WAV files at any rate except 44.1kHz have messed up sound

Question

I'm using ALSA on Ubuntu to try to play a WAV file. Presently, I'm able to read the wav header to figure out the file's sampling rate and such then set the parameters on ALSA to correspond. This works perfectly for files with a 44.1kHz sampling rate, but other files with rates at ~11kHz or ~22kHz do not play correctly. I'm not sure that I am setting the sampling rate correctly.

val = realSampleRate;
    //Sampling rate to given sampling rate
    snd_pcm_hw_params_set_rate_max(handle, params, &val, &dir);
    cout << "sampling at " << val << " Hz \n";

This gives the correct output ("sampling at 22050 Hz") but if I follow it with this:

val = realSampleRate;
snd_pcm_hw_params_set_rate_min(handle, params, &val, &dir);
cout << "sampling at " << val << " Hz \n";

the output proceeds to say "sampling at 44100 Hz" which is obviously contradictory. I also tried using snd_pcm_hw_params_set_rate_near but that doesn't work either, it says sampling at 44100 Hz on a 22050 file, and the audio throughout all of those were very messed up.

EDIT: One issue is incorrect sampling rates, which will speed up the playing, but the real issue comes from mono tracks. Mono tracks sound really distorted and very off.

EDIT: 8 Bit files are off too

Answer 1

Looks to me like your hardware is not capable of handling a 22.05Khz sampling rate for playback. The fact that the API function returns a different value is a clue.

ALSA is just an API. It can only do what your current underlying hardware is capable of supporting. Low-end, bottom-of-the-barrel, el-cheapo audio playback hardware will support a handful of sampling frequencies, and that's about it.

I had some custom-written audio recording and playback software, that was sampling and recording audio at a particular rate, then playing it back using ALSA's aplay. When I got some new hardware, I found that the new hardware was still capable of supporting my sampling rate for recording, for playback it didn't, and aplay simply proceeded to play back the previously recorded audio at the nearest supportable playback level, with hillarious results. I had to change my custom-written stuff to record and playback at the supported rate.

If the hardware does not support your requested playback rate, ALSA won't resample your raw audio data. It's up to you to resample it, for playback.

Answer 2

snd_pcm_hw_params_set_rate_max() sets the maximum sample rate, i.e., when this functions succeeds, the device's sample rate will not be larger than what you've specified.

snd_pcm_hw_params_set_rate_min() sets the minimum sample rate.

snd_pcm_hw_params_set_rate_near() searches for the nearest sample rate that is actually supported by the device, sets it, and returns it.

If you have audio data with a specific sample rate, and cannot do resampling, you must use snd_pcm_hw_params_set_rate().

Answer 3

Using "default" instead of "hw:0,0" solves this, including the sampling rate being too slow. "plughw:0,0" works as well, and it's better because you can select the different devices/cards programmatically whereas default just uses the default.