Add a perceptual loudness model, which should implement A-weighting and possibly other models (e.g. ISO 532)
(following #128)
ideoforms
