Indeed, AndyCap is porting to the same processor like Akso! Ksoloti Core prototypes with that processor are already in our hands.
2 Likes
In a noise related incident `sss/noise/colouredâ causes the new prototype to crash in an abrupt and uncalled for way. With no optimisation it is absolutely fine, with optimisation of any type then disaster, debugger dies, no breakpoints hit, no idea of what is going on at all.
I only noticed this because I am making a âNoise Machineâ for my terrible sleeping, the Noise app I have on Alexa (like most software systems) has gradually been degraded to the point that they disable all other Alexa functions when the app is running, so if you want to lower the volume, canât do that if the app is running, want to turn the light off, canât do that is the app is running, etc, etc.
Using one of these: Picade Max USB Audio for the board, no FPU so stealing lots of axoloti/ksoloti Q27 code to make the ultimate noise machine (something I will fiddle with forever and will never finish)
What could possibly go wrong!
Hmm, hard to say, the opject is using SINE2TINTERP() and GenerateRandomNumber(), and the local ALLPASS() has some two-dimensional array action going on.
It is crashing on my 1.1.0 Ksoloti as well - could be some array going out of bounds? Or could also be the good old âbitshift without bracketsâ danger.
AH, thanks for the info. I should have tested it on the normal Ksoloti as well.
So I guess its down to something silly with the optimiser againâŚ
Getting a hunch that something in ALLPASS() is wrong, possibly the SINE2TINTERP() phase input causing the sine2t table (firmware) go out of bounds?
Maybe you could focus on the axoloti-factory objects first.. there is so much unchecked stuff in axoloti-community barely stumbling along
Edit: thread discussing a noise bug in v1.0.12 firmware.