The difference between the sample counts of readCB and writeCB

[je1rav]

Hi,
The following is the serial monitor output of “audio_headset.ino”.
Total Microphone samples: 6256 / Speaker samples: 4762 / Sample rate: 44100
Total Microphone samples: 98348 / Speaker samples: 93050 / Sample rate: 44100
Total Microphone samples: 190440 / Speaker samples: 181338 / Sample rate: 44100
Total Microphone samples: 282532 / Speaker samples: 269626 / Sample rate: 44100
Total Microphone samples: 374624 / Speaker samples: 357914 / Sample rate: 44100
Total Microphone samples: 466716 / Speaker samples: 446202 / Sample rate: 44100
Total Microphone samples: 558808 / Speaker samples: 534492 / Sample rate: 44100
Total Microphone samples: 650900 / Speaker samples: 622780 / Sample rate: 44100
Total Microphone samples: 742992 / Speaker samples: 711068 / Sample rate: 44100
The difference from one previous output of the microphone is always 92092 and that of the speaker is 88288.
For each of readCB and writeCB, the number of samples for readCB(92) is 2 or 4 more than for writeCB(88 or 90).
In the case of the sample rate of 48000, 98 for readCB and 96 for writeCB.
Seems like too much for readCB.
Do you have any idea about this?

[je1rav]

HI,
I tested "audio_microphone.ino" with slightly change the readCB(){} as follows.

"size_t readCB(uint8_t* data, size_t len, Adafruit_USBD_Audio& ref) {
int16_t* data16 = (int16_t*)data;
size_t samples = len / sizeof(int16_t);
size_t result = 0;
// generate random data
for (int j = 0; j < samples; j++) {
data16[j] = (sample_count % 20) * 100; //simple saw wave
result += sizeof(int16_t);
sample_count ++;
}
return result;
}"
The result was

Discontinuous changes are seen in the third and eighth waves.

After change the code to
for (int j = 0; j < samples-2; j++) {

No discontinuous changes are seen.