Implement voice-processing

[JayFoxRox]

Pseudocode:

      
    // Create a buffer for this voice
    int32_t voice_samples[NUM_SAMPLES_PER_FRAME];
    int voice_samples_left = ARRAY_SIZE(voice_samples);

    // Figure out how many samples we'll require from the input, to fill this
    // frame.
    int voice_buffer_sample_count = 1337; //FIXME!

    // Get the samples
    #define MAX_VOICE_SAMPLES 100
    int32_t voice_buffer_samples[MAX_VOICE_SAMPLES];

    // Decode some input samples for this voice
    input_sample_count = decode(voice_buffer_raw_samples, voice_buffer_samples, voice_buffer_sample_count);

    // Resample voice input
    while(voice_samples_left > 0) {

      // This needs to play back at 48kHz
      resample(f, 48000, &input_samples, &input_sample_count, &voice_samples, &voice_samples_left);

      // Handle end of input buffer
      if (input_sample_count == 0) {
        if (voice_loops) {
          // Voice loops; go back to start of buffer
          input_samples = voice_buffer_samples;
          input_sample_count = voice_buffer_sample_count;
        } else {
          // Voice finishes; fill with silence
          while(voice_samples_left--) {
            *voice_samples++ = 0;
          }
        }
      }

    }
#if 0
    //FIXME: Generate envelope for this voice
    uint8_t amplitude_level = get_envelope_level(voice);

    //FIXME: This buffer generation is not necessary.
    //       The envelope counter runs at 1500 Hz = 1 update per frame.
    //       We simply have to figure out if there's linear interpolation.
    //       Eitherway: We can apply it in the mixing step below
    uint8_t amplitude_envelope[NUM_SAMPLES_PER_FRAME];
    memset(amplitude_envelope, amplitude_level, sizeof(amplitude_envelope));

    // Apply amplitude envelopes
    multiply(voice_samples, amplitude_envelope, NUM_SAMPLES_PER_FRAME);
#endif

    // Add the output to each of the MIXBINs for this voice
    for(int i = 0; i < 8; i++) {
      mix(mixbin[voice_mixbin_target[i]], NUM_SAMPLES_PER_FRAME, voice_samples, ARRAY_SIZE(voice_samples), voice_mixbin_volume[i]);
    }