Hi everyone, I hope you are all doing well. I am new to this and I have entered the forum so that you can help me to know how to convert wav files to files that are compatible with Syntakt. Since there are few packs, I would like to upload my own sounds. Regards.
Syntakt uses analog and digital synthesis. It does not support samples or waves, so you cannot do this
Ok, but how can I create my own sounds for Syntakt to receive?
Create them by adjusting the synthesis parameters on the device or by editing the device via the Overbridge plugin which exposes all the synthesis parameters on a computer
Synthesis requires user input but not that much expertise because the synthesis algorithms are broadly simplified (sorta grouped into macros) to produce musical results with fewer parameters to adjust than would be required
No wavs, there’s no translation from wavs to synthesis settings either
1 Like
Can these files be created on another device such as Analog Rytm MkII?
No. The Syntakt’s sounds all come from manipulating the internal machines. There is no uploading or creating files via another machine.
You can save presets of sounds but those only work between Syntakts.
Thanks for your prompt reply, I’ll try to somehow create my favorite sounds to add to Syntakt or even share.
I would recommend reading the manual for Syntakt: https://cdn.www.elektron.se/media/downloads/syntakt/Syntakt_User_Manual_ENG_OS1.21_231108.pdf
Syntakt unfortunately has no ability to accept wav/aiff or any audio files.
The packs like this ( https://www.elektron.se/us/tone-spells ) sold on Elektron are just synth presets packs of the internal machines, no samples or audio files.
2 Likes
You can only create sounds for a Syntakt on a Syntakt.
It’s not a sampler. It can’t play wav files.
It’s a synthesizer.
1 Like
If you’re looking for a device like Syntakt that allows you to add your own sounds as .wav files, you should look into the Digitakt instead of the Syntakt 
Would be quite nice to have some sort of a cross device standard like the dx7 files seem to be in FM. All synths are different, but there are also lots of similarities and it could go for the closest match based on a device having categories / tags for all its filters, oscillators etc. Like MIDI but for a patch. Would also allow for software editors / librarians to be built easier.
As with MIDI, it’d need someone to step up and implement it and then talk to other manufacturers.
I made a Neural Network that converts wav files to ST patches…I’ve only trained it with 5 of the digital machines though…it’s still early days.
j
0=0
4 Likes
Neural nets
Fascinating! I just recently saw a video demonstration of Synplant, and this seems like a really exciting idea, would be keen on hearing more about it as you progress. (I know just enough to imagine setting up something in PyTorch and MIDI, but nothing firsthand about audio/DSP neural nets, but was starting to poke around GuitarML, a NN-based amp modeler…)
1 Like
basically the flattened layer of the NN are the parameters of the Syntakt.
A common method would be to make very extreme and random patches for the ST and get mel spectrograms of the output and make your model that way. You can then convert your waveform to a Mel Spectrogram (using the same settings) and it can interpolate the date to get something to generate as close to that image as possible using the model.
2 Likes