Passive Mastering EQ

[ruairioflaherty]

I used an EQ with a 27k option for 8 years, and the current EQ has a 40k band.  I have never once used either. 

 

[ruffrecords]

Interesting comments. Thanks guys. 27k and/or 40K should be possible. A 1mH inductor would give a Q of about 1.25 at 40K so the bell would start at about 20K. With a Q of 0.7 it would start at around 11K. The shape in the audio band would be more like the start of a gentle shelving boost.

I must admit it made me smile when I heard of an EQ with a 27K option because my 3 band Pultec has an unintentional boost at about this frequency due to the self resonance of one of the Carnhill inductors. I was not aware of it until someone who uses a 48K bandwidth pointed it out. Maybe that is why so many people like the sound of it??

Which brings us neatly onto the topic of frequencies and bands. At present I envisage 5 band. A high, a low and three mids each with 10 selectable frequencies.  o we need that many and what frequencies do people find most use for mastering?

Cheers

Ian

 

[Recording Engineer]

I was hoping for a hi-pass... Also, I specifically like 18k with a 0.5 Q.

 

[ruffrecords]

I was hoping for a hi-pass... Also, I specifically like 18k with a 0.5 Q.

A high pass may well be possible. There is a spare position on the 12 way boost/cut switch - 5 are used for boost, 5 for cut and one for flat.

The way the EQ works means the lowest band can be made a shelving type and the way shelving cut works, the amount of cut can be anything you like. So, bottom line is the spare position on the boost cut switch could be made a HPF. Slope would be 6dB/octave.

18K is no problem. Seems everyone wants Q to be in the 0.5 to 0.7 region.

Cheers

Ian

 

[salomonander]

did i read correctly that you are able to offer variable q? or did i misread that? would be awesome. q to me always depends on utility. for boosting i personally like super wide pultec qs. for cutting i like narrow etc...

regarding the gain steps i honestly thing that a grayhill switch simply wont do in a mastering application. it has too little steps. having 3db boost max is too little imo and any bigger steps would be too much for mastering. but i think anyone wanting a proper mastering eq could get a proper elma 04 switches with 24 positions and put resistors on tere. that way one could do cool stuff like 0.5db steps in the first three 3db and then 1db steps. as seen on the buzz mastering eq for instance . we'd just need the values.

or maybe the switch could use all 12 positions and have a cut/boost toggle as in the mid band of your pultec? that would work as well... we could even bypass bands. which would be very handy.

 

[ruffrecords]

did i read correctly that you are able to offer variable q? or did i misread that? would be awesome. q to me always depends on utility. for boosting i personally like super wide pultec qs. for cutting i like narrow etc...

Yes variable Q is a definite option already built in to the PCB design. It works in just the same way as it does on a Pultec EQP1A by adding series resistance to the LC circuit. This does mean that lowering the Q also lowers the amount of boost

regarding the gain steps i honestly thing that a grayhill switch simply wont do in a mastering application. it has too little steps. having 3db boost max is too little imo and any bigger steps would be too much for mastering. but i think anyone wanting a proper mastering eq could get a proper elma 04 switches with 24 positions and put resistors on tere. that way one could do cool stuff like 0.5db steps in the first three 3db and then 1db steps. as seen on the buzz mastering eq for instance . we'd just need the values.

or maybe the switch could use all 12 positions and have a cut/boost toggle as in the mid band of your pultec? that would work as well... we could even bypass bands. which would be very handy.

I have a lot of sympathy with this view. Obtaining the right boost and cut with just 12 steps is difficult and 24 would open up so many possibilities it is well worth considering. The only issue is the cost but for mastering it is probably justified.

The Buzz I looked at only does half dB steps for the first 2dB. With 24 steps to play with we have 11 for each of boost and cut. So we could do 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 4.0,  5.0,6.0, 8.0 and 10.0 which should satisfy everybody. 

Another note on Q. With the +-3dB version, the cut and boost Q will be identical. With a 10dB range, the cut Q will automatically be several times higher than the boost Q exactly like it is in the Helios69 for instance.

Cheers

Ian

 

[salomonander]

sounds pretty awesome Ian! variable q like on pultecs - amazing

 

[ruffrecords]

sounds pretty awesome Ian! variable q like on pultecs - amazing

The fly in the ointment is the wiring from the 24way switch. Things are simplified using the Grayhill because the switch pins are in straight line. However, the Elma 24 way switch pins are in a circle and an absolute nightmare to track on the PCB. Now if there was a 24 way switch with the pins in a row like the Grayhill we would really be in business.

Cheers

ian

 

[pvision]

I haven't looked hard at the schematic but would think the switchable boost / cut resistors are in a chain. Would it be possible to have twelve 1 db steps plus an additional 1/2 db switch?

It'd be fiddly to use at first but, once you got the hang of it, you'd run through boost or cut settings quickly on the knob then toggle the additional 1/2 db to find the exact value

Maybe the 1/2 db would be the first resistor in the chain? So the switch would short it

Nick Froome

 

[salomonander]

The fly in the ointment is the wiring from the 24way switch. Things are simplified using the Grayhill because the switch pins are in straight line. However, the Elma 24 way switch pins are in a circle and an absolute nightmare to track on the PCB. Now if there was a 24 way switch with the pins in a row like the Grayhill we would really be in business.

Cheers

ian

i was thinking of either soldering the resisotrs to pins old school. or you could design a switch pcb like this to come with the kit:
https://www.don-audio.com/Classi-Audio-Elma-04-SHUNT-PCB
but i assume that still leaves you with the same problems. that said maybe elma can provide you with the layout if you contact them? id think so..

would a cut/bypass/boost toggle work as well? it would keep the cost down and do the same thing essentially. like in the 3band pultec... or check out the massive passive mastering edition. it does the same. i think it would be the best solution if it works with your circuit. cheap, all features plus a bypass per band.

 

[ruffrecords]

i was thinking of either soldering the resisotrs to pins old school. or you could design a switch pcb like this to come with the kit:
https://www.don-audio.com/Classi-Audio-Elma-04-SHUNT-PCB
but i assume that still leaves you with the same problems. that said maybe elma can provide you with the layout if you contact them? id think so..

would a cut/bypass/boost toggle work as well? it would keep the cost down and do the same thing essentially. like in the 3band pultec... or check out the massive passive mastering edition. it does the same. i think it would be the best solution if it works with your circuit. cheap, all features plus a bypass per band.

This EQ is different in that there is a single resistor ladder rather than one per switch. Think of it as a big ladder of resistors across the input, with the output taken from a tap at 10dB attenuation. Each EQ band is an RLC with one end connected to the output. The other end has a flying lead that can be used to connect it to any tap on the ladder. Connect the flying lead to a tap between the output and input and you get a boost. Connect it to a tap between output and gnd and you get cut. The beauty of it is you have one precise resistor ladder to set the steps; and the characteristic impedance , which sets the Q, stays the same no matter how many bands you add. Turning a band off is just a matter of disconnecting it from the output. With a ladder per switch you have to work out the values depending on the number of bands and you cannot disconnect a band without affecting the Q of the others.

The practical downside of the big ladder is you have to wire every switch to the same ladder.  Needs more thought.

Cheers

Ian

 

[ruffrecords]

Just had a thought. We could have separate 12 way switches for boost and cut level and frequency a la Pultec. So for each band you have a boost level and boost frequency, plus a cut level and cut frequency. You could have different or the same Q for cut and boost. Probably a good idea to make the cut and boost frequencies different. The possible combinations would be unbelievable but whether it would be any use as an audio tool is anyone's guess.

Cheers

ian

 

[salomonander]

Just had a thought. We could have separate 12 way switches for boost and cut level and frequency a la Pultec. So for each band you have a boost level and boost frequency, plus a cut level and cut frequency. You could have different or the same Q for cut and boost. Probably a good idea to make the cut and boost frequencies different. The possible combinations would be unbelievable but whether it would be any use as an audio tool is anyone's guess.

Cheers

ian

sounds pretty intersting to me. could cut and boost be used together? if so i think it would be pretty amazing. lets see what te others say. but i like the idea.

 

[ruffrecords]

sounds pretty intersting to me. could cut and boost be used together? if so i think it would be pretty amazing. lets see what te others say. but i like the idea.

They could be used together but if they were identical frequencies they would simply cancel out. In the EQP1A, although it has a common frequency select for bass boost and cut, the actual boost and cut frequencies are not the same so they don't cancel out and this is what causes the famous Pultec 'bump'. The mastering EQ would have separate frequency selection for boost and cut so we need to deliberately make them different frequencies to avoid cancellation.

Cheers

Ian

 

[darkus]

Interested in this. +/- 3dB range is more than enough in my opinion.

 

[salomonander]

it baffles me how many users want no more than a 3db range. maybe its just me but i think its way too little. looking at any mastering eq (sontec, gml, massive passive, buzz etc) none of them has this little play. they all go much further.
if one gets a near perfect mix for mastering than its maybe enough. but thats not the rule
or maybe  people suggesting this are used to itb eqs mainly? on a plugin id never boost past 3db either. but passive hardware eqs react much more subtle in my experience.

 

[ruairioflaherty]

it baffles me how many users want no more than a 3db range. maybe its just me but i think its way too little. looking at any mastering eq (sontec, gml, massive passive, buzz etc) none of them has this little play. they all go much further.
if one gets a near perfect mix for mastering than its maybe enough. but thats not the rule
or maybe  people suggesting this are used to itb eqs mainly? on a plugin id never boost past 3db either. but passive hardware eqs react much more subtle in my experience.

I don't make my suggestion of 3dB range from looking at what Sontec, GML, Massive and Buzz do, I make it from the experience of mastering records every day for many years.

Sure, If I can have extended gain range and 0.5dB steps I will take that (and I have that on the Porter I use every day) but if I'm limited to a 12 way switch then 0.5dB steps is more important that higher boost/cut, especially at the lower end of the range.

 

[ruairioflaherty]

While I'm here I should say that having 5 bands with such low Q does not make sense to me, they will just be working in opposition,  There has been a great deal of fetishizing of the Pultec cut and boost trick over the years but in real life on real pultecs I never found it particularly useful, it's certainly not something I would be looking for in other bands of an EQ.

Lastly, while I'm being a boring party pooper I don't think a passive EQ is the place to implement a high pass filter, getting 18dB/Oct with any frequency selections will mean multiple inductor taps and get complex quickly.

 

[salomonander]

...just shows how different we all work. i master a lot too and for me personally 3db would not be enough. this will be an expensive build and id need more options than 3db.  it would be great to find a solution that fits everybodys needs.

two ideas i have left - maybe both BS. just trying to contribute

maybe the boost/bypass/cut switch would work after all if Ian uses a 2x12 grayhill. one wafer would be responsible for cut, one for boost. the toggle selects the wafer to be used. or maybe i did still not understand how this circuit works. in this case please ignore this suggestion.

other option would be a proper 1x24 rotary pcb mout. i did find one after some search. i cant find it in stock anywhere but maybe its an option after all. i attached the datasheet for Ian. update: i just called electroswitch and they actually never produced this rotary. they can do though. ian feel free to pm me about this in case it is interesting. we'd need a group buy though....

 

[Recording Engineer]

I was thinking 12dB/Oct. Also, I like 0.5 Q for hi-mids and hi-end for one band only... I'm for 1-2 Q for the rest of the bands; maybe up to 3 for some specific low-mids cuts.