antiAliasFilter is giving -nand about rs-met

[elanhickler]

ill keep testing, not sure if it's happening after I reset the Spiral Generator.

[RobinSchmidt]

whoa! that's weird! what are the filter coefficients (a- and b- arrays)? what is the cutoff freq (i.e. subdivision)? how can i reproduce this?

[elanhickler]

you'll need my specific RS-MET code......... maybe tomorrow I will make a separate spiral generator repository and put my RS-MET code in there.

[elanhickler]

[RobinSchmidt]

that actually shouldn't make a difference (unless you edited the filter class). hmm - so it looks like the y2 value is already nan. i need to look at this tomorrow. maybe i forgot an initialization. i again notice that i should really switch to use std::vector instead of raw c-style arrays. then, we could look at the whole array in the debugger. ...these are things one only learns through experience. i don't want to use raw - c-arrays anymore. or juce::Array, for that matter. i say that, so don't have to make my mistakes again :-)

[elanhickler]

I know where the nan is coming from. It's coming from asin() where anything above 1 or below -1 results in NaN... from the Spiral Generator code... but... how do I solve this without do a limiter....

[elanhickler]

got it. I debugging by doing a conditional break if tmp != tmp (which catches -nand)

Then I "went up stream" until I found the first occurance of a bad value (-inf and nand), and it lead me to a value in spiral generator.