EQ Challenge

[NOYB]

Who can come up with the flattest 1/3 octave smoothed full range EQ for the attached measurement, using only a 2/3 octave graphic EQ with the following 15 center frequencies.

25, 40, 63, 100, 160, 250, 400, 630, 1k, 1.6k, 2.5k, 4k, 6.3k, 10k, 16k

[Wayne A. Pflughaupt]

Moving to BFD/Equalization Devices Forum.

Having trouble opening the file. Can you give us a jpg of the graph?

Regards,
Wayne

[brucek]

using only a 2/3 octave graphic eq with the following 15 center frequencies.

I don't know if the response is bad enough to apply any EQ. There's not a lot wrong with it.



The problem is that you don't offer information on how the graphics equalizer calculates bandwidth (i.e -3dB down, half gain points.......? ).

As an example of unusual, some EQ's like the DSP1124P, it defines bandwidth as:

Bandwidth (Hz) = centre frequency*(BW/60)*sqrt(2)

So, the Q formula becomes: Q = 60/[(BW/60)*sqrt(2)].

But, if you assume a standard definition, then 2/3 octave BandWidth would be a Q of ~2.1.

Then enter all the filters center frequencies into REW and set the Q to 2.1 (select the generic equalizer) and then play with the gains until the mdat file looked good. Then you have to assume the gains translate exactly to the type of graphics equalizer you have.

Below is what I'm talking about. I attached the filter.req file (with all the gains at zero) if you just wanted to load it and play.



Here's a fairly common definition of Bandwidth. This is how I calculate the Q of 2.1 to enter into the REW generic equalizer.



You could also just measure the equalizer with REW and establish the relationship between REW filters and the graphics equalizer filters, then you would be better prepared to setup the filters.

brucek

[Wayne A. Pflughaupt]

Thanks for providing the picture, brucek.

NYOB, to get an idea of what brucek is talking about...

The problem is that you don't offer information on how the graphics equalizer calculates bandwidth (i.e -3dB down, half gain points.......? )

...open up REW and select the BFD as the equalizer, and dial in a filter say, at 30 Hz with a 1/3-octave (20/60) bandwidth with a 6 dB boost or cut. Then switch the equalizer selection to either the FBQ or a generic PEQ. You'll notice that the bandwidth setting for either of those equalizers is 3/4-octave. So, what the BFD "calls" a 1/3-octave filter is a 3/4-octave filter on other EQs. Startling, huh?

Like brucek said, your graph looks pretty good. I think it could use some help in a few areas, though - the peak at ~500 Hz, and the depressions at 2 kHz and 5.5 kHz.

The problem is going to be your equalizer. A 2/3-octave EQ is going to affect that wide on both sides of the filter center, so it's cutting a path 1-1/3 octaves wide. That's pretty broad, so there isn't much you can do to address any of those three problem areas because they are much narrower than that. On top of that, none of the problem frequency centers line up with your available filter centers. IOW, the EQ will only mess up areas beyond the problematic sections, that don't need any attention.

By comparison, all three of those anomalies I mentioned would be an easy fix for a parametric EQ. For instance, the 500 Hz problem is about an octave wide, so a 1/2-octave filter would excise it nicely. The 2 kHz problem is also about 1-octave wide, so another 1/2-octave filter there. The 5.5 kHz dip is a bit less than 2-octaves wide, so something like a 3/4-octave filter would be a good start.

As you can imagine, I'd highly recommend ditching the 2/3-octave EQ for a good parametric model. You can't get anything resembling precision with a 2/3-octave graphic EQ. It's just a sophisticated tone control, IMO.

Regards,
Wayne

[NOYB]

Lets work with bandwidth being calculated at the half points. So for a unit with +/- 12 db of gain/cut the bandwidth would be at +/- 6 db. That would be a Q of 1.5, correct?

[NOYB]

I do plan on ditching the last century 2/3 octave gEQ when I purchase a new A/V receiver with built in pEQ. But until then…those few areas are very noticeable, even irritating with programs containing strong material in those areas. Especially in the 200 and 500 Hz area.

[brucek]

Lets work with bandwidth being calculated at the half points. So for a unit with +/- 12 db of gain/cut the bandwidth would be at +/- 6 db. That would be a Q of 1.5, correct?

No, that's not what I meant by definition of bandwidth. The problem is how it is defined in relation to the Q calculation (√2/BW). One equalizer may define its 2/3 octave as the total width of the filter at its -3dB endpoints, (so a 2/3 octave filter would formulate to Q of 2.12). Another may define only the positive half gain endpoint, so a 2/3 octave filter actually covers 4/3 octave overall. There are many different definitions of bandwidth.

You need to simply insert your equalizer into the line-out to line-in loop of REW and measure one of the graphic equalizers filters. Then match it to one of REW's equalizer selections. Once you understand how to translate what REW filters say in relation to your own equalizer, then you can have confidence in the values REW tells you to enter...

brucek

[Wayne A. Pflughaupt]

NOYB wrote:

I do plan on ditching the last century 2/3 octave gEQ when I purchase a new A/V receiver with built in pEQ. But until then…those few areas are very noticeable, even irritating with programs containing strong material in those areas. Especially in the 200 and 500 Hz area.

I don't see anything particularly wrong at 200Hz. You could use the equalizer's 2.5 kHz and 6.3 kHz filters to help the 2 kHz and 5.5 kHz depressions. Just don't overdo it because the filters are wider than the depressions. Likewise, a slight cut from the 400 Hz filter may help the 500 Hz peak.

Regards,
Wayne