Smart Intercom using ESP32

[Philomath88]

Hey guys,

am trying to make my STR 200 TV intercom smart.

I am currently working on picking up the microphone audio (which is the easiest since I don't need another person at the door).

The output is 4.3V so I build a voltage divider circuit using resistors (the probe divides by 10):

This is in theory my script to output the audio signal to a web browser (just for testing): https://youtu.be/rurvSKYao3U

`

#include "AudioTools.h"
#include <WiFi.h>
#include "driver/i2s.h"
#include "driver/adc.h"

// WIFI
const char* ssid = "gigacube-46F37C";
const char* password = "7fLf2t5t5j233839";

// Static IP configuration
IPAddress staticIP(192, 168, 1, 200); // The static IP you want to assign to your ESP32
IPAddress gateway(192, 168, 1, 1); // Your router's IP address
IPAddress subnet(255, 255, 255, 0); // Subnet mask
IPAddress dns(8, 8, 8, 8); // DNS (you can use Google's public DNS)

AudioWAVServer server(ssid, password);

// ADC and I2S Configuration
#define SAMPLE_RATE 16000 // 16kHz
#define I2S_NUM I2S_NUM_0
#define DMA_BUF_COUNT 8
#define DMA_BUF_LEN 512

class ADCStream : public AudioStream {
public:
ADCStream() : AudioStream() {
// I2S configuration
i2s_config_t i2s_config = {
.mode = (i2s_mode_t)(I2S_MODE_MASTER | I2S_MODE_RX | I2S_MODE_ADC_BUILT_IN),
.sample_rate = SAMPLE_RATE,
.bits_per_sample = I2S_BITS_PER_SAMPLE_16BIT,
.channel_format = I2S_CHANNEL_FMT_ONLY_RIGHT,
.communication_format = I2S_COMM_FORMAT_I2S_MSB,
.intr_alloc_flags = ESP_INTR_FLAG_LEVEL1,
.dma_buf_count = DMA_BUF_COUNT,
.dma_buf_len = DMA_BUF_LEN,
.use_apll = false,
.tx_desc_auto_clear = false,
.fixed_mclk = 0
};

    // Install and start I2S driver
    if (i2s_driver_install(I2S_NUM, &i2s_config, 0, NULL) != ESP_OK) {
        Serial.println("Failed to install I2S driver");
        return;
    }

    // Set I2S ADC mode
    if (i2s_set_adc_mode(ADC_UNIT_1, ADC1_CHANNEL_4) != ESP_OK) {
        Serial.println("Failed to set I2S ADC mode");
        return;
    }

    // Enable ADC
    if (i2s_adc_enable(I2S_NUM) != ESP_OK) {
        Serial.println("Failed to enable I2S ADC");
        return;
    }
}

size_t readBytes(uint8_t *data, size_t len) override {
    size_t bytes_read;
    if (i2s_read(I2S_NUM, data, len, &bytes_read, portMAX_DELAY) == ESP_OK) {
        return bytes_read;
    }
    return 0;
}

};

ADCStream adcStream;

void setup() {
Serial.begin(115200);
AudioLogger::instance().begin(Serial, AudioLogger::Info);

// Configure static IP
if (!WiFi.config(staticIP, gateway, subnet, dns)) {
    Serial.println("STA Failed to configure");
}

// Connect to Wi-Fi
WiFi.begin(ssid, password);
while (WiFi.status() != WL_CONNECTED) {
    delay(1000);
    Serial.println("Connecting to WiFi...");
}

// Print ESP32 IP address
Serial.println("Connected to WiFi");
Serial.print("IP address: ");
Serial.println(WiFi.localIP());

// start server
server.begin(adcStream, SAMPLE_RATE, 1);  // 1 channel

Serial.println("ADC stream server started");

}

void loop() {
server.copy();
}
`

Unfortunately, I don't hear anything. I tried outputting the ADC values and if I speak, the go up to a max of 300.

Does anyone have an idea what I am doing wrong?

[pschatzmann]

Why adding all this complexity ? Just use an AnalogAudioStream in input mode!
Here is an example that reads from an analog mocrophone...

ps. your code will not work anymore if you upgrade to an actual ESP32 core version because the I2S API has changed!

[Philomath88]

@pschatzmann Thank you for the fast reply! What code should I use? I am using an ESP32-WROOM-32 board.

The ultimate goal is to send the audio data via a web socket to my server

[Philomath88]

I have updated my code to this:

`#include "Arduino.h"
#include "AudioTools.h"
#include <WiFi.h>

// WIFI credentials
const char* ssid = "gigacube-46F37C";
const char* password = "7fLf2t5t5j233839";

// Static IP configuration
IPAddress staticIP(192, 168, 1, 200);
IPAddress gateway(192, 168, 1, 1);
IPAddress subnet(255, 255, 255, 0);
IPAddress dns(8, 8, 8, 8);

// TCP settings
WiFiServer server(8888);
WiFiClient client;

AnalogAudioStream in;
AudioInfo info(8000, 1, 16);

unsigned long sampleCount = 0;

void setup() {
Serial.begin(115200);

auto cfgRx = in.defaultConfig(RX_MODE);
cfgRx.adc_pin = 32;
cfgRx.copyFrom(info);
in.begin(cfgRx);

// Configure static IP
if (!WiFi.config(staticIP, gateway, subnet, dns)) {
    Serial.println("STA Failed to configure");
}

// Connect to WiFi
WiFi.begin(ssid, password);
while (WiFi.status() != WL_CONNECTED) {
    delay(1000);
    Serial.println("Connecting to WiFi...");
}
Serial.println("Connected to WiFi");
Serial.print("IP address: ");
Serial.println(WiFi.localIP());

server.begin();
Serial.println("TCP server started");

}

void loop() {
if (!client.connected()) {
client = server.available();
if (client) {
Serial.println("New client connected");
}
}

if (client.connected()) {
    int16_t sample;
    if (in.readBytes((uint8_t*)&sample, sizeof(sample)) == sizeof(sample)) {
        char buffer[50];
        snprintf(buffer, sizeof(buffer), "%lu,%d\n", sampleCount, sample);
        client.print(buffer);
        sampleCount++;
    }
}

// Small delay to prevent overwhelming the system
delay(1);

}`

And am using this python script to save the values:

import socket
import csv
from datetime import datetime

TCP_IP = "192.168.1.200" # ESP32's IP address
TCP_PORT = 8888

Generate a unique filename based on current timestamp

filename = f"audio_data_{datetime.now().strftime('%Y%m%d_%H%M%S')}.csv"

with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
s.connect((TCP_IP, TCP_PORT))
print(f"Connected to {TCP_IP}:{TCP_PORT}")

with open(filename, 'w', newline='') as csvfile:
    csvwriter = csv.writer(csvfile)
    csvwriter.writerow(['Sample', 'Value'])  # Write header
    
    try:
        buffer = ""
        while True:
            data = s.recv(1024).decode('ascii')
            buffer += data
            lines = buffer.split('\n')
            for line in lines[:-1]:
                sample, value = line.strip().split(',')
                print(f"Received: Sample {sample}, Value {value}")
                csvwriter.writerow([sample, value])
            buffer = lines[-1]
            csvfile.flush()  # Ensure data is written to file immediately
    except KeyboardInterrupt:
        print("Recording stopped by user")

print(f"Data saved to {filename}")

audio_data_20240716_172132.csv

Is this what I should expect? Given that the above code is not working anymore for the board, what approach would you recommend so I can broadcast the digitised audio to web to listen to it (just to verify it is working)?

[pschatzmann]

https://github.com/pschatzmann/arduino-audio-tools/wiki/It's-not-working

Just test the value range of the ADC signal first with the sketch that I have sent you!
In the optimum case the voltage should oszillate between 0 and 3.3V which would give values from - 32768 to +32767

[Philomath88]

@pschatzmann I have set it to D32 pin. This is the serial output:

https://pastebin.com/FTTLyiFW

What i have seen in the Oscilloscope is that it hovers around 2.1V and then fluctuates by a maximum of 0.1V

[pschatzmann]

Just record a sine wave and display it with the Arduino Serial Plotter.
This will give you a better feeling if you get a proper signal that you can postprocess with a ConverterScaler (shift and multiply)

ps: send out the raw data and convert it on the desktop.
Sending ascii is adding too much overhead....