Why are my wavetables out of tune on DTII?

Hi Elektronuats,

I’m in the process of making wavetables for use in the DTII’s Grid Mode. For the most part this has been a success, but when I play the wavetables back on on the DTII they are not tuned to the C note I exported the wavetables as. I need to set the tune parameter to -4.74 semitones to get the pitch back to middle C

Does anyone know why this is happening? Playing the wavetables as .wav files I can hear they are out of tune raw as well, but when loading them into Pigments they are tuned correctly.

I assume somewhere there is a format/bitrate/cycle length mismatch but I’ve been fooling around with various setting and havent made a difference. Phase Plant and Node are both spitting out 44.1k .wavs, not sure if that’s causing issues but I would assume not since the raw .wavs sound out of tune as well.

Any insight into why this is happening and a possible remedy? I’d be satisfied enough just repitching/respeeding the wavetables manually but I cannot find a tool that will batch repitch without timestretching.

I’ve encountered similar behaviors with different SCWs and Wavetables, and I think it comes from the standardization of sample lengths to 2^x numbers rather than to frequencies.

I haven’t noticed any loss of fidelity retuning them (presuming that’s what happens when they’re played anyway), or saving a preset on DT2 with the right tuning also works.

1 Like

Hey thanks, but I’m a bit unclear. What do you mean standardizing the sample lengths? Is this something that happens upon wavetable creation or something that happens during playback in a dedicated wavetable synthesizer?

I haven’t noticed any loss of fidelity retuning them withing the DTII either, but would much rather have the tuning baked in to the sample. Still looking for a tool that can do that.

1 Like

You can just open the wavetable as a wav file in your daw or audacity and follow instructions on repitching it there, then bounce it out as a new .wav.

Well I’ve gotten close by using Bitwig’s Raw mode and altering the tempo, but every time I bounce audio Bitwig seems to be applying an auto-fade out regardless of my settings.

Can’t seem to get there with audacity either, every repitch mode seems to time-stretch as opposed to respeed.

Repitching in Ocenaudio introduces all kinds of issues as well

I think I saw a bug about things being a semitone out on the bug thread. … check in there

Additionally altering pitch in audacity to 76.04% corrected the pitch but introduced artifacts

Thanks, I will look around. Although the fact that the .wav and the WT on the Digitakt play at the same pitch has me thinking the DT is not necessarily the issue

The culprit can be auto zero crossing snap for slices.
That was my conclusion with DT1 concerning some wavetables I made, going out of tune. Random wavetables didn’t work properly, a fews waves were out of tune.

I made a wavetable with the same sine wave for the whole wavetable, it was working, tune wise.
Ideally you may need a fade out for each wave in order to have zero crossings.

I thought I remember reading that zero-crossing snapping was disabled in the DTII’s grid mode, and I don’t see it referenced in the manual. As well, all slices are at the same pitch, which makes me think that it is not snapping? Not certain though.

Where ?

Someone from Elektron said so for DT1.

Not sure where I remember but I could be wrong. Thought I remember a post from someone at elektron as well. Do you have the post you’re referencing?

Regardless, the wavetables seem to work remarkably well on DTII. I’m not noticing any tuning other thank the inital pitch.

And exporting from bitwig vs. bouncing seems to solve the issue with the fadeout, so for now I can create usable wavetables. I would love to learn why this is happening though.

@Eak I probably missed that one : :content:

Good news if so…

1 Like

Nice! Absolutely good news, makes the grid machine much more usable as a WT synth. The DTII has some built in wavetables that sound great

1 Like

So it worked better for you with a fade out on each wave ?
If so it seems like a zero crossing problem…
Have to check…

No, a fade out on each wave is not needed. The issue was with Bitwig, when bouncing audio an auto fade out was applied so the last few cycles of the entire wavetable were at a lower level than the others. I’ve gotten around it by exporting the files vs. bouncing (which is probably faster anyway)

1 Like