Discussing the EQ

[jfromel]

On Bjorns suggestion I am starting a thread on EQ so perhaps we can all learn about EQ together.

I thought it may be fun and educational if we all learned about EQ, what it is and what it is not, it's effectiveness relative to it's placement in the signal path, perhaps start with something simple like the standard Fender tone stack in a pedal and then move on to more complicated things like a parametric.

A good place to start is the Duncan Tone Stack calculator, it's a Free program and a lot of fun.

http://www.duncanamps.com/tsc/

[BJF]

Hi,

I found this in my drafts folder and something I posted in a technical section somewhere

Perhaps you can look at the tonestack you are refering to as a modified T-filter. To analyze such a circuit would require about 24 pages of computation to see the effects of loading and source, just given one set of values and just varying the treblepot.

The first question to ask is what is it that needs to be removed for this thing to sound good?
With passive filtering you'd always remove something to gain somethingelse........

If you are familiar with T-filters and Baxandal you can take a shortcut.
Like this a T-filter is a midcut filter that gives one trebleflank and one bassflank and you can adjust the heights of these flanks. Oh and the point being to use a midcutfilter primarily to extend the available outputpower by lowering intermodulation. Of course if you'd cut mids at say 400Hz, you'd get more treble aswell since much of the energy in a guitarpickup is in the midrange and as you dampen a part of that the amplification in the treble and bass will increase. You could see how the typical tonestack as used in Fenders would be derived from a Baxandal only that you'd have treble on full and the 'reference resistor' ( the one that sets the reference damping) is now variable and then you custom tune the filter to the response you'd like.

What I am saying here is that if you know the effects of these circuits you may not have to compute each step- for this a computorprogram would be prefered as it would involve a long series of iterations. Yet if you follow the explinations on how the passive Baxedal net works you have also the tools for the T-filter and the function of the T-filter is also available at least at any decent library.

You can hear if you replace the resistor often refered to as slope resistor how the midnotch changes.

OK so try this start with a Baxendal and then allow tuning of the reference resistor.......and voila you have added a midrange control.

If you'd like to go fancy you can construct active filters employing feedback and you could for instance place Twin T Notch filter in the feedbackloop and get a midboost. By offsetting the flanks from a true TTN-filter you can alter the shape of the midboost..........and then just listen to what you want to add.

Active filters always give distortion and usually low headroom and will need to be carefully constructed to avoid side effects.

Feedback systems with vacuumtubes is no more difficult than that of single transistors allthough the impedancelevels are different and the open loop gain is lower in tubes.

Checkout feedback systems for tubes for extended fun.

I'd say though that effective filtering employs the least amount of filters that removes, or sometimes adds the desired range and in this there's no hocus pocus but just the same kind of thinking you'd employ when eq-ing a recording: What is it that sucks and how can that be most easily removed?

Effectiveness is maybe then the easiest controls to desired settings and again what is it that needs to be controlled.

Further more bass-and trebleflanks are limited in various positions in the amplifier and this may also be an effective way of controlling what distorts and when and how.

Oh well just a few thoughts

Have fun
BJ

[ibodog]

I found this in my drafts folder and something I posted in a technical section somewhere

Perhaps you can look at the tonestack you are refering to as a modified T-filter.

What was the "tonestack you are refering"?


The Duncan tone stack software has something similar to the Baxandal but slightly different. Functional equivalent? I think it's called the "Dave".

[BJF]

I found this in my drafts folder and something I posted in a technical section somewhere

Perhaps you can look at the tonestack you are refering to as a modified T-filter.

What was the "tonestack you are refering"?


The Duncan tone stack software has something similar to the Baxandal but slightly different. Functional equivalent? I think it's called the "Dave".

Hi,

Oh, I believe it was a schematic of a Fender amp AB something and the tonestack in that. When looking at that it can be clearly seen how it was derived from a Baxandall but it can also be viewed as a T-filter and compute as one but that's an other story. However in that thread someone also posted a link to the Duncan site- I know what kind of computations it takes to draw these curves and it is of course something suitable for a computor.
Some 30 years ago I asked an engineer to compute a filter similar to this but one I had tuned to what I liked to hear and I wanted to know what it did so I could make a connection between what I could hear with numbers- something that later became more important to me.

You know like when you record a guitar track and you need to tell the technician what frequencies you want attenuated when setting a sound.Such things saves time and creates a language and builds a bridge of words.

Of course later with more experience, if you can relate a feedbacking note to a frequency you can also remove that with a graphic eq that has sufficient numbers of bands such as maybe 32....

One can wonder a bit why graphic eq's for guitar are tuned the way they are?

Personally I liked mostly the MXR 6-band allthough that had a waterfall of noise so I had to rebuild that with lownoise transistors for all gyrators and an NE5532 for the EQ amplifier and thus reduced the noise by 40 dB!

However the tuning of the bands......!.1,6KHz though I hardly ever touched as then that would move the whole balance but I retuned that band for sharper Q and could then use it......

Parametric EQ's I do like especially since they can be most competent in finding just the right spot and you'd need many bands on a graphic eq to equal that....

However I have seen that many musicians have easier understanding of graphic EQ's since also those are supposed to show eq curve graphically while with a parametric eq you really are supposed to know what you are up to;)

Ah there's much more to say about this but it's a start and perhaps this can be a fun thread with many fun posts

Have fun
BJ

[Donner]

Probably the most useful thing I know about EQ is something Bjorn pointed out to me long ago and I have used often.

The most musical treble boost is often an appropriate bass/low mids reduction, NOT a boosting of treble.


Or on a graphic EQ rather than pushing up the treble freqs to get clarity/cut thru ~ try lowering the low mid bands a bit....and then pushing up the volume level a hair to compensate.

[ibodog]

Probably the most useful thing I know about EQ is something Bjorn pointed out to me long ago and I have used often.

The most musical treble boost is often an appropriate bass/low mids reduction, NOT a boosting of treble.


Or on a graphic EQ rather pushing up the treble freqs to get clarity/cut thru ~ try lowering the low mid bands a bit....and then pushing up the volume level a hair to compensate.

Yeah, I'm used to working with studio hardware and software graphic and parametric EQ. In software I normally work with the Waves plugins. Rarely does boosting more than +3dB sound good with most of these and with digital recording will cut into your headroom quickly. If you don't like the sound of something you identify what it is about the sound that you don't like and then try to zero in on that with a cut using a parametric, shelf, or pass filter. Even better to use a spectrum analyzer before and after the EQ so that you have a visual reference to what it was that made the sound bad and what you did that made the sound good. It helps you to remember better the next time.

But I don't know so much about the electronics side of it.

[BJF]

Hi,

On the electronics side of it I'd often think of how could this be simplified with a passive filter when using an eq to set a sound-I don't know eq's were expensive and if you could make the same response with a passive filter you could also optimize the amplifier......

Note that in many guitarpedals you'd find built in graphic type eq sections which shows signs of how they were developed like a motor and then graphic eq's ahead and after and then just the selected bands used with an EQ amp.
I mention this only as it shows uses of eq to shape for instance an overdrive sound before and after.

In recording an analog recording consol may take a little higher overload and in fact there are examples of and recordings of e.g. Eric Clapton playing through an overloaded tube semiparametric eq for distortion.
Still when distortion levels were pushed to lowest possible then also onset of distortion as a consequence had to be abrupt.....

While that's also in a rock 'n' roll guitarists' soul the question of wether this thing can distort?

EQ's seldomly do that pleasently and that really depends on the amplifiers used and the circuit.

Most often I have found live and not using my own amplifier and speakers it useful to have an eq to compensate for what I percieved as deficiencies and then most often just small changes could save the day and it could be something like a 3dB change.

Once I did a a measurement on the absolute bandwidth I'd percieve or rather at which point I'd hear or percieve a loss in treble using tuneable lowpass filter of Bessel type of the 3:rd order through a threeway speaker system and that I could conclude at about 7Khz which by an odd coincidence is about the same upper bandwidth as a tape echo recording head has and also most guitar speakers drop significantly after that.

Ah just some thoughts here

Have fun
BJ

[ibodog]

I found this and was thinking it could be adapted to guitar frequencies. There is adjustment for mid frequency, but it must have a fixed "Q"? Probably the Bass Frequency could be fixed rather than variable for guitar (or toggle for detuned guit or bass). But it might be nice to have a variable Treble frequency instead.

[BJF]

Hi,

Oh my goodness Q is the gentleman that sets up the gadgets for 007!

Right or rather the 'goodness of the filter ahum, known otherwise as the sharpness or width at the -3dB point of the peak.

For guitar often sharp Q's can change the tonality drastically and generally mild eq can be most effective and for treble that's hard one and it requires study since this is a region amplified greatly in guitar amplifiers....

Food for thought

Have fun
BJ

[ibodog]

So in the Studiomaster EQ schematic, is the Mid Frequency dual ganged control changing the center frequency with a fixed width, or is it changing the bandwidth with a fixed center frequency? I suppose the thing to do is to build it up and try it.

At a previous job I had regular access to one of these bad boys http://ap.com/products/2700 which would be perfect for analyzing changes to EQ circuits visually. Anyone use any similar hardware/software measurement tools that are available to the DIY'er?