TobyAM

in CodeBox~ to try and create a new fftstream state object, but I get an error every time.

Can anybody throw me a bone on how to use this?

I've tried many variations on :
@state fftleft = new fftstream();
in CodeBox~ to try and create a new fftstream state object, but I get an error every time.

I've tried many variations on :
@state fftleft = new fftstream();
in CodeBox~ to try and create a new fftstream state object, but I get an ... 

error-compiling-codebox~-fftstream

No matter what arguments I enter into fftstream() I get the same compile errors:

I've been trying this for hours so I would love just a simple example of this CodeBox FFT.

image.png

No matter what arguments I enter into fftstream() I get the same compile errors:
I've been trying this for hours so I would love just a simple example of this CodeBox FFT.

oh, i see now, you tried it as a ‘stateful operator’… maybe it isn’t one? and maybe codebox doesn’t actually do audio in rnbo-naked(?)… the helpfile for codebox in rnbo-naked doesn’t show any audio processing, it shows utilizing ‘peek’/‘poke’/‘interp’ ops as though they’re already proper ‘ops’ within the codebox system, but the example still translates out to events only(no audio in that example coming out of the codebox itself, the comment says ‘this codebox is event rate’ …but then that confuses me some more as i go looking into its inspector to see how to turn an rnbo-naked codebox into an audio-rate codebox(?))... think they just need to add a bit to those rnbo-naked docs.

not sure it’s a stateful-operator to begin with 🤔 (?) …good tries, tho! you got me curious :)

i saw your other thread's reply, now i get it:

codebox~(with the tilde) is where you do it! ha, i still have much to learn about rnbo~. will try it some more, thank you!

i saw your other thread's reply, now i get it:
codebox~(with the tilde) is where you do it! ha, i still have much to learn about rnbo~. will try it some more, thank you!

All options are optional, so the simple instantiation "@state fftleft = new fftstream();" should work, right? Is this operator broken?

Thanks for the report @tobyam! We'll look into it.

this seems to now be broken in a new way in 1.0.2

codebox~: No option of name: fftsize in object: codebox_tilde

given other bug reports, im wondering if going back to 1.0.1 is the way forward...

can I do this even if Ive updated to 8.5.2

@state f= new fftstream();
codebox~: No option of name: fftsize in object: codebox_tilde

was this tested?
given other bug reports, im wondering if going back to 1.0.1 is the way forward...
can I do this even if Ive updated to 8.5.2

The new error isn't related to any changes to fftstream, though the errors here are something we're aware of and looking into.

In the meantime, if you'd like to try fftbuffer in codebox~, I have an example you can try out so you can use FFTs in codebox~.

I get very different results depending on my IO Vector Size and Signal vector size, but it's maybe just restarting the audio. Sometimes 0 peaks, sometimes 4 peaks, rarely the expected 2 peaks.

Is something off with this example?
I get very different results depending on my IO Vector Size and Signal vector size, but it's maybe just restarting the audio. Sometimes 0 peaks, sometimes 4 peaks, rarely the expected 2 peaks.

Yep, definitely issues with the codebox example when restarting the audio.

At least with [fft~], the result is correct after audio rebuilds:

Yep, definitely issues with the codebox example when restarting the audio.
At least with [fft~], the result is correct after audio rebuilds:


Thanks for pointing out the signal and vector size changes causing bugs, Jean-Francois. There's indeed something going on here with the codebox example.

I don't experience any issues with the codebox example when my signal and vector size are 512 and the FFT size is 512. If they're not, I can reproduce the fact that the RNBO object works but the codebox example doesn't work reliably. I will investigate further.

error compiling Codebox~ fftstream