EQ bandwidth question...

F

floppsybunny

New member
i have a four-band eq effect patch with variable Q of value: 0.5-10.0. i understand that higher values of 'Q' mean narrower bandwidths but i'm wondering if anyone knows (or if there is a standard) of how wide in KH is the bandwidth curve in the cases, for 0.5 and 10.0? or does it vary by effect to effect? (my effect does not provide me with these specs so i'm assuming there is some standard?).

also, does anyone know an approximate value of Q that would be used in standard 12-band 'graphic' equalizers? (i'm assuming it's a fairly narrow spread).

thanks!

f. bunny.
 
Q changes depending on the center frequency. It works in octaves, so it will be a spread of multiples and fractions of octaves.
 
Farview said:
Q changes depending on the center frequency. It works in octaves, so it will be a spread of multiples and fractions of octaves.
Click to expand...
This is true.
 
There is no standard as measured by frequency, because frequency is not measured on a linear scale. 10Hz is a wide range down inthe bass regions but is just a tiny speck of a distance up in the high frequencies.

Some EQs measure a band's bandwidth by octave instead of a number. There, obviously there is an inverse relationip to Q in that the narrower the bandwidth, the smaller the octave range value (e.g. a 0.5-octave spread is narrower than a 1-octave spread.)

However, if there is a common standard that relates Q width to octave width, I have yet to find one. Part of the problem is that the actual bandwidth changes with amount of gain. On any given EQ, the Q width at a gain of +12dB will be quite different than the width of the same Q setting at +2dB. I believe it's for this reason that there is no actual metric assigned to Q value; a Q value of 1 doesn't equate to "X octaves" or "Y kHz" because there is no direct equation that doesn't first have to take both the amplitude and the slope of the bell curve into account.

It's also important to remember that as a standard, "Q" defines not the band width of the entire bell curve from 0dB on one end to 0dB on the other, rather it only defines that central portion of the curve defined by the points on either side of the center frequency where the curve drops 3dB below the amount of gain at the center frequency. See post #9 in this thread for a graphic plainly illustrating this aspect of Q. That whole thread talks about Q as a mater of fact.

I know this gets really confusing, but I HTH anyway.

G.
 
As SouthSide said, higher Q numbers mean wider bandwidth. So in your example .5 would be a narrow bandwidth and 10 would be a wide bandwidth.

These things vary from unit to unit, so it really would be a bit inaccurate to say that a "Q" of 1 equals one octave. That's kind of the rule of thumb, but there are *a lot* of variables.

It's something you need to experiment with one the eq that you own, listening and learning so you can apply it in the future.
 
Jeezit's fuck! WRONG WRONG WRONG WRONG WRONG!

.5 is actually a WIDE bandwidth, and 10 is a VERY NARROW bandwidth.

Get it straight guys before you dole it out to newbs.
 
thanks guys, very interesting and helpful stuff

can anyone then give me a rough approximation of the Q in octaves in the 0.5-10.0 notation my patch uses? are we talking a difference of 1-10 octaves or 1-2 or 3? (i knew lower Q values meant a wider bandwidth just don't know the scale of values...)

thanks!

f.bunny
 
Actually what I said was that Q number and octave range had an inverse relationship; meaning that as Q value goes up, octave range (or bandwidth) goes down. A "high-Q" signal traditionally means a very tight bandwidth.

Flopsy, to the best of my knowledge there is no standard for converting Q value to bandwidth because the slope of the curve and the amplitude of the curve both have to be taken into account as well.

Once again, the"Q value" typically defines the area of the slope 3dB either side of the center frequency. The "width" of that central part of the bell curve depends upon the formula for the overall slope of the curve, which changes depending upon EQ type and upon the gain (unless it's what's called a "constant Q design".)

G.
 
floppsybunny said:
thanks guys, very interesting and helpful stuff

can anyone then give me a rough approximation of the Q in octaves in the 0.5-10.0 notation my patch uses? are we talking a difference of 1-10 octaves or 1-2 or 3? (i knew lower Q values meant a wider bandwidth just don't know the scale of values...)

thanks!

f.bunny
Click to expand...
It doesn't matter. Just listen.
 
Ford Van said:
No need to guess, or approximate anything, this is just science.

http://www.mhsoft.nl/QN.html
Click to expand...
But that still is only accurate when the gain is set at 3db or -3db. The bandwidth and the Q are measured at the 3db down point. If you add 10 db of gain, the bandwidth measurment stays the same, but you will be affecting a much larger bandwidth.
 
OK, that guy is using the folowing equation for converting bandwidth to Q value:

form.Qout.value = Math.round(10000*Math.pow(2,BWin*0.5)/(Math.pow(2,BWin)-1))/10000;

A pretty straightforward formula. Does that apply to a constant Q EQ only? If so (and I think that may be the case), it'll give inconsistant values for other types of EQ.

G.
 
Farview said:
But that still is only accurate when the gain is set at 3db or -3db. The bandwidth and the Q are measured at the 3db down point. If you add 10 db of gain, the bandwidth measurment stays the same, but you will be affecting a much larger bandwidth.
Click to expand...

Errrrrrrrr...well, it is a better representation than the manufactures of eq's give data for! ;)
 
Ford Van said:
Jeezit's fuck! WRONG WRONG WRONG WRONG WRONG!

.5 is actually a WIDE bandwidth, and 10 is a VERY NARROW bandwidth.
Click to expand...

Not the way it is labeled on my Orban 642B. Narrow is the small numbers and wide is the big numbers.

Same on my Klark-Teknik DN410, the small numbers are narrow bandwidth and the big numbers are wide bandwidth.

Most eq's are built so that you turn the knob left for narrower eq and right for wider, but I have one eq where it's the opposite of that.

So it really is as Farview says: it doesn't matter, just listen.
 
SonicAlbert said:
Not the way it is labeled on my Orban 642B. Narrow is the small numbers and wide is the big numbers.

Same on my Klark-Teknik DN410, the small numbers are narrow bandwidth and the big numbers are wide bandwidth.

Most eq's are built so that you turn the knob left for narrower eq and right for wider, but I have one eq where it's the opposite of that.

So it really is as Farview says: it doesn't matter, just listen.
Click to expand...

Well personally, I think the bandwidth should be represented like you say. .33 sure looks like 1/3 octave to me, and 5 could be 5 octaves. Fair enough!
 
Ford Van said:
Errrrrrrrr...well, it is a better representation than the manufactures of eq's give data for! ;)
Click to expand...
I wasn't arguing. I was just pointing out how it was measured.
 
floppsybunny said:
...also, does anyone know an approximate value of Q that would be used in standard 12-band 'graphic' equalizers? (i'm assuming it's a fairly narrow spread).
Click to expand...
I've noticed that all EQs aren't created equally (haha-big surprise!) and in the case of software EQs, maybe due to the FFT size or whatever, when you dial in a high Q and take a cut you sometimes just don't get what the knob indicates in db. In other words at 3kHz,Q=6, cut=-9db (maybe you're trying to kill a "sproink" in a mandolin or something...) but the EQ actually only cuts about 3 or 4 db. This can lead to all kinds of confusion when you compare notes with your buddies on forums... :D

This is one reason I like to haul out Voxengo SPAN (free VST) RTA to kinda 'calibrate' myself and the controls of a new EQ. You can also find how big a Q you need for a 12-band type setting on your particular EQ - I'd have to guess and fiddle around to figure it out. But you can shoot some pink noise or whatever you want thru it and determine some settings while watching the RTA. I'm assuming you don't have the center freqs aurally memorized - I don't!

If you were going for a 15-band or 2/3 octave spaced center freqs you could use this ISO chart I got out of the specs section of a 15band dbx manual online:
15 ISO Standard, 2/3 octave bands per channel from 25 Hz to 16 kHz: 25, 40, 63, 100, 160, 250, 400, 630, 1k, 1.6k, 2.5k, 4k, 6.3k, 10k, and 16k
Click to expand...

Or you could work it out from a 10-band 1 octave spaced EQ also. My comment is just that it's probably different on your particular EQ and depends if the EQ type uses constant-Q or whatever.
 
Why would you want to equate a parametric to the controls on a graphic?

It doesn't make any difference what the numbers say. Twist the knobs until it sounds right, then look at the numbers and write them down. Not the other way around.
 
Farview said:
Why would you want to equate a parametric to the controls on a graphic?

It doesn't make any difference what the numbers say. Twist the knobs until it sounds right, then look at the numbers and write them down. Not the other way around.
Click to expand...

Yeah. All this Q talk is making me hungry of Pad Thai noodles!
 
You must log in or register to reply here.
Back
Top