quick qay to perform multiple operations?

[Tronic]

another version to make spaced string
and added pad zeroes to number

[tester]

Thanks, that what I was looking for.

There is still (at least?) one error to track (non-zero f4 can't be between f2-f4; descriptor combining problem), but I start to check on this one later.

[tester]

I think I fixed it, using one of these "strange ways". But I've lost that cool formatting.

Generally, there is a strange issue in non-fixed version. TrogzTools shows, that separately - values and strings are outputed correctly, i.e. value, string, value, string, ... all indexes match. But when both threads go into module that rejects some entries - values start to shift, connecting with wrong descriptors. Tried various ways, none worked correctly (or I missed something). So if you know how to fix the previous version - help is appreciated.

Or reformatting this one?

//edit:

Okay, know how to back the cool formatting again. Issue and "strange fix" remains... and maybe it will stay that way.

[tester]

Hmm... There is a conceptual error. I see it's results, know it is there, and yet - don't see it. :-/

I would have to substract from input only these midpoints, which didn't originated from that input. Somehow. Tricky. Or block results, according to string similarities (input vs midpoint's beginning/end).

[tester]

I think I made it.

[tester]

Having created resonance calculator for midpoints, now it's time to focus on virtual harmonic interferences. It goes like this. If you have 200Hz, then 404Hz will intefere with virtual (not audibly present) 400Hz. But there is a catch. Each tone can be considered as a base for virtual harmonics (octaves) - or as physical harmonic (octave) of a virtual fundamental. So 200Hz will also interfere with 104Hz in similar way.

To avoid going too deep, I guess we can focus on maximum 4 parts (experimentally selected according to listening tests and mixing graphs). It will go like this:

[reedited]
1f -- 2f -- 3f -- 4f ==> "f" is a first in the series
1/2f -- f -- 3/2f -- 2f ==> "f" is second in the series
1/3f -- 2/3f -- f -- 4/3f ==> "f" is third in the series
1/4f -- 1/2f -- 3/4f -- f ==> "f" is fourth in the series

"f" is a base, everything else are virtual harmonic points of interference with the second tone.

Elliminating all duplicates, we will end with a list of 11 items, from which one will contain physical differential beats. All other will represent "h-points" around which vibration will happen. So in this step - each physical frequency should be compared with 11 items of all other frequencies.

p.s.: probably in above setting, there is at least one conceptual error. Or maybe not? This topic is foreign to me.

[tester]

Okay, let's correct my thinking...
Proper table for interval based resonances seem to look like this (after elimination of duplicates):

f -- 2f -- 3f -- 4f --> no correction for manual recalcs
1/2f -- 3/2f -- 5/2f -- 7/2f --> x1/2 for manual recalcs
1/3f -- 2/3f -- 4/3f -- 5/3f -- 7/3f -- 8/3f -- 10/3f -- 11/3f --> x1/3 for manual recalcs
1/4f -- 3/4f -- 5/4f -- 7/4f -- 9/4f -- 11/4f -- 13/4f -- 15/4f --> x1/4 for manual recalcs

"Manual recalc" means, that there is no simple adding of frequency to a h-point in order to get the vibe.

[tester]

Okay, I finished it, but... it appears that due to use of loops (for making arrays and comparing values between arrays) - it works extremly slow, especially at higher load and with bigger arrays. Can these loop based calculations be changed into something else, not triggered step-by-step?

Initially I ask about green things, because for ruby - I would have to point the whole calculation and filtering theme (I would not know what to do in ruby code).