Thorsten
I would vote for best sound. Something transparent enough to be 'left in' all the time in my main system.
Now let's see... Palette, or passive Baxandall with active input/output, or Hoshuyama, or Carl Huff's, or UREI 537, or Nite EQ3D, or East German EQ, or ZEQ? This is quite a choice
. Really I don't feel qualified to choose one over the other. I do appreciate the elegance of the East German EQ. It looks like it achieves more with less. Regardless of design, some buffering and gain to make up the insertion loss will be needed. I tend to like discrete solid state. I brought up opamps because they simplify design. I also happen to have some pcbs for the simple circuit shown in post #4 of this thread (thank you moamps and Nelson Pass):
http://www.diyaudio.com/forums/analog-line-level/27651-mox-active-crossover.html
PCB layout is shown on post #19 of this thread (thank you Tiroth):
http://www.diyaudio.com/forums/solid-state/32036-dual-discrete-opamp-pcb-dip8.html
Reading that others may want to try ICs, it would seem easy to accommodate many of us through 8-pin DIP connectors.
I'll keep reading on the various EQ circuit approaches.
Cheers - Pierre
Hi,
I do not think there is anyone who would disagree on "best sound" and insist that we make it "lo-fi". Sadly what constitutes "best sound" tends to vary widely.
On person seems to insist that unless it has paralleled JRC 4560 it will soundly beastly and that tubes will "add texture". Others may posit that unless it has zero feedback tube circuitry it is a lost cause and so on.
I think we are also mostly agreed on having bands at around 20Hz; 120Hz; 500Hz; 2KHz; 5KHz and 20KHz (give or take a little) with a "telescoping" adjustment range, that is the outermost bands have the largest adjustment range and innermost midrange bands the smallest.
Well, PERSONALLY, if I where to make such a device, I would likely select something akin to the East German "ZEQ" but with tubes.
For example one may use a single section 5687 as the Input/Eq combined with the second section in a "Euridice"style transformer coupled line stage with the volume control load "designed in" into the Anode circuit impedance. A Gain control (as opposed to volume) may be inserted before the input tube as may be an input transformer, then the whole EQ/Linestage would have balanced in/out as well. The filters would probably used real inductors for the higher bands and gyrators where values become unavailable from ready stock.
However this would now constitute a tightly optimised single project which on grounds of either "not using the right type of tube", "not using negative feedback", "using too much local negative feedback", üsing coupling cap's and not direct coupling", "using tubes and not discrete transistors", "being too expensive" or "not using paralleled 4560 Op-Amp's (brrr, what a revolting thought that)" and so on ad infinitum would not be acceptable to most.
So I think as a group of interested people we must decide what we want.
One option is to apply our effort to making a purely passive EQ, universal in use that may be completed by the builder using whatever parts and active electronics they fancy (so I can use my beloved 5687 while Daniel can use his beloved parallel 4560 - I suspect each would have shiver of disgust running down each back at the others choice).
The other would be to make one complete EQ design with embedded electronics (Op-Amp's, Discrete, Tubes), optimised, in which case we would have to all agree on the electronics.
This may be best solved with a poll...
Ciao T
I do not think there is anyone who would disagree on "best sound" and insist that we make it "lo-fi". Sadly what constitutes "best sound" tends to vary widely.
On person seems to insist that unless it has paralleled JRC 4560 it will soundly beastly and that tubes will "add texture". Others may posit that unless it has zero feedback tube circuitry it is a lost cause and so on.
I think we are also mostly agreed on having bands at around 20Hz; 120Hz; 500Hz; 2KHz; 5KHz and 20KHz (give or take a little) with a "telescoping" adjustment range, that is the outermost bands have the largest adjustment range and innermost midrange bands the smallest.
Now let's see... Palette, or passive Baxandall with active input/output, or Hoshuyama, or Carl Huff's, or UREI 537, or Nite EQ3D, or East German EQ, or ZEQ? This is quite a choice
Well, PERSONALLY, if I where to make such a device, I would likely select something akin to the East German "ZEQ" but with tubes.
For example one may use a single section 5687 as the Input/Eq combined with the second section in a "Euridice"style transformer coupled line stage with the volume control load "designed in" into the Anode circuit impedance. A Gain control (as opposed to volume) may be inserted before the input tube as may be an input transformer, then the whole EQ/Linestage would have balanced in/out as well. The filters would probably used real inductors for the higher bands and gyrators where values become unavailable from ready stock.
However this would now constitute a tightly optimised single project which on grounds of either "not using the right type of tube", "not using negative feedback", "using too much local negative feedback", üsing coupling cap's and not direct coupling", "using tubes and not discrete transistors", "being too expensive" or "not using paralleled 4560 Op-Amp's (brrr, what a revolting thought that)" and so on ad infinitum would not be acceptable to most.
So I think as a group of interested people we must decide what we want.
One option is to apply our effort to making a purely passive EQ, universal in use that may be completed by the builder using whatever parts and active electronics they fancy (so I can use my beloved 5687 while Daniel can use his beloved parallel 4560 - I suspect each would have shiver of disgust running down each back at the others choice).
The other would be to make one complete EQ design with embedded electronics (Op-Amp's, Discrete, Tubes), optimised, in which case we would have to all agree on the electronics.
This may be best solved with a poll...
Ciao T
Folks,
I have opened a new thread, specific to the Cello palette Functional Clone, including said poll:
http://www.diyaudio.com/forums/analog-line-level/209644-cello-palette-style-eq-design-high-end-tone-control.html
Please use this thread for anything specific to the functional clone project (as opposed to more general discussions regarding high end capable tone controls) and vote in the poll, it is multi-choice, so perhaps vote first and second choice...
Ciao T
I have opened a new thread, specific to the Cello palette Functional Clone, including said poll:
http://www.diyaudio.com/forums/analog-line-level/209644-cello-palette-style-eq-design-high-end-tone-control.html
Please use this thread for anything specific to the functional clone project (as opposed to more general discussions regarding high end capable tone controls) and vote in the poll, it is multi-choice, so perhaps vote first and second choice...
Ciao T
That was a very good idea.
P.S.
I'd never consider running a bunch of JRC in series. No way.
As lovely as the low pitches and bari might be, they're not especially good at treble.
In fact, for treble, I'd like to see definitely something else, and definitely class A so that turning UP the treble doesn't turn up crossover distortion.
P.S.
I'd never consider running a bunch of JRC in series. No way.
As lovely as the low pitches and bari might be, they're not especially good at treble.
In fact, for treble, I'd like to see definitely something else, and definitely class A so that turning UP the treble doesn't turn up crossover distortion.
Last edited:
Your choices are fine with me. I think you got the point that I'd like generous capacity for current so the op-amp doesn't strain.
I don't care what model op-amp you actually choose.
But there is an exception--I really hate DRV134. Likewise, I also dislike "unity stable in all conditions" buffer products since that is compensation overdo, dull results, and unfortunately "sealed" into many buffer purpose op-amps, where the cheap peel-n-stick convenience comes at a cost to quality. As there are 2 of those inside DRV134, it is like two litter trays instead of just one. That particular prospect and class of products is what I find unnervingly nauseating. So, if you wanted to know what I don't like, that was it.
-------
Other topic:
2 bands is faster to operate than 6, but 6 is more useful.
I'm curious why don't we consider 8?
40, 120, 500, 1k, 2.2k, 4k, 7k, 12k
I'm asking about this because the raspy 7k is a totally different need than ear canal resonance at 4k, and I really don't think that one knob is going to do both. Likewise, to sort the midrange, you really might want to be able to dial down the glare at 2.2k without reducing 1k.
Since the year is 2012, I don't forsee the tone amp ever used for remastering; however, it looks extremely useful for speakers/rooms/ears compensations for music replay purposes, with high fidelity equipment.
Last edited:
Hi,
The 5687 is an op-amp only in the loosest possible terms, as, while it has indeed positive and negative inputs as well as an output, it actually is a Tube.
Why don't we consider 61 then? I would say the best number for people to handle easily is 5 - 7. Six is the original palette and from my experience works fine.
The bands are quite broad, so you may find the 5K band suffices.
Everything is a compromise, switchable centre frequencies are a possibility, but detract from ease of use.
I do see it used for re-mastering. If at your place you adjust the tonal balance you ARE remastering.
On the other hand given that speaker/room and hearing problems are rarely wideband something the Palette would singularly unsuitable for the Job.
Ciao T
The 5687 is an op-amp only in the loosest possible terms, as, while it has indeed positive and negative inputs as well as an output, it actually is a Tube.
Why don't we consider 61 then? I would say the best number for people to handle easily is 5 - 7. Six is the original palette and from my experience works fine.
The bands are quite broad, so you may find the 5K band suffices.
Everything is a compromise, switchable centre frequencies are a possibility, but detract from ease of use.
I do see it used for re-mastering. If at your place you adjust the tonal balance you ARE remastering.
On the other hand given that speaker/room and hearing problems are rarely wideband something the Palette would singularly unsuitable for the Job.
Ciao T
In this case, please consider 7.
40, 120, 600, 2.2k, 4k, 7.5k, 12k
It is the treble area that has the greatest need--average ear sensitivity peak at 2.2k and average ear canal resonance at 4k. I think it is important to pay attention to audiology data when discussing a tone amp, and its potential for usefulness.
EDIT: ONE additional dial instead of the complications of adjustable frequencies. . . there was much discussion over pushing the 5k about lower or higher and for very good reason. But, let's accommodate that in a streamlined manner instead of a band-aid fix. Right?
Last edited:
Hi,
We are not talking narrow band parametric EQ's here, so frequency
bands don't really matter and the above is not very symmetrical.
You've got gaps of over an octave, over 2 octaves, nearly 2 octaves
followed by ~ an octave, another ~ octave and less than an octave.
Its simply work work well at the top end, too many knobs too close
together, that are impossible to intuitively set at low settings.
Its seems fairly clear what Cello are up to, ~ symmetrical 20Hz to 20KHz.
Raising the bass frequency a little would make it even more symmetrical.
rgds, sreten.
Last edited:
Do you consider ears important?
Well, yes, just like your chart. . . except that I wanted to split the 5k control into 2 controls so as to give ear canal resonance its own knob.
See also ear frequency response data to see that auditory band sensitivity center is 2k, not 1k, and therefore I believe that the range of boost/cut should approximately follow the range of compensation indicated as necessary within a Fletcher Munson or ISO chart.
There's just a couple of minor differences.
If matching an audiology chart, all of the treble controls would have a slightly smaller boost/cut but the 500hz (or 600hz) would have a slightly larger boost cut than shown on your chart. It isn't a big difference, but shouldn't we think about matching transmitted signal to its intended receiver, or is that an outdated notion?
Well, yes, just like your chart. . . except that I wanted to split the 5k control into 2 controls so as to give ear canal resonance its own knob.
See also ear frequency response data to see that auditory band sensitivity center is 2k, not 1k, and therefore I believe that the range of boost/cut should approximately follow the range of compensation indicated as necessary within a Fletcher Munson or ISO chart.
There's just a couple of minor differences.
If matching an audiology chart, all of the treble controls would have a slightly smaller boost/cut but the 500hz (or 600hz) would have a slightly larger boost cut than shown on your chart. It isn't a big difference, but shouldn't we think about matching transmitted signal to its intended receiver, or is that an outdated notion?
Last edited:
Hi,
Audiology data? How do you equalise your ears during a performance at the concert hall?
I think you still misunderstand the purpose of the Palette. Maybe read this article regarding the lower cost "Palette Preamp"...
Cello Palette Preamplifier | Stereophile.com
Anyway, if you want to have these frequencies and controls, design your EQ for them...
Ciao T
Audiology data? How do you equalise your ears during a performance at the concert hall?
I think you still misunderstand the purpose of the Palette. Maybe read this article regarding the lower cost "Palette Preamp"...
Cello Palette Preamplifier | Stereophile.com
Anyway, if you want to have these frequencies and controls, design your EQ for them...
Ciao T
Thorsten - this resonates with me. If you go this way, I will follow. Perhaps this will be my first tube project?
Pierre
The arranger and conductor perform this task as well as possible by ear.
Thank you for the article.
Hi,
What they do is rather different to equalisation, with respect.
In the end you select the numbers of bands and frequencies you like, however to have a distinct EQ band at 4KHz and at 7KHz you would need quite high filter Q with all that means sonically.
However, looking at the ISO2003:226 equal loudness curves I do not see any indication that specific needs exist for a 4KHz and 7KHz band, rather it would seem that our 2KHz and 5KHz bands should shift to 1.5KHz and 3KHz, however I do not really see any particular direct need to place EQ bands precisely on peaks/dips in the equal loudness curves.
If we on the other hand compare the Palette EQ frequencies with the "spatial bands" according to Blauert et al., we find a pretty good match with original palette, it may be worth lowering 500Hz and 5KHz to 400Hz/4KHz, but we are really splitting hairs here.
Ciao T
What they do is rather different to equalisation, with respect.
In the end you select the numbers of bands and frequencies you like, however to have a distinct EQ band at 4KHz and at 7KHz you would need quite high filter Q with all that means sonically.
However, looking at the ISO2003:226 equal loudness curves I do not see any indication that specific needs exist for a 4KHz and 7KHz band, rather it would seem that our 2KHz and 5KHz bands should shift to 1.5KHz and 3KHz, however I do not really see any particular direct need to place EQ bands precisely on peaks/dips in the equal loudness curves.
An externally hosted image should be here but it was not working when we last tested it.
If we on the other hand compare the Palette EQ frequencies with the "spatial bands" according to Blauert et al., we find a pretty good match with original palette, it may be worth lowering 500Hz and 5KHz to 400Hz/4KHz, but we are really splitting hairs here.
An externally hosted image should be here but it was not working when we last tested it.
Ciao T
Thank you for considering all of that.
Meanwhile, back to this:
The entire middle section, with what appears to be 6 channels, looks extremely easy since you just make 1 and then 5 copies, with, apparently minor filter variations. So, most of what fills that box is not frightening, not complicated, not difficult and not expensive at all.
It does have a couple of frights though, the three boards at far left, three boards at far right and large collection of the most expensive controls. . . have got me bluffed a bit. Can those areas be done at reduced complexity and costs?
Meanwhile, back to this:
The entire middle section, with what appears to be 6 channels, looks extremely easy since you just make 1 and then 5 copies, with, apparently minor filter variations. So, most of what fills that box is not frightening, not complicated, not difficult and not expensive at all.
It does have a couple of frights though, the three boards at far left, three boards at far right and large collection of the most expensive controls. . . have got me bluffed a bit. Can those areas be done at reduced complexity and costs?
Hi,
Well, there are a lot of around USD each J-Fets used in the discrete circuitry. It gets expensive quick...
I have reason to believe that the circuit is BROADLY (not exactly) related to Project 75 on the ESP Site. Generally this kind of design is an (IMNSHO unnecessary) Op-Amp-Grave with little actual benefit (performance or other) over different structures with much greater simplicity.
Well, that depends on what you want.
In effect you can build a Project 75 using NE5532 and TL072, a bunch of cheap pots and a bad power supply call it "Functional Clone". Will the difference be audible or material? Your call.
Ciao T
Well, there are a lot of around USD each J-Fets used in the discrete circuitry. It gets expensive quick...
I have reason to believe that the circuit is BROADLY (not exactly) related to Project 75 on the ESP Site. Generally this kind of design is an (IMNSHO unnecessary) Op-Amp-Grave with little actual benefit (performance or other) over different structures with much greater simplicity.
Well, that depends on what you want.
In effect you can build a Project 75 using NE5532 and TL072, a bunch of cheap pots and a bad power supply call it "Functional Clone". Will the difference be audible or material? Your call.
Ciao T
Hi,
I agree you are splitting hairs. But FWIW if you want to consider
audiology then think about the Bark or Mel scales which indicate
in octaves the lower controls should be closer together and the
higher controls further apart if anything, not the other way around.
"It is the treble area that has the greatest need" is simply not true,
unless you narrowband EQing headphones or the like with high Q
filters. Here the centre frequencies are largely academic with the
very low Q's involved at the subtle settings used most of the time.
If anything 1KHz used as the centre frequency for perception of
tone is wrong. It really all happens about an octave lower AFAICT.
Don't ask me to explain what it means, a lot happens below 1KHz.
FWIW 2 octave changes above 1KHz sound more like 1 to most.
But you'd be a marketing fool to mess with the ubiquitous 1KHz.
An interesting proposition for 6 bands would be ~ :
20Hz, 100Hz, 333Hz, 666Hz, 1.33KHz, 4KHz, 20KHz.
x5, x3, x2, ~x2, x3, x5. could work very well. YMMV.
Symmetrical and a halfway house compared to normal.
rgds, sreten. (ex-telecoms speech quality expert)
Note that you have two centre frequencies widely regarded
as the minimum bandwidth for telephony, 333Hz to 4KHz.
I agree you are splitting hairs. But FWIW if you want to consider
audiology then think about the Bark or Mel scales which indicate
in octaves the lower controls should be closer together and the
higher controls further apart if anything, not the other way around.
"It is the treble area that has the greatest need" is simply not true,
unless you narrowband EQing headphones or the like with high Q
filters. Here the centre frequencies are largely academic with the
very low Q's involved at the subtle settings used most of the time.
If anything 1KHz used as the centre frequency for perception of
tone is wrong. It really all happens about an octave lower AFAICT.
An externally hosted image should be here but it was not working when we last tested it.
Don't ask me to explain what it means
, a lot happens below 1KHz.FWIW 2 octave changes above 1KHz sound more like 1 to most.
But you'd be a marketing fool to mess with the ubiquitous 1KHz.
An interesting proposition for 6 bands would be ~ :
20Hz, 100Hz, 333Hz, 666Hz, 1.33KHz, 4KHz, 20KHz.
x5, x3, x2, ~x2, x3, x5. could work very well. YMMV.
Symmetrical and a halfway house compared to normal.
rgds, sreten. (ex-telecoms speech quality expert)
Note that you have two centre frequencies widely regarded
as the minimum bandwidth for telephony, 333Hz to 4KHz.
Last edited:
Hi,
One can stick to the ISO standard frequencies and avoid infortuitous numbers.
So:
16, 125, 315, 630, 1260, 5040 and 20160
However, I think the single 500Hz control is okay instead of 315 & 630 and we do not really need the 1KHz ish control, but one at a frequency a bit higher.
Interestingly the quite legendary NitePro EQ3 has four bands below 1KHz and only one above (plus "Air").
We can argue long and hard, I personally feel 6 bands are already borderline, I'd rather four bands with a few centre frequencies on the middle bands, but this bound to confuse many.
What I can say is that for mastering 10 bands are already way too many...
Seven would be a "good number" though.
Ciao T
One can stick to the ISO standard frequencies and avoid infortuitous numbers.
So:
16, 125, 315, 630, 1260, 5040 and 20160
However, I think the single 500Hz control is okay instead of 315 & 630 and we do not really need the 1KHz ish control, but one at a frequency a bit higher.
Interestingly the quite legendary NitePro EQ3 has four bands below 1KHz and only one above (plus "Air").
We can argue long and hard, I personally feel 6 bands are already borderline, I'd rather four bands with a few centre frequencies on the middle bands, but this bound to confuse many.
What I can say is that for mastering 10 bands are already way too many...
Seven would be a "good number" though.
Ciao T
Last edited:
You might want to check the Mouser price just to be sure that's accurate. There's many options and they're not counterfeit and the small signal parts are a few cents.
Project 75? Heck no! There's so much going on with it that its more of a sound processor and it loses the #1 feature of the palette. . . an eq with un-colored rendering. So, the functional clone is dysfunctional. Well, let's not do that.
It seems that we've successfully narrowed the options down to 2:
Nite3d and East German EQ.
Meanwhile, back to this thing:
Now, on the East German EQ, I really like the 3 simple transistors and the simplicity of the passive circuits as well. The inductors have me bluffed but I think that a quick briefing on where to buy them or what to do for them would succeed.
On the Nite3d is where I'd like to see some Class A bias. One wouldn't want to crank up the air control with crossover distortion present. I don't really care what sort of op-amp's you'd pick for this, but discrete might be and stay more specific.
Which one of the remaining two options do you like the best for the new 6 band tone amp, East German EQ or Nite3d?
Maybe it is time for some fun? Perhaps we could have a look and a build of an inexpensive, uncomplicated small scale 3 band version of each? The physical units can be compared head to head, which could be fun.