Unoffical empeg BBS

Quick Links: Empeg FAQ | RioCar.Org | Hijack | BigDisk Builder | jEmplode | emphatic
Repairs: Repairs

Topic Options
#86992 - 12/04/2002 02:41 What do the Q Parameter numbers actually represnt?
CurlyKicker
member

Registered: 04/01/2002
Posts: 135
Loc: Orange County, CA
I've read the FAQ and understand what the Q Parameter is. But I'd like to know what the numbers represent. Would Q 0.50 represent a half an octive? Or maybe the ratio of db to frequency? What is it?

Thanks,

Curly
_________________________
http://www.ubernet.org/

Top
#86993 - 12/04/2002 05:46 Re: What do the Q Parameter numbers actually represnt? [Re: CurlyKicker]
genixia
Carpal Tunnel

Registered: 08/02/2002
Posts: 3411
IIRC, Q is defined as the center frequency divided by the bandwidth of the filter between the 3db points wrt to the center boost/cut.

_________________________
Mk2a 60GB Blue. Serial 030102962 sig.mp3: File Format not Valid.

Top
#86994 - 12/04/2002 09:07 Re: What do the Q Parameter numbers actually represnt? [Re: genixia]
mtempsch
pooh-bah

Registered: 02/06/2000
Posts: 1996
Loc: Gothenburg, Sweden
That is my recollection too...

/Michael
_________________________
/Michael

Top
#86995 - 12/04/2002 11:37 Re: What do the Q Parameter numbers actually represnt? [Re: mtempsch]
tfabris
carpal tunnel

Registered: 20/12/1999
Posts: 31563
Loc: Seattle, WA
I'm actually very interested in the answer to this question, too, as it directly affects my ability to use the RTA software to adjust my EQ.

If I wanted to program a fixed Q number into each of the empeg's default bands, what should they be set to?

My goal: To have a given band, when modified, affect only the neighboring frequencies, and only barely (if at all) affects the next band. For instance, my goal would be to set the Q so that the frequency modifications dovetail perfectly into each other, while using the default bands, like so:



From what I saw with the RTA program, this is not the case with the default "auto" q values. When the Q is automatic, it HEAVILY affects SEVERAL neighboring bands. This makes it very difficult to use the EQ in its default mode to make precise corrections.

For refrence, the default bands are nicely laid out in terms of the frequency spectrum. They go:

50 100 180 350 700 1300 2500 5000 10000 18000

The automatic Q works as follows:

0db: q=0.55
+/- 5db: q=0.98
+/- 10db: q=1.75
+/- 20db: q=5.53

It is interpolated as you adjust different levels of each band, you can see this on the screen. Dunno if the interpolation is linear between those datapoints, didn't do the math.

Interestingly, it seems to use the same Q adjustments for each band. Is that even right? Wouldn't the Q need to be different for each band, or are the default bands already spaced correctly for those auto-Q values to work properly?


Attachments
85503-q.gif (95 downloads)

_________________________
Tony Fabris

Top
#86996 - 12/04/2002 13:33 Re: What do the Q Parameter numbers actually represnt? [Re: tfabris]
tfabris
carpal tunnel

Registered: 20/12/1999
Posts: 31563
Loc: Seattle, WA
Okay, I've messed with it some more, and I see the problem.

In order for it to do what I want it to do, the Q has to change with distance from 0db. Hence the need for the auto-q feature.

The problem is that the auto-q feature isn't suiting my tastes well. To have it be only at 5 q when you've got -20db of gain, doesn't seem to be enough for the 10-band EQ. I haven't tried the 5-band EQ, maybe it's correct for that one.

It's all a question of tweaking it to what you need.

So can someone translate into english the explanation of the formula for Q given above?
_________________________
Tony Fabris

Top
#86997 - 12/04/2002 13:57 Re: What do the Q Parameter numbers actually represnt? [Re: tfabris]
mtempsch
pooh-bah

Registered: 02/06/2000
Posts: 1996
Loc: Gothenburg, Sweden
I'll try, though english isn't my first language

If genixia and I remember correctly it's like this:

Q=f_center/(f_high-f_low) <==> f_high-f_low=f_center/Q

With a Q of 1 and a center frequency of 1000Hz, the width of the affected band (>= -3db effect) will also be 1000Hz ( f_center/(f_high-f_low) = 1) -> f_high-f_low = 1000

With a Q of 5 you'd have an affected band that is 200Hz wide.
With a Q of 10 you'd have an affected band that is 100Hz wide.
With a Q of 0.5 you'd have an affected band that is 2000Hz wide (not sure how getting down towards DC affects things though).


Unfortunately I don't recall how the f_high and f_low (the frequencys where the effect is down 3db compared to the center frequency) are placed, but I don't think it's f_center ± (f_high-f_low)/2. IIRC, it's that f_center is the geometric average of the f_high and f_low (ie f_center=sqrt(f_high*f_low)), but I'm definately not certain on that point. If so, f_high can be written as f_low+f_center/Q so you get a 2nd power(?) equation to solve to find f_high and f_low.

Did that help any?

/Michael
_________________________
/Michael

Top
#86998 - 12/04/2002 14:04 Re: What do the Q Parameter numbers actually represnt? [Re: mtempsch]
tfabris
carpal tunnel

Registered: 20/12/1999
Posts: 31563
Loc: Seattle, WA
/me moves hand over the top of his head and makes a whooshing noise.
_________________________
Tony Fabris

Top
#86999 - 12/04/2002 14:07 Re: What do the Q Parameter numbers actually represnt? [Re: tfabris]
mtempsch
pooh-bah

Registered: 02/06/2000
Posts: 1996
Loc: Gothenburg, Sweden
Sorry...

/Michael
_________________________
/Michael

Top
#87000 - 13/04/2002 00:38 Re: What do the Q Parameter numbers actually represnt? [Re: mtempsch]
CurlyKicker
member

Registered: 04/01/2002
Posts: 135
Loc: Orange County, CA
In reply to:

With a Q of 5 you'd have an affected band that is 200Hz wide.
With a Q of 10 you'd have an affected band that is 100Hz wide.
With a Q of 0.5 you'd have an affected band that is 2000Hz wide (not sure how getting down towards DC affects things though).

Did that help any?




That helps. Now I can get an idea of how much Q I should use and what I should be listening for.

Thanks,

Curly
_________________________
http://www.ubernet.org/

Top
#87001 - 13/04/2002 03:19 Re: What do the Q Parameter numbers actually represnt? [Re: CurlyKicker]
mtempsch
pooh-bah

Registered: 02/06/2000
Posts: 1996
Loc: Gothenburg, Sweden
Remember, this is the -3db point - you'll still effect frequencies above and below that band, but not as much...

And it's for a center frequency of 1000 Hz. Move up to 10 kHz and the width of the effected band will also widen with a factor 10.

/Michael
_________________________
/Michael

Top