Reading .wav files - Not reading header?

Question

I am trying to read the header information from a .wav file, and I'm reading the following properties of the file:

    type 
    size
    Format Tag
    Format Length
    Channels
    Sample Rate
    Average Bytes Second
    Block Align
    Bits Per Sample
    Data size

Now, everything reads perfectly up until reading in Sample Rate, then, the output is either "0" value or just NULL.

Here is the code: (I have opened the text file, and passed the pointer to the member function)

bool Wav::readHeader(FILE *dataIn)
{
   char type2[4];
   fread(this->type, sizeof(char), 4, dataIn);

   fread(&this->size, sizeof(int), 1, dataIn);

   fread(this->type, sizeof(char), 4, dataIn);

   fread(this->type, sizeof(char), 4, dataIn);

   fread(&this->format_length, sizeof(short), 1, dataIn);

   fread(&this->format_tag, sizeof(short), 1, dataIn);

   fread(&this->channels, sizeof(short), 1, dataIn);


   fread(&this->sample_rate, sizeof(unsigned int), 1, dataIn);

   fread(&this->avg_bytes_sec, sizeof(unsigned short), 1, dataIn);

   fread(&this->block_align, sizeof(short), 1, dataIn);

   fread(&this->bits_per_sample, sizeof(short), 1, dataIn);

   return true;
}

And the class members are defined as:

    char type[4]; 
    int size;
    unsigned short format_tag;
    unsigned short format_length;
    unsigned short channels;
    unsigned int sample_rate;
    unsigned short avg_bytes_sec;
    unsigned short block_align;
    unsigned short bits_per_sample;
    unsigned int data_size;

Does anyone have any solutions or suggestions?

(P.S. Apologies if this question is not right. I have tried to write it the best I could!)

Is it a Windows question? If so, why not use proper API to separate headers from payload? — Roman Ryltsov
– Roman Ryltsov, Commented Aug 13, 2012 at 14:50
avg_bytes_sec is 32-bit value and not unsigned short you have, see http://msdn.microsoft.com/en-us/library/windows/desktop/dd390970%28v=vs.85%29.aspx — Roman Ryltsov
– Roman Ryltsov, Commented Aug 13, 2012 at 14:55
Your header fields look slightly off. Unless I missed something, you appear to be trying to read the samplerate from offset 22. The samplerate is located at offset 24, see ccrma.stanford.edu/courses/422/projects/WaveFormat . You might also want to use types with guaranteed sizes (int32_t and friends). — Michael
– Michael, Commented Aug 13, 2012 at 14:57
Aha thank you. Roman. So, when it says "WORD" on the url provided, do I assume this means char? — Phorce
– Phorce, Commented Aug 13, 2012 at 15:01

Cornstalks · Accepted Answer · 2012-08-13 21:51:20Z

2

If you had tightly packed structs representing the (PCM) WAVE's header, it would look like this:

struct Riff
{
  char chunkId[4]; // "RIFF" (assuming char is 8 bits)
  int chunkSize; // (assuming int is 32 bits)
  char format[4]; // "WAVE"
};

struct Format
{
  char chunkId[4]; // "fmt "
  int chunkSize;
  short format; // assuming short is 16 bits
  short numChannels;
  int sampleRate;
  int byteRate;
  short align;
  short bitsPerSample;
};

struct Data
{
  char chunkId[4]; // "data"
  int chunkSize; // length of data
  char* data;
};

struct Wave // Actual structure of a PCM WAVE file
{
  Riff riffHeader;
  Format formatHeader;
  Data dataHeader;
};

Essentially, your problem is that format_length (Format::chunkSize in my struct) and avg_bytes_sec (Format::byteRate in my struct) are 2 bytes each in your code, but they should each be 4. You're also not reading the data subchunk's header, but I assume that's intentional?

To read, you can do something like this:

void readWave(std::ifstream& file, Wave& wave)
{
  // First read the RIFF header
  file.read(wave.riffHeader.chunkId,                                   4);
  file.read(reinterpret_cast<char*>(&wave.riffHeader.chunkSize),       4);
  file.read(wave.riffHeader.format,                                    4);

  // Now read the FORMAT header
  file.read(wave.formatHeader.chunkId,                                 4);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.chunkSize),     4);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.format),        2);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.numChannels),   2);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.sampleRate),    4);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.byteRate),      4);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.align),         2);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.bitsPerSample), 2);

  // Now read the DATA header
  file.read(wave.dataHeader.chunkId,                                   4);
  file.read(reinterpret_cast<char*>(&wave.dataHeader.chunkSize),       4);

  // The actual data goes in wave.dataHeader.data, so you can allocate it
  // and then read direclty into it now
}

Michael's link is the one I always use when I need to read/write a WAVE file (which I've had to do several times). I suggest you read it carefully.

If you've got extra chunks in your WAVE file before your "data" chunk, this code should be able to skip them:

void readWave(std::ifstream& file, Wave& wave)
{
  // First read the RIFF header
  file.read(wave.riffHeader.chunkId,                                   4);
  file.read(reinterpret_cast<char*>(&wave.riffHeader.chunkSize),       4);
  file.read(wave.riffHeader.format,                                    4);

  // Now read the FORMAT header
  file.read(wave.formatHeader.chunkId,                                 4);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.chunkSize),     4);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.format),        2);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.numChannels),   2);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.sampleRate),    4);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.byteRate),      4);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.align),         2);
  file.read(reinterpret_cast<char*>(&wave.formatHeader.bitsPerSample), 2);

  // WAVE files are just a special type of RIFF file, so it's possible
  // there are other chunks, like "fact" chunks. We'll skip over these
  // extra chunks until we find a "data" chunk
  char chunkId[4] = {0};
  int chunkSize = 0;
  while (file.read(chunkId, 4) &&
         (chunkId[0] != 'd' ||
          chunkId[1] != 'a' ||
          chunkId[2] != 't' ||
          chunkId[3] != 'a'))
    {
      file.read(reinterpret_cast<char*>(&chunkSize), 4); // Read the chunk's size
      file.seekg(chunkSize, std::ios_base::cur); // Skip the chunk
    }

  // We've found the DATA chunk and header
  wave.dataHeader.chunkId[0] = chunkId[0];
  wave.dataHeader.chunkId[1] = chunkId[1];
  wave.dataHeader.chunkId[2] = chunkId[2];
  wave.dataHeader.chunkId[3] = chunkId[3];
  file.read(reinterpret_cast<char*>(&wave.dataHeader.chunkSize),       4);

  // The actual data goes in wave.dataHeader.data, so you can allocate it
  // and then read direclty into it now
}

edited Aug 13, 2012 at 21:51

answered Aug 13, 2012 at 16:22

Cornstalks

38.5k20 gold badges85 silver badges151 bronze badges

Sign up to request clarification or add additional context in comments.

21 Comments

Phorce Over a year ago

Hello, so can I use this type of structure inside a class? Or, could I possibly have different methods "riff", "Format", "Data" that passes an ifstream to get data? -- Confused

Cornstalks Over a year ago

Sure you could. I'm not sure I entirely understand what it is you're asking though...

Cornstalks Over a year ago

I just added a method about reading. Hopefully it answers your question.

Phorce Over a year ago

Hello, thank you :D it worked.. I think!! Quick question.. If I had data = new char[SIZE_OF_DATA] then can I just read it in.. E.g. file.read(this->data, this->dataSize); ??

Phorce Over a year ago

For example.. What should the data type be for the actual data? When I have it as char, I seem to just get "? #" or something.. Also, with a 41KB file, and, after the header.. Should the data size by 4?

|

Collectives™ on Stack Overflow

Reading .wav files - Not reading header?

1 Answer 1

21 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

21 Comments

Your Answer

Sign up or log in

Post as a guest

Related