OggOpusEncoder bug

[smokesea]

Problem Description

read audio from microphone, encode it using OGG OPUs, transfer it to PC via network, and check the encoded file.

$ opusinfo test.s
Processing file "test.s"...

New logical stream (#1, serial: 11b6e847): type opus
Encoded with Arduino
WARNING: Samples with negative granpos in stream 1
WARNING: Sample count ahead of granule (3840>1959) in stream 1
WARNING: Sample count ahead of granule (5760>2000) in stream 1
WARNING: Sample count ahead of granule (7680>2036) in stream 1
WARNING: Sample count ahead of granule (9600>2072) in stream 1
WARNING: Sample count ahead of granule (11520>2104) in stream 1
WARNING: Sample count ahead of granule (13440>2142) in stream 1
WARNING: Sample count ahead of granule (15360>2182) in stream 1

Device Description

borad: esp32s3，psram enabled

Sketch

audio_encode.h
#include "Arduino.h"
#include "AudioTools.h"
#include "global.h"

//inmp441 pin
#define AUDIO_I2S_MIC_GPIO_WS 4
#define AUDIO_I2S_MIC_GPIO_SCK 5
#define AUDIO_I2S_MIC_GPIO_DIN 6

// I2S port
#define I2S_PORT I2S_NUM_0

#define AUDIO_SAMPLE_RATE     16000  // 44100
#define AUDIO_OPUS_FRAME_MS   60    
#define AUDIO_OPUS_BITRATE    16000 
#define AUDIO_OPUS_COMPLEXITY 0    
#define AUDIO_CHANNELS 1


void EncodeTask(void *param);



audio_encode.cpp

#include "audio_encode.h"
#include "AudioTools/AudioCodecs/CodecOpusOgg.h"
#include "AudioTools/Communication/WebSocketOutput.h"

AudioInfo encodeInfo(AUDIO_SAMPLE_RATE, AUDIO_CHANNELS, 16);
I2SStream mic;
WebSocketOutput wsOut(webSocket);

OpusOggEncoder opusOggEncoder;
EncodedAudioStream audioEncoder(&wsOut, &opusOggEncoder);
StreamCopy copier(audioEncoder, mic);

void startAudioRecord(){
    // AudioLogger::instance().begin(Serial, AudioLogger::Debug);
    webSocket.sendTXT(":micOn");

    auto cfg = mic.defaultConfig(RX_MODE);
    cfg.setAudioInfo(encodeInfo);
    cfg.sample_rate = AUDIO_SAMPLE_RATE;
    cfg.bits_per_sample = I2S_BITS_PER_SAMPLE_16BIT;
    cfg.channels = I2S_CHANNEL_MONO;
    cfg.i2s_format = I2S_STD_FORMAT;
    cfg.channel_format = I2S_CHANNEL_FMT_ONLY_LEFT;
    cfg.use_apll = false;
    cfg.is_master = true;
    cfg.fixed_mclk = 0;
    cfg.pin_ws = AUDIO_I2S_MIC_GPIO_WS;
    cfg.pin_bck = AUDIO_I2S_MIC_GPIO_SCK;
    cfg.pin_data_rx = AUDIO_I2S_MIC_GPIO_DIN;
    cfg.pin_data = I2S_PIN_NO_CHANGE;
    mic.driver()->setWaitTimeReadMs(AUDIO_OPUS_FRAME_MS);
    mic.begin(cfg);
    //wait for mic to start ??
    vTaskDelay(pdMS_TO_TICKS(50));

    opusOggEncoder.setAudioInfo(encodeInfo);
    auto& cfgEncoder = opusOggEncoder.config();
    cfgEncoder.application = OPUS_APPLICATION_VOIP;
    cfgEncoder.bitrate = AUDIO_OPUS_BITRATE;
    cfgEncoder.sample_rate = AUDIO_SAMPLE_RATE; 
    cfgEncoder.complexity = AUDIO_OPUS_COMPLEXITY; 
    cfgEncoder.frame_sizes_ms_x2 = OPUS_FRAMESIZE_60_MS;
    cfgEncoder.channels = AUDIO_CHANNELS;
    opusOggEncoder.begin();
    
    while(globalData.getMicStatus()){
        int begin = millis();
        size_t cc = copier.copy();
        int delay_ms = millis() - begin;        
        if (delay_ms < AUDIO_OPUS_FRAME_MS){
            vTaskDelay(AUDIO_OPUS_FRAME_MS - delay_ms);
        }
    }
    
    copier.end();
    vTaskDelay(10);
    webSocket.sendTXT(":micOff");
    Serial.println("Mic Off");
}


void EncodeTask(void *param){

    while(1){
        if (!WiFi.isConnected()){
            vTaskDelay(100);
            continue;
        }
        if ( globalData.getSpeaking() || !webSocket.isConnected())
        {
            vTaskDelay(10);
            continue;
        }
        if(globalData.getMicStatus()){
            Serial.println("Start Audio Record");
            startAudioRecord();
            Serial.println("End Audio Record");
        }
        vTaskDelay(100);
    }
}

Other Steps to Reproduce

My master, please guide me

What is your development environment (incl. core version info)

framework: arduino
platform: platform.io

I have checked existing issues, discussions and online documentation

• I confirm I have checked existing issues, discussions and online documentation