Equalizers question

Good people,

What is the difference in concept, use, and results between "parametric equalizer" and "dynamic equalizer"?

I also have found "Graphic equalizers" that resemble what I am more familiar with from hi-fi use...So are those dynamic or parametric? I mean those that have a set of 7 or so sliders that one moves up and down and can immediately hear how they affect the audio.

Thanks!

A dynamic EQ modifies its filters automatically in response to the spectral content, so that filtering adapts on the fly to the source material. They can be tricky to use, since it's hard to predict how they'll respond moment to moment.

GlissEQ is unlike other dynamic EQs in that it has a time delay built in before the dynamic modification kicks in. The concept is that it helps preserve transients. Personally, I turn that feature off for lower frequencies and only use the dynamic setting for highs.

While you're collecting terms, here's another one for you: paragraphic equalizer. Fancy name for a parametric equalizer with a spectral display overlaid. Like GlissEQ and Ozone.

On the subject of graphic equalizers, they are laid out that way (with fixed bands) to make it easier to quickly guess which slider to hit when feedback starts onstage. They are almost indispensable for live sound, but have no place in a recording studio because they are not flexible enough to serve any practical purpose.

Another place dynamic equalisers turn up is in the Behringer Ultra Curve Pro Mastering processor. What goes on there is that they split the spectum into three bands and you have compressor like controls. When certain thresholds are met within these bands in level, then the sound of the equaliser can ease its way into the sound. You might have this scoop out some mid range when certain levels are met because at some point the sounds starts to sound excessive mid range wise and the dynamic eq can cut in and pull things back.

A multiband compressoer can do a similar thing except there is more control over how much shift or attenuation there is after the threshold is met with the dynamic eq.

Likewise you can get it to boost as well when a certain threshold is also met. This is handy when you want to add some eq and push things a bit more after a threshold is met.  ie you may want the bottom end and top end to just nudge up nicely once a certain threshold is met.

It is quite complex to program and you have to have a real sense of what you are editing in terms of its parametrs.

Obviously one has to do this sort of thing well below 0 db FS to allow for all these increases in levels here and there in the spectrum. You can still follow it with a compressor as well and that will even out the overall level shifts but you will still get the dynamic change of EQ.

The Behringer is obviously hardware but can easily patch into the digital signal chain. I have only used it when I needed something special to happen EQ wise when a certain level was exceeded and that was pretty high up as well. It did it well.

What Dave is talking about is also correct. Some dynamic eq's respond according to the spectral content of the material. There is more than one type. The level sensitive ones work better on dynamic tracks or mixes. ie classical music. Music that is not very dynamic does not make them do much.

Graphic EQ's can have a sound like anything else with analog inputs and outputs. I used to have this Fostex EQ which just added a smooth sound all of its own even with all the sliders flat. So I see that as a useful tool. It made drum overheads sound warm and smooth. It was distorting things I imagine as it was not expensive but it must have been nice distortion. I also have an expensive Yamaha 31 Band stereo EQ and it too sounds good.

It can be quite handy for things like taking a mono signal and creating a stereo image from it. Patch a mono signal into both inputs of both stero halves. Set all the sliders opposite to each other top and bottom. A good way to split the spectrum on L and R individually. I have always found graphic sliders easy to read as they represent the frequency response. Join up imaginary sliders to create the impression of the response. Any EQ that shows off the response so well is not all bad.

Easy to do variations with the stereo effect creation too. eg 3 up 3 down etc..

Years ago, when I had almost no budget for gear, I bought my own $800 Ashly something or other 31 band EQ... the fancy one... to use in my personnel swiss army knife fix it kit.

The Ashly was the cheapest model that had full bypass when each slider hit the mid point indent.

That was considered good... in a world where NO GRAPHIC EQUALIZER was ever regarded as good... they all sucked... well, maybe not the Clark Technik 360... they were tools used to do things you had to do that could be done in other ways but usually weren't because of time constraints.

Having said that... it's hard to think of any good reason to use a digital version of something that most people wished they didn't ever have to use.

The multi node parametric EQs made today are what we always wanted.

That's all just my opinion... but if can modestly suggest... I am expert at operating a 31 band EQ... and I never liked having to patch one into the chain... although I was always glad to have them when it was a time when I needed one.

One thing to consider when comparing either GUI layout is too keep your comparisons matched. In other words if the graphic is +/- 12dB you want to know that so the parametric is set to a similar level. The same goes for Q settings. A cheap graphic has a wide Q. A great graphic has a variable Q that narrows as you increase the level change. Today's parametric designs let you dial in the Q. Make sure you are comparing like settings.

Frankly most analysis of EQ dsp indicates it is all remarkably similar unless there is something very special like phase linear math, the dynamics we have just talked about, or an IR of some actual hardware device hidden in the mix.

Just some stuff to think about.

best regards,
mike



edit spelling

Cjay,

You may want to elaborate on the description of the graphic EQ.

A cheap 31 band graphic usually does NOT have a true 1/3 octave Q even under the best of circumstances. There's lots of over lap which creates all kinds of side effects when you use adjacent controls. The wider range of effect and the summing of adjacent ranges makes it easier to hear some change when make an adjustment... but it's rarely the change you think you made.

A top of the line modern 31 band graphic EQ has a self adjusting Q that actually narrows as you adjust the levels further up or down so that they DO effectively remain confined to 1/3 octave and so the controls more or less effect what they appear to be labeled for. This makes them more useful to someone who is listening for a subtle effect when problem solving.

Is that your website? That's a lot of writing... well done.

very best,
mike

Good writeup CJ.

Thanks for linking it too.

"The multi node parametric EQs made today are what we always wanted."


When you arrive at the conclusion that almost all dsp EQ sound the same (unless purposefully made to be different in some way) because the math is practically standardized you are left with an opportunity to realize that a parametric with graphics EQ offers so much more control and opportunity for understanding if you adjust EQ while listening carefully.

I have observed a lot of people operate a EQ in the absence of listening... they just use a "system" or conventions they adopted after learning of some other persons habits.


If one considers my explanation about the nature of an inexpensive outboard hardware 31 band EQ...  one should be concerned that the knowledge they think they are gaining by having a detailed display on the face of the unit may be tainted by the very fact that the EQs rarely respond as the "graphic" might suggest.

In other words, you may think you know are cutting the lows 3dB by pulling down each of a few sliders on the left side of the unit to a place where the labels say -3dB.... but in reality you have probably summed a bunch of adjacent overlaps and actually pulled the wide band down -12dB... or something like that.  Consider that if you grab a few adjacent sliders in the mids and nudge them... the effect is greater than the labeling suggests... it may sound good... but is the "graphic" really helpful when it is so misleading?

It's not valuable knowledge... unless you really know what you did.

Grab a Klark Technik 360 and check it out... they ARE surgical and let you do what you think you are doing... assuming the caps are fresh.

My dad, the EE, use to get livid every time I used the term "graphic" EQ... he'd mutter "there's nothing graphic about that thing"... of course my dad never got to drive a modern Ashly or a Klark Technik 360. It took me a long time to under stand that his frustration was the fact that I might be deluding myself by thinking the "graphic" on a cheap EQ represented reality.

So, in summary... today's para graphic EQs can sound exactly like ANY graphic EQ ever made and are incredibly useful as a learning tool... especially if you really want to correlate your listening ear to the actual changes you are introducing into the signal flow.

When you learn to adjust a parametric graphic EQ to match that sound you like on the 31 band EQ you'll possess a more accurate understanding of what you did to get the sound you want.

I know this seems pedantic... but I learned hard lessons down in the pits... it left a strong impression... and I want to  pass on the benefits.

all the best,
mike


edit for spelling and clarity

Both of these filters are known as Paragraphic equalizers.

Which one is more graphically representative of the way you hear sound?

IMO, the newer one is much more graphically accurate... and there fore more useful as a tool for gaining experience and understanding.

The old kind offers a very crude graphical representation.






That's why I say the modern paremetric-graphic EQ is what we always wanted.


I would like to add that the reason a top flight hardware analog EQ costs so much is that it attempts to be accurate with regards to it's labeling.

To achieve accuracy the builder needs to source capacitors that are both consistent and accurately labeled.

It's very expensive to test and sort capacitors down to the low percentage tolerances required to ensure accuracy.

When assembling a EQ the labels on the front are only accurate when the capacitors actually match the design values. It's really expensive when you need to be so picky.

Most hardware EQs are kinda sorta accurate... so the idea that the graphic on a front panel is accurate enough to be useful for learning about sound is undermined by the practicalities of the situation.

all the best,
mike