DSP Robotics Support

Hello!

Just wanted to know if anyone did experiments with a wave-table-osc that is changing the wave-form by time by interpolating between 2 waves?

Thank you!
C.Hackl

I only know of crossfading solutions with arithmetic oscillators, not wavetables. Those were made by Spogg and somebody else that I forgot

tulamide wrote:I only know of crossfading solutions with arithmetic oscillators, not wavetables. Those were made by Spogg and somebody else that I forgot

I think this is about this synth:

viewtopic.php?f=3&t=4419&p=24991&hilit=quilcom+nine#p24973

It's on my list to make a new version at some point.

But I don't know anything about scanning wavetable oscillators, and I would like to. I think Adam did this in his Viper but that's a commercial product.

what we could do is make a short audio of different cycles one after another,
these are read back with a counter with which you could program to different kind of reads
the audio is a mem or float array, and 'wave/mem read' is used to read the counter index

OceanSwift have done such a thing previously with FS

Why not just read both wavetables and crossfade then

chackl wrote:Just wanted to know if anyone did experiments with a wave-table-osc that is changing the wave-form by time by interpolating between 2 waves?

martinvicanek wrote:Why not just read both wavetables and crossfade then

From where I'm sitting, those sound like exactly the same thing.

As you pronounced (wrote maybe better) the word dynamic,
my guess is that you probably will be changing table sizes...
I've spotted years ago, and then DWB confirmed, a "bug" when changing array sizes...
I've come up with a solution, as well as DWB, and it seems that DWB's fix was the one retained...
Here is a link to the fix:

viewtopic.php?f=2&t=2694&p=14009&hilit=dwb#p14009

Hope this helps to put you in the right track
Take care

More info needed - the OP is too vague.

Hello!

Thank you for the massive input - yes it seems that i realy failed explaining at some point

dynamic is not suposed to be "dynamic wave size" - the wave-table size is fixed

martinvicanek wrote:Why not just read both wavetables and crossfade then

Yes this is what i asked for.

As i experimented with FFTs some time ago i came across the idea to analyse a sound's wave form every example 0.05 second and generate multible wave-tables out of it. Then load each wave to a wave-table and put them into a mem array - and read the mem array by an interpolation - like two wave-tabel-read modules.

I have seen a project some time ago that had an optimiced wave-table-read with 2 interpolated wave-read modules after it. So i was thinking about extending this to a wave-array-read module with wave tables. (But i do not find the project - maybe someone else knows it)

I would be interested if this would leed to a possibility to do a "more complex" system here.

Thank you

The download in this post should give you a start with the wavetable reading parts. The thread is really about envelopes, but if you dig down into the dowload you'll find a "wavetable array" (mem array containing multiple wavetables), and an oscillator which can read any wavetable out of the array.