claritychallenge / clarity

Clarity Challenge toolkit - software for building Clarity Challenge systems
https://claritychallenge.github.io/clarity
MIT License
130 stars 54 forks source link

Audiograms - this exists as an audiogram class but is not being used consistently in the code base #277

Closed groadabike closed 1 year ago

groadabike commented 1 year ago

@jonbarker68,

Is the clarity/enhancer/gha/ the best place for the class Audiograms? If we want to use it consistently in Cadenza/Clarity recipes, it shouldn't be part of the enhancer/gha as it is not only for camfit.

I feel that this class fit better in a different location like clarity/utils/commons or clarity/commons or similar.

What do you think?

jonbarker68 commented 1 year ago

@jonbarker68,

Is the clarity/enhancer/gha/ the best place for the class Audiograms? If we want to use it consistently in Cadenza/Clarity recipes, it shouldn't be part of the enhancer/gha as it is not only for camfit.

I feel that this class fit better in a different location like clarity/utils/commons or clarity/commons or similar.

What do you think?

I agree. It's actually defined in a couple of places enhancer/gha and evaluator/msbg. The plan will be make a single class that they both use and that is placed somewhere more obvious in the module hierarchy.

It needs a bit of careful thought though: some functions are binaural and expect an audiometric threshold for a left and right ear, others are monaural. The simplest option would be that audiograms are always for one ear, and that binarual functions need a pair of audiogram objects. But that doesn't capture that binaural audiograms have a shared frequency axis but it would keep things simple and in principle, you could have different frequency measurements made in each ear.

You'll see this issue quite clearly if you compare the current enhancer/gha and evaluator/msbg designs.