CODEWITHSUNDEEP
caudiostm32wav
Antoine
Antoine

Reputation: 3

C / STM32 - Read and copy a data from .wav file

I am trying to copy a .wav file from my flash RAM memory.

#define AUDIO_BUFFER_SIZE          (1024 * 8)       /* Size (in bytes) of the buffer containing the PCM samples */

uint8_t                  Buffer[AUDIO_BUFFER_SIZE];    // Buffer containig the PCM samples to play

...

      /* Fill in the buffer with new data */
  if (f_read(&File, (uint8_t   *)Buffer, AUDIO_BUFFER_SIZE, &bytesRead) != FR_OK)
  {
    Error_Handler();
  }

  if (counter==1){
      HAL_GPIO_WritePin(LED_RED_GPIO_Port, LED_RED_Pin, GPIO_PIN_SET);
      //uint8_t string[20]="Hello, world!";
      //f_write(&OutFile, Buffer, (UINT)sizeof(Buffer),&bytesRead);
      for(int i = 0; i <= sizeof(Buffer); i++){
          f_printf(&OutFile, "%d\n",Buffer[i]);
          osDelay(10);
      }
      counter++;
  }
  else{
      HAL_GPIO_WritePin(LED_RED_GPIO_Port, LED_RED_Pin, GPIO_PIN_RESET);
      f_close(&OutFile);
  }

When i do this, I get a file with some values like that (right part of this screenshot) Output file

How can i get the correct values as we can see them on the left part of my screenshot ?

Regards

Upvotes: 0

Views: 984

Answers (1)

MikeCAT
MikeCAT

Reputation: 75062

Based on the values, your samples looks like being encoded in signed 16-bit little endian format.

To decode the format, you can do like this (assuming that the format specifier of f_printf is just like the standard printf):

// 2 bytes per sample, also use < instead of <=
for(int i = 0; i < sizeof(Buffer); i += 2){
    int value = Buffer[i] | (Buffer[i + 1] << 8); // merge the 2 bytes into one integer
    if (value >= 0x8000) value -= 0x10000; // because the samples are signed
    f_printf(&OutFile, "%.4f\n", value / (double)0x8000); // divide with the maximum value
    osDelay(10);
}

If you cannot print floating-point number via f_printf, you can print with rounding by doing:

  1. Multiply 10000 because there are 4 digits after the decimal point
  2. Multiply 2 and add or subtract 0x8000 (the value used for division) based of the sign of the value
  3. Divide by 0x8000 * 2
  4. Print the value before and after the decimal point
// 2 bytes per sample, also use < instead of <=
for(int i = 0; i < sizeof(Buffer); i += 2){
    int v;
    int value = Buffer[i] | (Buffer[i + 1] << 8); // merge the 2 bytes into one integer
    if (value >= 0x8000) value -= 0x10000; // because the samples are signed
    // divide with the maximum value
    v = ((value * 10000) * 2 + (value >= 0 ? 0x8000 : -0x8000)) / (0x8000 * 2);
    f_printf(&OutFile, "%s%d.%04d\n",
      v < 0 && v / 10000 == 0 ? "-" : "", // sign (because typical integers don't have -0)
      v / 10000, // value before the decimal point
      (v < 0 ? -v : v) % 10000); // value after the decimal point
    osDelay(10);
}

Upvotes: 1

Related Questions