Are there really 'magical' eq frequenices for guitar?

[gwpt]

I am planning on making an EQ stomp box for guitar. I was planning either 5,6 or 7 frequencies.
So, I started looking around for/thinking about what would be good frequencies to use.

In my searching there were many different opinions as to the 'optimal' bands (all the way up to the Rockman PGE-2 having the sacred frequency set!)
Here are some examples of ones used:

Mesa MK IV: 60, 240, 750, 2200, 6600
MXR 6: 100 200 400 800 1600 3200
MXR 10: 31.25, 62.5, 125, 250, 500, 100, 2000, 4000, 8000, 18000
Boss GE-6:100, 200, 400, 800, 1600, 3200
Boss GE-7: 100, 200, 400, 800, 1600, 3200, 6400
Rockman PGE-2: 32, 64, 125, 250, 350, 500, 700, 1000, 1400, 2000, 2800, 4000, 8000, 16000

What are people's views? What would you go for if you were to make an EQ pedal?
Thanks

[Transmogrifox]

Something including 250 Hz as a center is good.  I consider the 150 Hz to 350 Hz band to the be the "mud frequency" range, which often times you want to scoop to make your guitar fit into the mix.

Really these are just octave bands, or fractional octave bands. 

Probably a 32 Hz band is useless for guitar unless you regularly detune your guitar and use octave-down FX. 

My vote would be for the MXR10 lineup, but veto the 31.25 and 62.5 on the low end and 18k on the high end.  This would make for a useful 7-band EQ.  You could add in the 62.5 if you do metal and detune.

The MXR10 as it is probably is better suited to a mix or the drum kit.  It contains everything useful for guitar, and more that is not useful.

I think my ideal would be a 3-band parametric EQ.

[Brisance]

I second the parametric, speaking from long electronical music production/mixing experience, there's no one size fits all eq/compression (or any other) settings, all depends on the particular performance of the guitar as well as everything else in the band playing. Thus a parametric EQ should be best.

[btschock]

I would say if you are looking for a variable frq check out dial tone pickups.

[antonis]

I think my ideal would be a 3-band parametric EQ.

I completely agree...!! 

(even a 4 band EQ is - at most cases - practically useless... )


BTW, it would be interesting a stand-alone xyz band graphic EQ pedal for omittιng individual pedals separate tonestucks/controls/filters build..

[Mark Hammer]

There are likely "magic" frequencies for a given guitar and given sound.  These ideals are, of course, hard to predict, making a product suited to any unknown user tricky to develop and plan out.

There are two ways to tackle this.  One is to cram in as many bands as possible and make them relatively narrow, such that some of them will be either exactly what the user needs, or close enough.  The other way is to use parametric control, and allow for the adjustment of where the cut/boost is situated in the spectrum, and perhaps have focussed or spread-out that cut/boost is.

They each have their pros and cons.  One of the reasons why fixed-band units are called "graphic" equalizers is because the controls are typically laid out in a low-to-high/left-to-right arrangement, such that the user can roughly equate the mental "picture" of the controls with the intended tone.  That is, they can graphically "map" the tonal response.  As well, of only one knob is employed, per band, you can stick a lot more of them in a small space.  Parametric control, on the other hand, while allowing finer control, does require more knobs, and also precludes forming a mental map of the correspondance between control settings/locations and tone.  Although I suppose a multi-band parametric might have left-right sliders to indicate where the resonant band is centred, and up-down sliders for cut/boost.  Still, probably wouldn't be as instantly clear to the user as looking at a row of 31-sliders.

In some respects, the frequencies selected for fixed-band (graphic) types are a function not of the ones that are "magic", but of the mathematics of:
- how many sliders you can fit in the available space
- how much of the entirely usable spectrum you want to affect
- the nearest common cap value

For EQs included in amplifiers, those considerations are added to what sort of shelving controls are also available.  So, if there is a bass and treble shelving-type control, then the sliders can be focussed on the mids.  And if they can be focussed on a narrower chunk of the spectrum, then the bands can afford to be narrower and more closely spaced.

Finally, understand that the reason such circuits are called "equalizers" is that they were originally devised to compensate for listening-space and/or P.A. speaker resonances; the objective being to make any given listening space "equal" to another.  This is, of course why something so ill-suited to a performing musician's needs as a 31-band EQ exist; it is intended to be able to address as many different room/speaker characteristics as possible, and NOT to voice the instrument differently.

[anotherjim]

If I could only pick one -  it would be 1Khz, and cut only - a lot of cut.

[GGBB]

I find with guitar all the interesting frequencies are from about 1-2k down. Above that is basically just treble/brightness/presence. Being able to cut 250Hz "mud" and 1k "nasal" ranges is very helpful. I also find a slight boost around 400-600Hz can be interesting.  Additionally, being able to roll off or boost the low bass below 100Hz and treble. So for me that adds up to 5 controls - low and high shelves, and mid-bass, low-mid, and mid parametrics. Kind of a weird arrangement, but as I said the lower frequencies are far more interesting to me.

[Mark Hammer]

I half-built a 5-band unit, with one of those KIA EQ chips, but have yet to finish wiring it up.  The resonant bands range from around 200-1500hz or so.  I figured shelving controls elsewhere in the signal path would take care of that other stuff for me, so I concentrated on the upper bass and mids for purposes of revoicing the guitar.

[PRR]

Octave-wide bands are not magical. It is like pushing mice around with a large pillow. You can't do anything small or precise. Very blunt tool.

Octaves are the tool for general shape and balance, when you do not have time to custom-rig shelving filters. They can also push trouble-spots way down, at the cost of laming an entire octave to hit one hot-spot.

The internationally approved sequence has "125" and goes octaves up/down. Like the MXR10.

For the specific case of common guitar I might start at 100. This hits 82Hz as good as it gets. The MXR6 or Boss plan.

The PGE-2 is the octaves around 125, plus the half-octaves 350 700 1400 2800, the key midrange. If the half-octs are appropriately higher-Q (also more expensive), then this is musical. In many GEQ topologies more bands is more hiss, which suggests a boost in front and higher supply voltages.

The PEG-2 at 14 knobs is half-way down the road to a 27-Band 1/3rd octave GEQ, which is very powerful, psychoacoustically justifiable, and these days not so very expensive. (It is of course too much front panel to sit on a pedal-board, and is three times more stuff than your "5,6 or 7 frequencies" even if you trim the ends.)

[Mark Hammer]

I still think the ideal guitar tone-shaper consists of a fixed Baxandall-type bass shelving control, two frequency-tunable semi-parametric resonant boost/cut controls, and a 2-pole variable lowpass filter for treble adjustment.  Six knobs that can transform almost any guitar sound into a near infinity of other guitar sounds.

[bool]

528Hz ... everybody knows ...

[tempus]

I read this late last night and didn't get a chance to reply, but I can see you've got a lot of advice here. I agree with most of the other posters, and from personal experience have found fixed band graphic EQs to be less than ideal, simply because the bands are fixed. There s no way to anticipate the interaction of the following variables:

1. the tone the player is looking for
2.  his amp tone/tone stack
3. the guitar being played
4. the speaker being used
5. the style of music being played....

I've used a graphic EQ for a few years now, niggling to find the right frequencies (mostly in vain). A couple of weeks ago I pulled out a parametric, and boom... dial in the frequency, apply the boost/cut, and you're done. So (with the help of RG's geofex page) I designed a 5 (count em, 5) band fully parametric EQ, which I'm waiting on parts for. I chose 5 bands because I found I need to cut the low (<500) cut and boost 2 areas of mids (500-1.5K) cut upper mids/highs (3K ish) and boost some highs (6K ish).
Since you're actually building a custom EQ, you might as well do it right, particularly since the parametric is not much more complicated than the graphic. I can post my schematic here if you like, but I haven't actually breadboarded it yet. This whole mess should fit into a computer power supply case, and I'll be using it in the FX loop of my amp, with no need to switch it in and out (you originally suggested you wanted a stomp box).

If you have an FX loop on your amp, I highly recommend putting the EQ there. You have a much larger degree of tonal flexibility EQing the signal after distortion has been applied than before.