BYTEBEAT-XFADE.ino reads analog inputs without analogRead()

The BYTEBEAT-XFADE does not use analogRead(). The values of the potentiometers are read in the "ISR(TIMER2_OVF_vect)" function and stored in in1_pot1, in1_pot2, in3 and pot3.

#define in1_pot1 gp->cv[2]
#define in2_pot2 gp->cv[1]
#define in3 gp->cv[3]
#define pot3 gp->cv[0]
Unfortunately I am not able to explain more details because I don't understand them (yet).

I found this page: garretlab.web.fc2.com/en/arduino/inside/hardware/arduino/avr/cores/arduino/wiring_analog.c/analogRead.html
But I still don't understand all the details (yet).

Hi, from your link 

int analogRead(uint8_t pin){    if (pin >= 14) pin -= 14; // allow for channel or pin numbers    // set the analog reference (high two bits of ADMUX) and select the    // channel (low 4 bits).  this also sets ADLAR (left-adjust result)    // to 0 (the default).    ADMUX = (analog_reference << 6) | (pin & 0x07);    // without a delay, we seem to read from the wrong channel    //delay(1);    // start the conversion    sbi(ADCSRA, ADSC);    // ADSC is cleared when the conversion finishes    while (bit_is_set(ADCSRA, ADSC));    // ADC macro takes care of reading ADC register.    // avr-gcc implements the proper reading order: ADCL is read first.    return ADC;}
compared with this part of bytebeatxfade sketch

ISR(TIMER2_OVF_vect) {  gp->div32 = !gp->div32;                      // divide timer2 frequency / 2 to 31.25kHz  if(gp->div32)  {    gp->div16 = !gp->div16;     // sample channel 0-3 (CV) and 4 (audio) alternately so each channel is sampled with 15.6kHz    if(gp->div16)               // Get CV-data now    {      gp->pot_cv=ADCH;               // Store CV data detected for main-loop (will be processed with next sample from audio-in)      cbi(ADMUX,MUX0);               // Clear all CVs...      cbi(ADMUX,MUX1);               // ...on input-multiplexer                      sbi(ADMUX,MUX2);               // set multiplexer to channel 4 for audio    }    else                             // Get audio-data in this round    {      gp->cv[gp->pot_selector] = gp->pot_cv;  // store CV-data selected in last round for main loop      gp->audio_data=ADCH;                // get ADC channel 4      cbi(ADMUX,MUX2);                    // disable analogue input on channel 4            gp->pot_selector++;                // select next Pot or CV input      if(gp->pot_selector > 3 )          // Round-robin one of 4 possible CVs/pots        gp->pot_selector = 0;      ADMUX |= gp->pot_selector;        // Set current CV or Pot selected for non-audio input (round-robin)
     gp->new_sample=true;              // We selected audio and CV data now in interupt to be processed in next main loop    }    sbi(ADCSRA,ADSC);                   // ADC start conversion once, will be reset by the Atmega8, we have to set it each round...  }}
I start understand, more or less, how the ATmega "take a measure" from its analog inputs. Thanks

Here also section 23 from ATmega datasheet