Reading a wavefile into a signed float array

Question

With the following code, I can read the a wave-file into a float array. However, the negative side of the wave ends up on the positive side so: -MAX -> 0 and "-0" -> MAX. I can't quite wrap my head around how to do this properly. I do realize the header portion of the wave may be more than 44 bytes long, but for this application it's not going to happen. I hope one of you fine folks can help me :-)

string filename;
getline(cin, filename);

ifstream wavFile(filename, ios::binary | ios::ate);

if (!wavFile) 
    cout << "error reading file\n";

int size = wavFile.tellg();
wavArray = new unsigned char[size];
wavFile.seekg(0, ios::beg);
wavFile.read((char *)wavArray, size);
wavFile.close();


int j = 1;
for (int i = 0; i < 262; i++)
{
    unsigned int tmp = (wavArray[i + 44] | (wavArray[i + 45] << 8));
    printf("%d : %f \n", k, (tmp / 32768.0));
    j++;
    i++;
}

Answer 1

The problem is your use of unsigned for the tmp variable. So you never have a negative value in tmp, even when the sample is negative. Try this

int16_t tmp = (wavArray[i + 44] | (wavArray[i + 45] << 8));

You should #include <cstdint> to get the int16_t type.