Auto master EQ

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

gnd

Well-known member
Joined
Jan 24, 2006
Messages
285
Hi.

Automatic EQ, working as multiband compressor, but a bit different.

It would be like ordinary wide Q eq, 3 or 5 fixed freq/Q bands, but with eq levels changing slightly (+-3dB, or +-6dB) with incoming signal. Maybe LDRs could be there to change gains in EQ, with fixed attack/release. So there would be some compression. But it is not really compressor, it may be better called dynamic EQ or something.

What do you think of this idea?

I would build such a thing. Are there any schematics of something similar already?

...
 
[quote author="gyraf"]why isn't this a normal multiband compressor?[/quote]

If I remember properly, multiband comp is splitting signal to several bands with crossover, then applying compression to each band, and then mixing bands together again. This mixing of signals is problematic at crossover points.

Dynamic eq is working differently, it is just a eq. Like you would be manually driving gain pots while mixing.

Maybe some commercial products are called multiband comps, but are actually dynamic eqs? Would be interesting to study some schematics.

Where to get more info on this subject?

...
 
Where to get more info on this subject?

Good question - I've always thought of multibandcomps as dynamic eq'ing, but you're right, they're often done by using dividing filters and processing individual bands.

Jakob E.
 
What I have in mind is EQ topology like Pultec, or similar. Passive eq, with gain makeup stage. No tubes. I have some lundahl LL1517/LL1539, and would like to use them on output. I wouldn't bother with switchable frequencies, because in the end in my real life, they remain on some fixed positions all the time anyway. The only thing I would do is to add some LDRs to eq gain pots, to make them move off the pot's position a bit, 3dB up/down. I need to figure this out still, just an idea, but probably it is doable, is it? Attack/release would be slooow, real slow, maybe 1 second attack and 5 second release.

I would maybe try some other eq topologies. I'd go for really wide Q eqs. I also like overlapping eq bands, so that changing one freq band has influence also on other freq bands. Something organic, which lives by itself. :?

I always liked TL Audio EQ1. Maybe somebody has schematics of this one? At least filter section?

Any other recommendations for filter topology?

...
 
I have a similar idea I want to experiment with, although just with one band. A wide-Q LC series filter to ground through a vactrol.

A de-esser of sorts, but ment for wider "taming" of upper harsh frequencies. A few selectable frequencies: 2500Hz, 5000Hz, 7500Hz, 10000Hz..

Not much GR needed, some 2-3dB should be more than enough. Stereo tracking is going to be a problem though..

Hell, I'm going to try to put a few GE diodes in place of the vactrol and see what kind of a sound I can get. :thumb:
 
they already make a Dynamic EQ. it's made by bss and is model 901...

they have a good explination of how it works. check out these pages...
dpr901iium-16.jpg

dpr901iium-17.jpg
 
DPR-901.

Killer every time.

Complicated to set up sometimes, but when it's useful, NOTHING else works like it.

FORGET LDR's. You absolutely HAVE to have a very accurate forward-predictable gain behaviour, and LDR's simply don't do that at all well.

In fact, you are basically forced to use VCAs (unless youcan invent something which has eluded people for a good long while, and immediately steps well outside this topic), and you do get a complicated product to build as a result...

But in the end, not many people use them. the 901 is a FANTASTIC piece of gear, and trying to get people to use it -even when you have one righ there in the rack in front of them- meets with resistance more often than not.

But I love the one we have here. i know a few front-of-house live guys who won't leave home without one. -For live sound, controling stage "boom", being able to de-ess a little and guarding against midrange 'squawk' is just part of what they do with them: They rock.

If you can do it, by all means build one. -I bet it'll look a lot like a DPR-901 when it's finished however, and you might want to think about how much it'll cost to build versus how much you can get a 901 for... plus the 901 will be right the minute you open the box... a DIY design might need a LOT of tuning to make the ranges useful... then there's the display so the user can figure out what the hell the box is doing, how far gfrom threshold (before onset) etc...

Yes, the 901 is king.

Keith
 
Yes, dynamic EQ has value to generate a target spectral output.

I see application for this in next generation entry level sound systems (probably DSP based) where the circuitry detects the probable source identity (vocal or general instrument classification) and EQs for a target spectral balance.

The cost of DSP is dropping to point this may happen within the decade. At some point it will be cheaper to do this in software, than to put all the knobs and sliders on a cheap PA.

Or not.

JR
 
I did a dynamic EQ for vocal enhancement, for people without a voice. It works great. I can make a kit if it is demand on such devices. Later one said, "Anatoliy, we did record on studio, they can't record my voice without distortions!" However, I did not tell him that when I worked with them his voice was always distorted... :cool:
 
[quote author="Wavebourn"]I did a dynamic EQ for vocal enhancement, for people without a voice. It works great. I can make a kit if it is demand on such devices. Later one said, "Anatoliy, we did record on studio, they can't record my voice without distortions!" However, I did not tell him that when I worked with them his voice was always distorted... :cool:[/quote]

Can you maybe show the schematics?

thnx.
 
[quote author="gnd"][quote author="Wavebourn"]I did a dynamic EQ for vocal enhancement, for people without a voice. It works great. I can make a kit if it is demand on such devices. Later one said, "Anatoliy, we did record on studio, they can't record my voice without distortions!" However, I did not tell him that when I worked with them his voice was always distorted... :cool:[/quote]

Can you maybe show the schematics?

thnx.[/quote]

It was a long time ago, at the beginning of 1980'th. However, I can design a new version.
 
[quote author="Wavebourn"]
It was a long time ago, at the beginning of 1980'th. However, I can design a new version.[/quote]

You can design new version?
I'd be interested to se at least basic schematics, to get some idea...

I still have no idea what I will do, and any info is welcome.

...
 
[quote author="gnd"][quote author="Wavebourn"]
It was a long time ago, at the beginning of 1980'th. However, I can design a new version.[/quote]

You can design new version?
I'd be interested to se at least basic schematics, to get some idea...

I still have no idea what I will do, and any info is welcome.

...[/quote]

I will do that.
 
[quote author="SSLtech"]
FORGET LDR's. You absolutely HAVE to have a very accurate forward-predictable gain behaviour, and LDR's simply don't do that at all well.
[/quote]

Thanx, Keith. I have searched eBay for bss 901. I guess this thing goes beyond my intentions, and beyond my needs.

I'm still somehow into LDRs. Vactrols seem great. You say that I absolutely have to have a very accurate forward-predictable gain behaviour. But I will go for really slow attack/release. Why would I need predictable gain behaviour? For stereo accuracy? Or is problem somewhere else?

thnx.
 
[quote author="JohnRoberts"]Yes, dynamic EQ has value to generate a target spectral output.

I see application for this in next generation entry level sound systems (probably DSP based) where the circuitry detects the probable source identity (vocal or general instrument classification) and EQs for a target spectral balance.

The cost of DSP is dropping to point this may happen within the decade. At some point it will be cheaper to do this in software, than to put all the knobs and sliders on a cheap PA.

Or not.

JR[/quote]

It's already being done in higher-end car system. Someone told me they think it was a German manufacturer who was doing it... Higher end Volkswagons?

I also talked to a car audio shop, and they said they can get their hands on auto-EQing DSP gear pretty easily.
 
[quote author="SSLtech"]DPR-901.

Killer every time.

...i know a few front-of-house live guys who won't leave home without one...

Keith[/quote]



Yes, I'm one of those... Love it.

/Anders
 
Why do we need matching/predictable look-ahead?

Well, if you don't have that, the ONLY topology which you can go with is feed-back, -otherwise things are WILDLY out of control.

Feed-back is no small undertaking in this matter. You have to monitor ONLY the affected band, since out-of-process-band parts of the spectrum will skew your compression. Vactrols are a LOT more predictable than Cadmium-Sulphide optos, but nowhere near as well-behaved or predictable as VCAs...

If you're talking about band-splitting a signal, doing DOWNWARD gain control only, and re-combining, then you're really steering yourself back towards the kingdom of multi-band compression/limiting. If you want frequency selective gain increase/gain reduction to differentiate your project from a MBC/MBL, you're heading into 901 territory.

-Yes, the 901 looks expensive to build because it has about a hundred LEDs on the face of it, which looke like a frivolous lightshow, but when you USE the device, you HAVE to be VERY aware of when it kicks in and -more importantly- when it's GOING to kick in... If you have lots of upward gain expansion set, things can go into feedback live if your gain-staging isn't correct, and you need to have some warning of impending action.

But don't just take my word for it: -ask a user: Pucho or Kvintus for example. See if they think you could strip that box down at all and have anything useable. -I bet the only thing you MIGHT be able to lose is a fourth band, but sometimes even all 4 bands get used, other times one or two only. -Our particular unit is the later version which allows you to "split" the two mids away to a seperate path and therefore have a pair of 2-band devices if you don't need all four.

Having used it and knowing it's capability, i think they're VERY useful for problem-solving. However, If you blindfolded me, or reduced the displays, I would have a VERY hard time setting it. -and working fast especially, you NEED gear to be helpful and informative, NOT difficult or obstructive.

-And "Can't you HEAR when it's working how you like it?" doesn't cut it for me: this is NOT full band squashing, and the effect is VERY subtle at times, so it can be VERY difficult to hear indeed. Onset, ratio and amount can be difficult to judge when the source signal varies a lot: like the difference between a "Sh" sound on a vocal and a "Mmm" sound for example... is it compressing or is he singing softer? Am I doing too much? -Without informationthis stuff is impossible to use. -Now, how are you going to get information from a vactrol? -You can't meter only the narrow band sidechain audio signal and get anything meaningful, beause it won't represent anything the brain/eye can easily make sense of... (if you were set to high frequencies, it would ignore most singing and hit the end-peg as soon as he whispers an "S" for example...)

So you're pushed back to reading meaningful control voltages for Gain control, and perhaps a band meter for threshold/onset... and you're being pushed towards building something that worls like a 901 again...

Hey, prove me wrong. -PLEASE. -But I've looked at this before, and I really do think that BSS just got it right. I mean if you really want to build the ultimate analog console surface, you end up with something that looks a LOT like an SSL G series... if you want the ultimate dynamic equaliser, you end up with something that starts to look a LOT like the BSS.

Keith
 
Back
Top