RFE: Visualization of EQ frequency response

[vmatare]

Hey, I have an idea for a nice feature:

The EQ could really benefit from a visualization of the expected frequency response. Hz on the X axis and DB offset on Y.

So the background is I have a bi-amping setup which gives plenty of headroom in the bass range, but the limiting factor on getting an awesome bass out of it are the normal modes of my living room.

There is a bunch of those, and I can identify their peak frequency okayish, but I'd like to be more precise when trying to cout them out. Currently, I search through the frequencies with a sine wave and listen for where it gets out of hand. Try to smooth it out and then listen to the surrounding frequency range, find gaps, tune again and so forth. Slightly tedious.

So I think it might be pretty helpful to at least see what I'm changing, because listening takes much more time, and people also tend to get seasick from those sine waves.

I haven't tried "tuning" the room with a mic yet, but it's definitely something I'd like to try later on. I have a feeling pulseeffects can help me with that as well...

[wwmm]

I would like this too. But as you can see in #515 it is hard to get that when you do not use the native plugin graphical interface

[vmatare]

Oh yeah, I imagined the function described by the equalizer might be hard to find/model. So I thought, why not just measure it? Sample a sine through the spectrum and record the difference between input and output volume, bam. Should be possible to do this within a few seconds, maybe even less than one if the FFT and whatever magic is going on there converges quickly enough...

[vmatare]

Maybe it works just as well (or even better) with noise. That should be much faster. Oh, and you can even run it with full CPU power if you make sure it doesn't depend on a natural-time clock.

[vmatare]

Oh, this is an interesting read: https://lsp-plug.in/?page=manuals&section=para_equalizer_x32_lr
So we could even measure the EQ's quality in terms of impulse response... o_O
... And it can be done on live data O_o

[lee-tts]

@vmatare Did you manage to tame your room modes finally? I am here for this exact reason myself, and have one that's especially bad at 116Hz that I wanted to squash with pulseeffects. I realize that in the current version it's no longer possible to change the bandwidth / width of the EQ like in the past. Pity because my room has pretty sharp spikes and the EQ is a bit wide on a given parameter.

[wwmm]

@lee-tts you can change the bandwidth. You even have more filters available like notch filters

[vmatare]

Yeah, you now have this "Quality" control which affects the bandwidth. But to answer you question @lee-tts, I wouldn't say I've tamed anything finally. I have this particularly bad resonance around 85 Hz. The EQ band at that spot is set to -21.6 dB, but that's for modern mixes. Seems to me like old mixes (like music from 60s to 70s) anticipated more bass resonance. Anyways, there are other resonances as well, and while I feel I have cut out the worst of it, but have created some holes in the process. I.e. I'm cutting out too wide, and in other low frequency ranges I still notice minor resonances and gaps now and then.

So yeah, definitely far from perfect, but at least I'm getting some of that impressive ultra-low bass you can have nowadays.

[lee-tts]

Thanks guys.

So the quality somehow affects the bandwidth? I played with it a bit initially but the 'width' written in the middle of the dialog didn't change so I moved on. Are these documented somewhere? The help says Quality is the "Quality factor." so I'm not sure what that means. It also seems to refer to the old width control. I'd be happy to make a PR to update that help page :)

I've got one parameter set to 116Hz and lowering it more than about 6dB basically scoops all the bass in the music I'm listening to (classic rock, metal from all eras -- there's some classic rock that has plenty of 116Hz haha that Bb2 is deadly)

[wwmm]

You have a good reading here https://en.wikipedia.org/wiki/Q_factor. Long story short bandwidth = center_frequency / quality_factor

[lee-tts]

@wwmm Got it! Thank you. Not sure why I didn't see the width number changing before. Now I see that it does change when you adjust quality.