Changing the delay time

Hi Charles,
I think you are referring to this post: https://nootropicdesign.com/store/forums/topic/audio-delay-line/

The original poster wanted to use buttons to control the delay amount instead of a pot, so that’s why he introduced the delayValue variable.

To control the delay amount, change echoDelay and make sure it is a multiple of 3. You may choose to control echoDelay with a pot on A0. The original code is:

echoDelay = analogRead(0) * 30;

and that means that echoDelay will have a value from 0-30690.

Howdy, Michael —

I’m still having a bit of trouble changing the length of the delay to more than the 2 seconds it produced originally.

I change echoDelay = 24000 but that has no effect. Is there another delay modification that I’m not finding?

Any help would be appreciated.

Thanks, Charles

Here’s the code:

#include <AudioHacker.h>

#define DEBUG

unsigned int playbackBuf;
unsigned int sampleRate;
unsigned int readBuf[2];
unsigned int writeBuf;
boolean evenCycle = true;
unsigned int timer1Start;
volatile unsigned int timer1EndEven;
volatile unsigned int timer1EndOdd;
unsigned long lastDebugPrint = 0;
volatile long address = 0;
unsigned int echoDelay;
boolean echoWrapped = false;

void setup() {
#ifdef DEBUG
  Serial.begin(115200);        // connect to the serial port
#endif

  playbackBuf = 2048;
  sampleRate = DEFAULT_RECORDING_SAMPLE_RATE;
  timer1Start = UINT16_MAX - (F_CPU / sampleRate);

  AudioHacker.begin();

#ifdef DEBUG
  Serial.print("sample rate = ");
  Serial.print(sampleRate);
  Serial.print(" Hz");
  Serial.println();
#endif
}

void loop() {

  delay(300);
  // echoDelay is number of memory slots back into the past to read for the echo.
  // must be a factor of 3 because we pack 2 samples into each 3-byte sequence of SRAM.
  echoDelay = 24000; 

#ifdef DEBUG
  if ((millis() - lastDebugPrint) >= 1000) {
    lastDebugPrint = millis();

    //each 3 echoDelay2 is 2 samples.  
    unsigned int delayMillis = (float)(((float)((echoDelay * 2) / 3)) / (float)(sampleRate/1000.0));

    Serial.print("echo delay = ");
    Serial.print(delayMillis);
    Serial.print(" ms    even cycles remaining = ");
    Serial.print(UINT16_MAX - timer1EndEven);
    Serial.print("   odd cycles remaining = ");
    Serial.print(UINT16_MAX - timer1EndOdd);
    Serial.println();
  }
}
#endif

ISR(TIMER1_OVF_vect) {
  TCNT1 = timer1Start;
  unsigned int signal;
  unsigned int echo;
  int mix;

  AudioHacker.writeDAC(playbackBuf);

  // Read ADC
  signal = AudioHacker.readADC();

  if (evenCycle) {
    long echoAddress = address - echoDelay;
    if (echoAddress < 0) {
      echoAddress += MAX_ADDR;
    }
    AudioHacker.readSRAMPacked(0, echoAddress, readBuf);
    if ((!echoWrapped) && (echoAddress > address)) {
      // avoid reading from unwritten memory
      echo = 2048;
      readBuf[1] = 2048;
    } else {
      echo = readBuf[0];
    }
  } else {
    echo = readBuf[1];
  }
  if (echoDelay == 0) {
    echo = 2048;
  }

  if (evenCycle) {
    writeBuf = signal;
  } else {
    AudioHacker.writeSRAMPacked(0, address, writeBuf, signal);
    address += 3;
    if (address > MAX_ADDR) {
      address = 0;
      echoWrapped = true;
    }
  }

  
  playbackBuf = echo;

#ifdef DEBUG
  if (evenCycle) {
    timer1EndEven = TCNT1;
  } else {
    timer1EndOdd = TCNT1;
  }
#endif
  evenCycle = !evenCycle;
}

Changing the echoDelay value seems to have no audible effect at all. I’ve tried many different values from around 300 to 48000 (all divisible by 3), and I can’t hear, or measure with a stopwatch, any significant change in the delay, which is consistently around 2 sec.

But when I use a very high value, such as 90000, it produces a lot of feedback-like noise, so there definitely is a maximum limit.

I’m not using any pots, but just changing the value in the code and uploading to the Audio Hacker to test it out.

I have the DJ Shield (unassembled, unfortunately) so I’ll put that together and try using a pot with the echo example. Hopefully, I’ll get to it this weekend.

Thanks for the tip… I changed echoDelay to
unsigned long echoDelay;
and now have a 4 second delay maximum.

Do you have any thoughts on how I can reach the 8.6 second maximum, as in the documentation?

Well, that is disappointing.

Thanks anyway…