Hello all,

I asking for advice on the best way to read single cycle waveforms from a bank :

Every single cycle wavform is made of 2048 points (samples). The aim is to morph from waveform_1 to waveform_2.
In order to save computing time (this is made to run on a RPi), the morphing is saved into a "bank" where the transitions from Waveform_1 to Waveform_2 are stored one after the other one and the speed of this wavform change can be modulated.
The bank is a wavfile.

At this stage I use a phasor~ to read the wavform and move the offset (0, 2048, 4096, ...) in the bank for reading but the result sounds bad (kind of a buzzing sound like the phasor was being part of the sound).

Here is the elementary patch and a screenshot of what a bank look like.

Is there an efficient way for doing this please? (/uploads/files/1577112224693-capture-du-2019-08-24-21-49-55.png)

"Every single cycle wavform is made of 2048 points"

tabread4~ needs 3 points more. (see tabread4~ help-patch)

@JMC64 phasor~ is part of the sound, because when reading the wavetable, a jump is occuring at it's EDIT: falling edge . Better use a triangle.
Instead of using snapshot~ in the slow control domain, you could shift by the right [inlet~] with [+~] in the signaldomain faster.
To shift at 0, 2048, 4096 exclusively, run a [phasor~] synced to the triangle and flush a [samphold~] with it's falling edge at the start of a phase.
Or clip the triangle by a very high number to get a rectangular shape flushing [samphold~] at the falling edge: [*~ 9999999999999] connected to [clip~ 0 1] connected to [ samphold~] .
All this is switching or alternating, not morphing continiously.

@lacuna : I have trouble understanding. Could you please make a short patch explaining what you mean with the triangle? Thank you a lot .. I hope it might help others too.

@JMC64 I currently can't run pd here, as im writing on a phone, but I'll explain
@lacuna said:

@JMC64 phasor~ is part of the sound, because when reading the wavetable, a jump is occuring at it's rising edge . Better use a triangle.

Sorry, that was wrong: the jump is at the falling edge.
Looking at the shape of phasor~s waveform: / / / /
it is a ramp from 0 to 1. Then it jumps within one sample back to 0. That's a sawtooth.
With a triangle ^^^^
you can [tabread4~] parts of the array back and forth, without any glitch.

As you change its range from 0-1 to 0-2048 by multiplying with [*~ 2048] ,
[+~ 2048] is adding an offset: 2048-4096

snapshot~ and all other control-objects (everything without tilde ~ ) have a resolution of one blocksize, not updating the right inlet of [tabread4~] on each sample, but up to 64 samples too late.
Control objects calculate and output between the blocks with 64 as fastest rate (even if the blocksize is changed by [block~] )
(Bela is running a modified PD with 4 or 8 samples blocksize as minimum, I think - but still. )
Signal ~ objects calculate between the blocks and output continiously samplewise within the block of the realtime audio stream.

use an additional [tabread4~] to crossfade between the sounds too. I'd also recommend keeping the wavetable index a signal (then the fractional part is how much of the current vs. next wavetable).
you might have resolution issues with just using [+~ ] for the wavetable offset due to 32-bit resolution, but I don't think you'll have much difficulty for normal wavetable sizes.

there was a library iem_dp that had a tabread~ that allowed a signal to be used for the offset, but I don't see it in deken now
my library also has [tabread4hs~] which also has a signal inlet for offset (but uses different interpolation)

With a crossfader you can even do continious morphing.
Yes, with large arrays you'll get resolution issues. These already happen with [phasor~][*~ highnumber], too - because it does the rounding between the range of 0 to 1, some circles not reaching 0 or 1, while keeping its frequency perfectly stable
https://forum.pdpatchrepo.info/topic/12865/samphold-noise-phasor-noise-round-off-error-or/15
EDIT / correction: not perfectly stable, as it jitters with the samplerate.