Short-term and momentary loudness max measurements

[hiccup7]

I would really appreciate these additional features:

1. Short-term loudness max
2. Momentary loudness max

The first is the more important feature for me.

[csteinmetz1]

I will look into adding this soon. It should be fairly simple as it would only be a matter of removing the gating that is applied in integrated measurement, and then modifying the block size if I am not mistake. I am currently in the processing of improving some structure things so I will get to this after.

Thanks for your interest in pyloudnorm!

[csteinmetz1]

Looking to pick this up soon. Attaching the reference document EBU tech 3341 (see page 6). This will require modularizing some of the code that is in integrated_loudness(), which is a probably a good thing, since its quite large.

[fan455]

Hi, based on your implementation and BrechtDeMan's implementation (https://github.com/BrechtDeMan/loudness.py), I just implemented the short-term and momentary LUFS calculation. You can find it here: https://github.com/fan455/fan455-audio-processing-codes/blob/main/modules/loudness.py

I've tested the momentary LUFS calculation in this code using some test signals from the EBU tech 3341 (https://tech.ebu.ch/docs/tech/tech3341.pdf) and errors are within the tolerance range. I compared the results with the Youlean Loudness Meter's momentary LUFS and found no difference (may be not enough comparisons though). But it still needs more checking and optimization for better performance. However, the integrated LUFS calculation is significantly different compared to Youlean but close to pyloudnorm.

[alisonbma]

Loudness range (LRA) would also be a useful addition to the repo!

Essentia has the capability for momentary, short term, and LRA computation