Audio Subsystem Demo Code

[Dominicc1]

Audio Demo Features + Documentation for Audio Subsystem !!!

[Dominicc1]

Work Flow :

ESP-32 ( I2S, Wifi )
SPH 0645 ( I2S input, Data From Mic -> ESP 32 )
DF Robot Max 98357A ( I2S output, ESP 32 -> Speaker )
External RF Module ( Convert baseband audio data to RF signal and vice versa )

Transmit ( TX )

Sources modulated audio to integrated radio transceiver

1. Audio Sourcing of Human Voice (MEMS Mic)
2. Audio Processing into 'analog samples' understood by Write CC1200 Device Driver #81 (ESP32)
3. Transmit Human Voice Write RF Subsystem Demo Code #82

Receive (RX)

Sinks modulated audio from integrated radio transceiver

1. Taking Audio Analog Sample from RF Module Write CC1200 Device Driver #81
2. Audio Processing into I2S audio stream ( ESP32 )
3. Audio Sinking (MAX98357A)

[Dominicc1]

#include <Arduino.h>
#include "driver/i2s.h"

// ===== Pin Configuration =====
// I2S CLOCKS (shared by mic + amp)
#define I2S_BCLK 19 // BLCK of SPH and AMP share the same clock
#define I2S_LRCLK 18 // Left Right of SPH and AMP share the same clock

// MIC (SPH0645)
#define I2S_DIN 16 // Mic SPH 0645 Data to ESP32

// AMP (MAX98357A)
#define I2S_DOUT 21 // Amp data from ESP32
#define AMP_SD 17 // Shutdown pin for amplifier

#define SAMPLE_RATE 16000 // SPh works great on 8 KHz to 48K khz
#define BUFFER_SAMPLES 128 // indicating how many sample of buffer will be read and write during I2S cycle , 64-512 ( the bigger number the bigger latency, lower can be affected to CPU)

int16_t buffer[BUFFER_SAMPLES];

void setup() {
Serial.begin(115200);
delay(500);
Serial.println("ESP32 I2S Mic → Amp Pass-Through Starting...");

// ===== Enable amplifier =====
pinMode(AMP_SD, OUTPUT);
digitalWrite(AMP_SD, HIGH); // Amplifier ON

// ===== I2S Configuration =====
i2s_config_t i2s_config = {
.mode = (i2s_mode_t)(I2S_MODE_MASTER | I2S_MODE_RX | I2S_MODE_TX),
.sample_rate = SAMPLE_RATE,
.bits_per_sample = I2S_BITS_PER_SAMPLE_32BIT,
.channel_format = I2S_CHANNEL_FMT_ONLY_LEFT,
.communication_format = I2S_COMM_FORMAT_I2S_MSB,
.intr_alloc_flags = ESP_INTR_FLAG_LEVEL1,
.dma_buf_count = 8,
.dma_buf_len = BUFFER_SAMPLES,
.use_apll = false,
.tx_desc_auto_clear = true,
.fixed_mclk = -1
};

// ===== I2S Pin Mapping =====
i2s_pin_config_t pins = {
.bck_io_num = I2S_BCLK,
.ws_io_num = I2S_LRCLK,
.data_out_num = I2S_DOUT, // Data to be used to transfer to MAX98357A
.data_in_num = I2S_DIN // Data to be used voice then transfer to ESP32
};

// Install and start I2S
i2s_driver_install(I2S_NUM_0, &i2s_config, 0, NULL); //install I2S for audio
i2s_set_pin(I2S_NUM_0, &pins);

Serial.println("I2S Initialized. Listening...");
}

void loop() {
int32_t micBuffer[BUFFER_SAMPLES];
size_t bytes_read = 0, bytes_written = 0;

// === Read from microphone (32-bit samples) ===
i2s_read(I2S_NUM_0, micBuffer, sizeof(micBuffer), &bytes_read, portMAX_DELAY); // // I2S Sample get from to SPH 0645 and it is 32 bits, MAX_Delay Function is to wait until all the data is fully receive otherwise partial or empty data will be trasnfer to MAX 98357A
int samples_read = bytes_read / sizeof(int32_t);

long levelSum = 0;

// === Convert 32-bit mic data → 16-bit audio for amp === because AMP only accept 16 bit PCM value but intake is 32 bit if keeping output garbage
for (int i = 0; i < samples_read; i++) {
int32_t sample = micBuffer[i] >> 18; // Reduce amplitude (less feedback) // used to 14 and 16 but i am trying ti increase to 18 to add stronger anntuaator ( MAX 98357 A only read 16 bit if left justified ), or maybe ( add more gain reduction, filtering, or physical separation ), as Dr, Tacca recommendation sine it is single duplex Larsen noise should not be a big problem, also I did try move the speaker 1 meter away from the mic but nothing much changes
// 18 stronger antenuation
// 14-16 louder output, the more the gain reduction

// Clamp to 16-bit range trying to keep it in 16 bit range ( which is -32768<x< 32767)
if (sample > 32767) sample = 32767;
if (sample < -32768) sample = -32768;

buffer[i] = (int16_t)sample;
levelSum += abs(buffer[i]);

}

// === Send audio to MAX98357A amp ===
i2s_write(I2S_NUM_0, buffer, samples_read * sizeof(int16_t), &bytes_written, portMAX_DELAY); // I2S Sample sending from to Max 98357A to amplify, MAX_Delay Function is to wait until all the data is fully receive otherwise partial or empty data will be trasnfer to MAX 98357A

// Debug volume monitor
Serial.println(levelSum / samples_read);
}