Debra Peri

I searched the forums but couldn't find a solution for the

 errors. In my particular case the error is being generated by cycle~ not being happy with the frequency value I'm piping to it.

cycle~: cannot set frequency: number is not finite

What are the best practices for dealing with these errors?

Hello,

I searched the forums but couldn't find a solution for the "number is not finite" errors.   In my particular case the error is being generated by cycle~ not being happy with the frequency value I'm piping to it.  

cycle~: cannot set frequency: number is not finite 


What are the best practices for dealing with these errors? 

Hello,

I searched the forums but couldn't find a solution for the "number is not finite" errors.   In my particular case the error is being

handling-"number-is-not-finite"-errors

first try to avoid that it happens. it more often happens due to calculations you perfomed yourself than because of a buggy external.

then there should be objects which can filter it out. (percolate´s normalize? an apple silicon version is on the git now.)


first try to avoid that it happens. it more often happens due to calculations you perfomed yourself than because of a buggy external.

then there should be objects which can filter it out. (percolate´s normalize? an apple silicon version is on the git now.)


Excuse me, but why would you assume the calculations are from an external?

And, I'm not going to build something from github that can be, should be easily solved inside of Max. 

Excuse me, but why would you assume the calculations are from an external?   

And, I'm not going to build something from github that can be, should be easily solved inside of Max. 



because i have seen externals which do this. i can even show you commercial VST plug-ins which send denormal data or inf to the host.

if you caused infinite or -infinite yourself with some kind of nested higher math functions, then it should be more straight forward to avoid them than to fix it using a filter. 

a simple clip() or max() at the right position usually gets rid of it.

you might ask why cos~ or expr are not filtering it themselves?... i have heard it is a convention to allow that kind of output from compiled code.


because i have seen externals which do this. i can even show you commercial VST plug-ins which send denormal data or inf to the host.

if you caused infinite or -infinite yourself with some kind of nested higher math functions, then it should be more straight forward to avoid them than to fix it using a filter. 

a simple clip() or max() at the right position usually gets rid of it.

you might ask why cos~ or expr are not filtering it themselves?... i have heard it is a convention to allow that kind of output from compiled code.


The calculations were done internal to Max. I found where the error is being generated, 2 math ops (* and +) that was creating the non-finite number before it was piped into the cycle~. Once I found that calculation I used a clip object.

So really, all you needed to do was offer the suggestion of a clip or min/max object. Which by the way, I had already used but in the wrong location in the calculation rig. 


The calculations were done internal to Max.   I found where the error is being generated, 2 math ops (* and +) that was creating the non-finite number before it was piped into the cycle~.  Once I found that calculation I used a clip object.  

So really, all you needed to do was offer the suggestion of a clip or min/max object.  Which by the way, I had already used but in the wrong location in the calculation rig. 


Handling "number is not finite" errors