PRO-AUDIO PLUG-INS
You are not logged in.
Hi Gary,
I'd like to discuss the High Pass (or Low Cut, the term that's always made more sense to me). Mainly the Q curves of the section.
Using the principle of the Q curves on the bands, that a small q value is a wide bell and a large q value a narrow bell (fig 1) I'd assume the cut curves would be the same (fig 2)
however when I study the effects I get something like in Fig 3, a small q value produces a tight curve that actually boosts the freq at the value and a large q value is a wide curve that starts to cut the frequency higher up the spectrum than the set value.
Fig 4 is how I would expect the section to operate. Note that in fig 3 the frequency is not completely cut also.
I remember an older version of the EQ had a fixed curve with just a variable freq, in many ways I think this would be easier to operate and give the plugin a very nice gui simplicity (I always find concentric controls slightly finicky (although you've done a good job with the vintage EQ ;-) ) though perhaps it would be a step back maybe a legend like the one I've done in fig 2 would be useful on the section to make it visually understandable.
Hope this helps your continued improvements of some great products.
If you'd prefer I emailed you directly with thoughts like this please let me know.
Best,
Tom
Offline
Thanks for the feedback Tom.
The Q of the High Pass (the correct term for this type of filter) does currently work in the the reverse way to the peak filters, for consistency, i'll flip the control values.
The Highpass does cut to the extent off its slope see graph. Tighter Qs do result in a resonant peak occuring, which i like, i could get rid of it by reducing the Q range, but i think it would make it less versatile.
Offline
Is that screenshot after you've flipped the Q values?
I can see that the curve starts sloping at the frequency, do you think it would be better to keep the selected frequency intact and have a tighter slope, "harder knee"?
the resonant thing is a bit unexpected, maybe more suited to the mid range bands? I'd expect a low cut to be more clinical about it's job, but I see your point. associating it with a tighter curve would make more sense. I think flipping the q is a good thing.
what do you think about adding the legend, if so I think it should actually be more like;
Offline
The screen is from the current version.
The slope is currently 12bB per octave, i'm happy with that as is sound wise.
Re: the Q, i didn't mean resonance as in a moog filter where you are purposely implementing and accentuating it, merely that when you get to tight Q settings you naturally expect it to get more resonant, Peak Q as well.
legend: along those lines, but with curvey lines to match Peak Q style, the straight lines look more like the filter styles.
Offline
Ok I understand, I supposed I'm more used to a traditional mid-range mixing desk's fixed low cut button (on mine it is 80Hz).
Great stuff Gary, all the best
Offline
A question came up when I was having an engineering based conversation recently, is a Low Cut actually different from a High Pass? or is it a term not officially used?
Offline
"Low Cut" doesn't describe the actual type of filter being employed just the net result, it could be a low shelf, a high pass even a very wide band pass filter (though unlikely).
Offline
Just to stick my oar in one last time, from discussing this with someone on a current project it is a bit odd that a HP set at a frequency would start filtering above the frequency selected (ie the frequency you want to be "passing") and the resonance thing is a bit peculiar when you are trying to roll off bass not accentuate it, which you'd be using the bands for.
What is the tightest Q value before the resonance appears? I think this is what I'd like to keep it set at.
best
ps are you planning on flipping the Q and adding some legend in an update?
Offline
There is always a transition with filters starting either side of the frequency set, this gets smaller as the order of the filter increases, but usually sounds less "natural" or analogue as the order increases. Q at 1.3 gives the smallest transition width for the HP. The "bump" at high Qs is useful in a variety of situations, bass drums in particular, and as its optional i don't see the problem.
Flipping the HP Q and adjusting the legend will be in the next update, I may look at bumping the order of the HP up a bit to increase the steepness of the rolloff also.
--
Here you can see the transition for 500hz:
and 1khz:
currently Higher Q (control wise) gives a wider transition
currently lower Q (control wise) gives a smaller transition
Offline
Thanks Gary, that's all really interesting, especially to see the graphs, the gradient is a lot less steep than I'd assume for some reason, and doesn't change with the q value as I thought, but I'm no audio engineer, I closer represent the man on the street in this situation, and I'm big on interface design. So as long as it was expected I think it would be an interesting trait of the plugin, I don't want to be a party pooper about a potentially really interesting effect. maybe the resonant bump could be illustrated in the legend so it gets used more knowingly. Hope you don't mind these analities. Using the EQ and COMP on all my sessions with great enjoyment.
ps was my "fig 4" an inappropriate suggestion for a HP filter in a different case?
Last edited by tomv-k (2010-08-07 02:56:46)
Offline
Your fig4 would need the order of the filter changing rather than Q, sometimes you see slope
options in Digital Eqs ie: 12db, 24db etc. The highest setting off the Q does change
the slope as the transiton widens and flatttens, see last my post 3rd pic.
Offline
Ah ok, but the bell shapes of the mid frequency bands' slopes change steepness with changing the Q? I assumed it would be like a one sided bell shape.
Offline
HP filters work differently to Peak filters, Q can't give you a steeper slope only a slighty less steep one as it it flattens out. You need to up the filter order to get a steeper cut-off, which also increases cpu usage. Don't take my word for it though, do some googling on analogue filter types, lots of info out there.
Offline
I've noticed that cutting the LF is much more what I would expect the HP to be doing, at fully -20dB is cuts sub frequencies much more profoundly than the HP section, which now makes sense with the wide slope it you've shown in your graphs. I know I'm banging on about this but it just doesn't seem to be doing the right thing, you want to get rid of sub frequencies but it's not as mean as it should be perhaps?
It prompted me to check what the values of my mixing desk is and I found this graph in the manual
Which is what I kinda thought, that the bands were more gradual and musical whereas the HP is more clinical and extreme. what do you think on this Gary? Having a moveable Freq is completely great but a steep slope seems to make more sense?
best
Offline
As mentioned, I will look at increasing the HP slope in the next update from 12dB to 24dB per octave.
Offline
Hi Gary, great stuff, just noticed from your graphs that the resonant peak on full whack is 10dB? crazy, is that modelled on any piece of analogue gear?
I get it now re order of filter, that's what I thought the knob was doing basically, would I be in the minority there?
Just noticed from an old forum entry that the hp slope used to be 24db/octave at one point, so it got changed?
best
T
Offline
tomv-k wrote:
Hi Gary, great stuff, just noticed from your graphs that the resonant peak on full whack is 10dB? crazy, is that modelled on any piece of analogue gear?
This Q behavior is intrinsic to this kind of digital HP filter. I could limit the Q range more, but if you find it "crazy" at the tightest setting its easy to avoid that setting. I actually like the ability to do that, it allows simultaneous boost and low end cut just below the target frequency.
tomv-k wrote:
I get it now re order of filter, that's what I thought the knob was doing basically, would I be in the minority there?
Its always been a Q control.
tomv-k wrote:
Just noticed from an old forum entry that the hp slope used to be 24db/octave at one point, so it got changed?
It's always had a slope of 12dB.
If 24dB was mentioned it was a typo.
Offline
Yeah I do think it is quite interesting once you know to use it creatively, just hadn't realised it was 10dB, crazy is cool.
Offline
I have to say Gary that now I know how to use it the resonant bump on the HF is a rather cool feature.
Looking forward to the update.
Offline