sd mp3 to i2s: When playing a sound multiple times it plays the ending of whatever was played previously at the start.

[LawrencePritchard]

Hi,

I'm making some electronic air drums for my university project. An esp32 connected to some drumsticks (with an accelerometer and gyroscope) sends data to another esp32 to turn it into sound. For example, if this is sent:
myData.a = 'k', myData.b = 0.3;
It needs to be turned into a kickdrum at 0.3 volume.

My issue is that when I play a sound from the sd card it plays the ending of the previous sound before playing the current sound. The first time a sound is played this doesn't happen. Even if I only try and repeat the same sound it still plays the ending of the sound at the start. (I wonder if the buffer isn't fully cleared each time. Is there a way to clear the buffer?)

here is my implementation:

#include <SPI.h>
#include <SD.h>
#include "AudioTools.h"
#include "AudioTools/AudioCodecs/CodecMP3Helix.h"
#include <esp_now.h>
#include <WiFi.h>


// Structure example to receive data
// Must match the sender structure
typedef struct struct_message {
    char a;
    float b;
} struct_message;

// Create a struct_message called myData
struct_message myData;
// callback function that will be executed when data is received

int sck = 7;
int miso = 8;
int mosi = 6;
int cs = 14;
I2SStream i2s; // final output of decoded stream
VolumeStream volume(i2s);
EncodedAudioStream decoder(&volume, new MP3DecoderHelix()); // Decoding stream

StreamCopy copier; 
File kickFile;
File crashFile;
File snareFile;

// receiving data via espnow 
void OnDataRecv(const uint8_t * mac, const uint8_t *incomingData, int len) {
  memcpy(&myData, incomingData, sizeof(myData));
  
  Serial.println(String(myData.a));
  Serial.println();

  volume.setVolume(myData.b); // set the volume

 if (myData.a == 'k') { 
        Serial.println("kick:");
        kickFile.seek(0);
        kickFile = SD.open("/kickableton.mp3");
        copier.begin(decoder, kickFile);

      } else if (myData.a == 's') {
        Serial.println("snare");
        snareFile = SD.open("/snare_ableton.mp3");
        snareFile.seek(0);
         copier.begin(decoder, snareFile);
      }

}
 
void setup(){
  Serial.begin(115200);
  AudioToolsLogger.begin(Serial, AudioToolsLogLevel::Info);  

  // setup file
  SPI.begin(sck, miso, mosi, cs);
  SD.begin(cs);

  // setup i2s
  auto config = i2s.defaultConfig(TX_MODE);
  config.pin_ws = 9;//15 on ESP32 FEATHER a1
  config.pin_data = 4;//32 on ESP32 FEATHER a0
  config.pin_bck = 5;//14 on ESP32 FEATHER a2
  i2s.begin(config);

  // setup I2S based on sampling rate provided by decoder
  decoder.begin();

   // set initial volume
  auto vcfg = volume.defaultConfig();
  vcfg.copyFrom(config);
  //vcfg.allow_boost = true; // activate amplification using linear control
  volume.begin(vcfg); // we need to provide the bits_per_sample and channels

  // begin copy
  // copier.begin(decoder, audioFile);
kickFile = SD.open("/kickableton.mp3");
snareFile = SD.open("/snare_ableton.mp3");
crashFile = SD.open("/Crash_1.mp3");

  // Set device as a Wi-Fi Station
  WiFi.mode(WIFI_STA);

  // Init ESP-NOW
  if (esp_now_init() != ESP_OK) {
    Serial.println("Error initializing ESP-NOW");
    return;
  }
  // Once ESPNow is successfully Init, we will register for recv CB to
  // get recv packer info
  esp_now_register_recv_cb(esp_now_recv_cb_t(OnDataRecv));
}

void loop(){
 copier.copy();
}

[pschatzmann]

You forgot to call end() and begin() on the decoder before you start a new mp3 file.
I guess you could avoid this by just using pcm data.

You could also think about sending just some silence data to i2s if you don't provide any audio data...

[LawrencePritchard]

Thank you it works!

[pschatzmann]

Don't forget to close the file before reopening it again: otherwise you create a memory leak!
Alternatively you could just reposition to the beginning with lseek(0)