Custom audio element Samples

[CoolingCoder]

Hi. I'm currently developing code using ESP-ADF, where I created a custom audio element to process the samples of the right channel in order to implement a feedback process in some kind of passthru program. However, I don't know how to properly understand the samples since I expected that whenever I had silence detected from the microphone, the samples should be around the value of 32768 (which is half the maximum value). Instead, I am getting samples around the maximum and minimum values:

I (58355) PASSTHRU: RSample: 65146
I (58365) PASSTHRU: RSample: 64844
I (58385) PASSTHRU: RSample: 64878
I (58415) PASSTHRU: RSample: 85
I (58445) PASSTHRU: RSample: 765
I (58475) PASSTHRU: RSample: 867
I (58485) PASSTHRU: RSample: 204
I (58505) PASSTHRU: RSample: 64932
I (58535) PASSTHRU: RSample: 64773
I (58565) PASSTHRU: RSample: 286
I (58595) PASSTHRU: RSample: 131
I (58605) PASSTHRU: RSample: 651
I (58625) PASSTHRU: RSample: 184
I (58655) PASSTHRU: RSample: 398
I (58685) PASSTHRU: RSample: 64980

Additionally, it all sounds good when I plug the output into a speaker, so I know it's purely an interpretation issue.
I hope some of you can help me understand how these samples truly work so I can continue developing my program.


This is the code I'm currently using:

Code: [Select all] [Expand/Collapse]

1. ESP_LOGI(TAG, "[ 2 ] Start codec chip");
2.     audio_board_handle_t board_handle = audio_board_init();
3.     audio_hal_ctrl_codec(board_handle->audio_hal, AUDIO_HAL_CODEC_MODE_BOTH, AUDIO_HAL_CTRL_START);
4.     audio_hal_set_volume(board_handle->audio_hal, 60); //60
5.  
6.     ESP_LOGI(TAG, "[ 2 ] Create audio pipeline for playback");
7.     audio_pipeline_cfg_t pipeline_cfg = DEFAULT_AUDIO_PIPELINE_CONFIG();
8.     pipeline = audio_pipeline_init(&pipeline_cfg);
9.  
10.     ESP_LOGI(TAG, "[ 3 ] Create i2s stream to write data to codec chip");
11.     i2s_stream_cfg_t i2s_cfg_w = I2S_STREAM_CFG_DEFAULT();
12.     i2s_cfg_w.type = AUDIO_STREAM_WRITER;
13.     i2s_stream_set_channel_type(&i2s_cfg_w,I2S_CHANNEL_TYPE_ALL_LEFT);
14.     i2s_stream_writer = i2s_stream_init(&i2s_cfg_w);
15.  
16.     ESP_LOGI(TAG, "[ 3.1 ] Create i2s stream to read data from codec chip");
17.     i2s_stream_cfg_t i2s_cfg_r = I2S_STREAM_CFG_DEFAULT();
18.     i2s_cfg_r.type = AUDIO_STREAM_READER;
19.     i2s_stream_set_channel_type(&i2s_cfg_r,I2S_CHANNEL_TYPE_ALL_LEFT);
20.     i2s_stream_reader = i2s_stream_init(&i2s_cfg_r);

GeSHi © Codebox Plus Extension

Code: [Select all] [Expand/Collapse]

1. static int audio_inverter_process(audio_element_handle_t self, char* buf, int len)
2. {
3.     int rsize = audio_element_input(self, buf, len);
4.  
5.     if (len != rsize || (rsize % 4) != 0) {
6.         ESP_LOGW(TAG, "unexpected rsize: %d, len: %d", rsize, len);
7.     }
8.  
9.     uint16_t LSample;
10.     char * lsamplep = (char *) &LSample;
11.     uint16_t RSample;
12.     char * rsamplep = (char *) &RSample;
13.     float sample = 0;
14.  
15.     for (int i =0; i<rsize; i+=4) {
16.         rsamplep[0] = buf[i];
17.         rsamplep[1] = buf[i+1];
18.         lsamplep[0] = buf[i+2];
19.         lsamplep[1] = buf[i+3];
20.        
21.         buf[i] = rsamplep[0];
22.         buf[i+1] = rsamplep[1];
23.         buf[i+2] = lsamplep[0];
24.         buf[i+3] = lsamplep[1];
25.     }
26.  
27.     ESP_LOGI(TAG, "RSample: %hu",RSample);
28.    
29.     rsize = audio_element_output(self, buf, rsize);
30.  
31.     return rsize;
32. }

GeSHi © Codebox Plus Extension

[MicroController]

so I know it's purely an interpretation issue.
I hope some of you can help me understand how these samples truly work

Interpret the samples as signed values, i.e. int16_t.