midiratio could be more precise

[elgiano]

Not a big issue but sometimes annoying, especially in obvious cases:

12.midiratio // -> 1.9999999999945
2.pow(12/12) // -> 2.0 

When I think about Henry Partch (roughly) saying that equal temperament is a lie compared to just intonation, this gets particularly annoying :)

Looking at the source code, it seems like midicps/cpsmidi use more precise constants than midiratio/cpsoct... anyone know if this is a bug or intended for some reason I can't understand?

supercollider/include/plugin_interface/SC_InlineUnaryOp.h

Line 283 in f806ace

return (float64)440. * std::pow((float64)2., (note - (float64)69.) * (float64)0.08333333333333333333333333);

supercollider/include/plugin_interface/SC_InlineUnaryOp.h

Line 290 in f806ace

inline float64 sc_midiratio(float64 midi) { return std::pow((float64)2., midi * (float64)0.083333333333); }

Environment

• SuperCollider version: 3.11.1
• Operating system: Arch Linux

Steps to reproduce

12.midiratio // -> 1.9999999999945
2.pow(12/12) // -> 2.0 

Expected vs. actual behavior

I would expect 12.midiratio to be exactly 2.0

[mossheim]

this looks like a bug to me! possibly just missed in 4249e321. that other code is from the initial commit.

however, this stuff is in the public plugin interface (include/plugin_interface), which means we need to be a bit careful about how to fix things there...

[simonperrier11]

Just came across the same issue in 3.13.0, macOS 13.5.2