Iver Jordal

What is the motivation for adding this transform? The effect would be pretty much the same as applying a low-pass filter, right?

I personally just apply a LPF to emulate such behavior. Doing a downsampling (including the usual LPF that is involved in that process) and then an upsampling requires more compute,...

I guess it should randomize at least these things: * the difference between min gain and max gain * "the wideness" of the gain curve (think like: how fast does...

Great! Done

I guess the most important thing is the difference between the minimum gain and the maximum gain. In the future I want to establish a common post-gain pattern that can...

This needs default parameters, so I can do `my_transform = GainPeak()` and start using it just like that Also, I tried to illustrate some examples of gain curves (green) that...

Yeah, take the gain curve with a grain of salt, especially the min gain part :P I guess min gain isn't the most important feature here, but the difference between...

Yeah, the offset can be selected randomly

Sweet, thanks 👍

Good question! Let's consider the possibilities: 1. `Shift` and then `GainPeak`. In this case the peak will always be centered, so that's maybe not so realistic, and a ML model...