DominikK

I'm trying to build an external which allows for variable delay line pitch shifting, basically the functionality of the "transposer" abstraction.

Since it's only meant to be an exercise it's not that critical but I'd still like to get it right.

In general the whole thing works already, but there is still an error somewhere I can't find, which creates a noisy sound as well as making the original material sound somehow distant. Also as the amount of pitch shift gets bigger some skewed (modulation?) frequencies are audible. At first I thought that linear interpolation is not precise enough, so I tried implementing Lagrange interpolation but that didn't solve the issue (unless my interpolation function is flawed of course).

I've been spending a lot of time trying to fix this but I think I need some help. I'll post the main part of the perform routine as well as the interpolation function.

phasorfreq = (1 - pitch_factor) * samplerate / windowsizesamps;

si = phasorfreq * siFactor; // siFactor = windowsize in samples / sample rate

phasorValue2 = phasorWavetable[(iphase+halfwindowsizesamps)%windowsizesamps]; // 180° phase shifted

ddelay1 = phasorValue1 * windowsizesamps;

delay_line1 = lagrangeInterp(x, delay_line, delay_length, read_index1, ddelay1) * linInterp(x, windowWavetable, x->windowsizesamps, ddelay1);

//windowWavetable is half a sinewave period

//same code as delay_line1 using different indices


  double* delay_line = Array;
    int delay_length = ArraySize;
    double interpSample = 0;
    double p = 0;
    double fraction = ddelay - trunc(ddelay);
    
    for (int i = -1; i 

I'm trying to build an external which allows for variable delay line pitch shifting, basically the functionality of the "transposer" abstraction.
Since it's only meant to be an exercise it's not that critical but I'd still like to get it right.
In general the whole thing works already, but there is still an error somewhere I can't find, which creates a noisy sound as well as making the original material sound somehow distant. Also as the amount of pitch shift gets bigger some skewed (modulation?) frequencies are audible.  At first I thought that linear interpolation is not precise enough, so I tried implementing Lagrange interpolation but that didn't solve the issue (unless my interpolation function is flawed of course).
I've been spending a lot of time trying to fix this but I think I need some help. I'll post the main part of the perform routine as well as the interpolation function.

        pitch_factor = pow(2, cents/100/12);
        phasorfreq = (1 - pitch_factor) * samplerate / windowsizesamps;

        si = phasorfreq * siFactor;         // siFactor = windowsize in samples / sample rate
        iphase = trunc(phase);
        phasorValue1 = phasorWavetable[iphase];
        phasorValue2 = phasorWavetable[(iphase+halfwindowsizesamps)%windowsizesamps];       // 180° phase shifted
        phase += si;

        if (phase >= windowsizesamps) {
            phase -= windowsizesamps;
        }

        if (phase < 0) {
            phase += windowsizesamps;
        }

        delay_line[write_index++] = *input++;

        if (write_index >= delay_length) {
            write_index -= delay_length;
        }

        ddelay1 = phasorValue1 * windowsizesamps;
        idelay1 = trunc(ddelay1);
        read_index1 = write_index - idelay1;

        if(ddelay1 < 0) {
            ddelay1 = 0;
        } else if(ddelay1 > delay_length) {
            ddelay1 = delay_length - 1;
        }

        while (read_index1 < 0) {
            read_index1 += delay_length;
        }
       delay_line1 = lagrangeInterp(x, delay_line, delay_length, read_index1, ddelay1) * linInterp(x, windowWavetable, x->windowsizesamps, ddelay1);
    //windowWavetable is half a sinewave period

        //delay line 2
       //same code as delay_line1 using different indices

        *output++ = delay_line1 + delay_line2;

        if(write_index >= delay_length) {
            write_index -= delay_length;
        }

var-delay-line-pitch-shifter

I'm sorry, I seem to be too dump to use the "code" tags. I thought it was 

[e] Which is apparently the way it works in posts...

I'm sorry, I seem to be too dump to use the "code" tags. I thought it was  mycode  but that seems to create rubbish...

[e] Which is apparently the way it works in posts...


there is also [code] [code] [/code] here, but that tends to kill line breaks.

lol, you cant set the code tag in code tags ... here is it with additional spaces

Var delay line pitch shifter