AutoEQ incorrectly using LS instead of LSC

[coda101]

Describe the bug When comparing an EQ result that only use peaking filters to that of AutoEQ precomputed results (that use shelve filters) the sound of either is not identical. This is because AutoEQ is incorrectly labelling the filters as an "LS" instead of "LSC".

To Reproduce Use crin graph or squig to create a Harman EQ for example, and compare the resulting sound with AutoEQ's precomputed result. You can use only peaking filters if you want or a bass shelf because the default will be LSC for the online graph tool.

Expected behavior The bass should be the same, but it sounds different (fuller, in my case, because AutoEQ was boosting mid bass too much).

[coda101]

I think it may be more noticeable with high gain shelf filters. Low gain seems indistinguishable.

[coda101]

Also using HS instead of HSC

[jaakkopasanen]

Fixed! Thanks a lot for pointing this out to me!

[bhack]

@wwmm Is this going to break easyeffects APO parsing?

[wwmm]

Is this going to break easyeffects APO parsing?

@Digitalone1 was direct involved in the APO preset parsing implementation and will be able to give a better answer. But in a quick look at the code my guess is that if the label was wrong in the preset file its corresponding value will go the the wrong place. In other words the parsing will work but the value will not be saved where the person that created the preset intended because the label is wrong.

[wwmm]

Forget it. I think that our parsing is handling these labels in the same way. It is better to wait for @Digitalone1 answer.

[Digitalone1]

@wwmm Is this going to break easyeffects APO parsing?

I'm not an expert of audio filters, anyway looking at the code, LCS will apply Low-shelf filter.

https://github.com/wwmm/easyeffects/blob/master/src/equalizer_ui.cpp#L42

Then it will parse the gain and, if the quality is not present, it's set as 2/3 float:

https://github.com/wwmm/easyeffects/blob/master/src/equalizer_ui.cpp#L248-L253

Don't know if this is what you want, I hope it's enough. The code was imported from LSP equalizer plugin.

[wwmm]

Thanks @Digitalone1. I had forgotten we are doing the same as LSP. So it is probably fine. I just wonder why LS and LSC are both mapped to low shelf. @bhack what is the difference between them? I have never used APO.

[jaakkopasanen]

LS is a fixed slope low-shelf filter and is approximately 0.67 Q. LSC is the one where you can control Q. Using LS instead of LSC was simply a mistake on my part.

[bhack]

It seems that we have an extra parameter but I don't know if we have it in LSP equalizer. See the filter table at https://sourceforge.net/p/equalizerapo/wiki/Configuration%20reference/

There was also an old thread on reddit/oratory1990: https://www.reddit.com/r/oratory1990/comments/qkhunz/ls_or_lsc_setting_in_config/

[bhack]

LS is a fixed slope low-shelf filter and is approximately 0.67 Q. LSC is the one where you can control Q. Using LS instead of LSC was simply a mistake on my part.

Yes it is the extra Q so if we don't have a problem mapping LCS/HCS string and its Q the parser is working ok.