Answer 3

修改自 qiz 對 swift 5 的回答

func extractSubchunks(data:Data) -> RiffFile? {
    var data = data
    var chunks = [SubChunk]()
    let position = data.subdata(in: 8..<12)
    let filelengthBytes = data.subdata(in: 4..<8).map { UInt32($0) }
    let filelength: UInt32 = filelengthBytes[0] << 24 + filelengthBytes[1] << 16 + filelengthBytes[2] << 8 + filelengthBytes[3]
    let wave = String(bytes: position, encoding: .utf8) ?? "NoName"
    guard wave == "WAVE" else {
        print("File is \(wave) not WAVE")
        return nil
    }
    data.removeSubrange(0..<12)
    print("Found chunks")
    while data.count != 0{
        let position = data.subdata(in: 0..<4)
        let lengthBytes = data.subdata(in: 4..<8).map { UInt32($0) }
        let length: UInt32 = lengthBytes[0] << 24 + lengthBytes[1] << 16 + lengthBytes[2] << 8 + lengthBytes[3]
        guard let current = String(bytes: position, encoding: .utf8) else{
            return nil
        }
        data.removeSubrange(0..<8)
        let chunkData = data.subdata(in: 0..<Int(length))
        data.removeSubrange(0..<Int(length))
        let subchunk = SubChunk(name: current, size: Int(length), data: chunkData)
        chunks.append(subchunk)
        print(subchunk.debugDescription)
    }
    let riff = RiffFile(size: Int(filelength), subChunks: chunks)
    return riff
}

Answer 4

這是 Swift 的數據擴展，返回另一個數據，使用qiz 的答案制作。

extension Data {
    func wavValue: Data? {
        var numChannels: CShort = 1
        let numChannelsInt: CInt = 1
        var bitsPerSample: CShort = 16
        let bitsPerSampleInt: CInt = 16
        var samplingRate: CInt = 44100
        let numOfSamples = CInt(pcmData.count)
        var byteRate = numChannelsInt * bitsPerSampleInt * samplingRate / 8
        var blockAlign = numChannelsInt * bitsPerSampleInt / 8
        var dataSize = numChannelsInt * numOfSamples * bitsPerSampleInt / 8
        var chunkSize: CInt = 16
        var totalSize = 46 + dataSize
        var audioFormat: CShort = 1

        let wavNSData = NSMutableData()
        wavNSData.append("RIFF".cString(using: .ascii) ?? .init(), length: MemoryLayout<CChar>.size * 4)
        wavNSData.append(&totalSize, length: MemoryLayout<CInt>.size)
        wavNSData.append("WAVE".cString(using: .ascii) ?? .init(), length: MemoryLayout<CChar>.size * 4)
        wavNSData.append("fmt ".cString(using: .ascii) ?? .init(), length: MemoryLayout<CChar>.size * 4)
        wavNSData.append(&chunkSize, length: MemoryLayout<CInt>.size)
        wavNSData.append(&audioFormat, length: MemoryLayout<CShort>.size)
        wavNSData.append(&numChannels, length: MemoryLayout<CShort>.size)
        wavNSData.append(&samplingRate, length: MemoryLayout<CInt>.size)
        wavNSData.append(&byteRate, length: MemoryLayout<CInt>.size)
        wavNSData.append(&blockAlign, length: MemoryLayout<CShort>.size)
        wavNSData.append(&bitsPerSample, length: MemoryLayout<CShort>.size)
        wavNSData.append("data".cString(using: .ascii) ?? .init(), length: MemoryLayout<CChar>.size * 4)
        wavNSData.append(&dataSize, length: MemoryLayout<CInt>.size)

        wavNSData.append(self)

        let wavData = Data(referencing: wavNSData)

        return wavData
    } 
}