uint32_t p1 = 0;
uint32_t p2 = 0;

bool rtrig = 0;

int32_t f1;
int32_t f2;
MTOFEXTENDED(inlet_master + param_master, f1);
MTOFEXTENDED(inlet_slave + param_slave, f2);

if (inlet_reset && !rtrig) {
  rtrig = 1;
  p1 = 0;
  p2 = 0;
} else if (!inlet_reset) {
  rtrig = 0;
}

/*
 *
 *                          /|
 *                         / |
 * 				     /  sum1
 *                     /     |
 *                   /_______|                         d1
 *                 /|        |			a =   -----------
 *               /  |        |                  	 sum1 + d1
 *             /    |        |
 *           /     d1        |
 *         /        |        |
 * 	    /          |	    |
 *     /______a_____|__1-a___|
 */

// here i do some geometry, trying to avoid irregular phase resets
// p1 and p2 are unsigned 32 bit int variables.. You can see from simplicity
// them as going from 0 (0x00000000) to 1 (0xFFFFFFFF)
uint32_t sum1 = p1 + f1;       // integrating frequency in phasor 1
uint32_t sum2 = p2 + f2;       // integrating frequency in phasor 2
uint32_t d1 = 0xFFFFFFFF - p1; // distance between p1 (not integrated) and 1
uint32_t sum1_d1 = sum1 + d1;  // distance between p1 (integrated) and 0 . This
                              // is a positive number and makes sense only if the
                              // phasor has reset (by overflowing the variable)
int32_t a =
    134217728.f *
    ((float)d1 / (float)sum1_d1); // this is a variable that goes from 0 to 2^27
                                  // (i'm confortable working with Q27) and tells
                                  // you where the reset exactly happens
int32_t a_compl = (1 << 27) - a;               // 1-a
uint32_t d2 = ___SMMUL(a_compl << 4, f2 << 1); // see the triangle

if (sum1 < p1) {
  p2 = d2;
} else {
  p2 = sum2;
}

p1 = sum1;

int32_t r1;
int32_t r2;

SINE2TINTERP(p1, r1);
SINE2TINTERP(p2, r2);

outlet_o1 = r1 >> 4;
outlet_o2 = r2 >> 4;