Is there a simple De esser diy?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

simonsez

Well-known member
Joined
Nov 14, 2008
Messages
679
Location
Jakarta, ID
I think de esser is important gear when deal with vocal track. Is there any simple diy de esser that just using ordinary component (not hard to find, not "That" Ic). I really love Orban de esser, i find this schematic below, but don't know what that module is. Any suggestion? Thank you very much..
 

Attachments

  • 526A.gif
    526A.gif
    88.3 KB · Views: 344
simonsez said:
Yes, i read that, I can't find that "THAT" ic here...

Please edit your profile so that it shows where "here" is, as that greatly increases the odds of anyone being able to help you with that. Several suppliers ship these parts worldwide.

Most de-essers need an RMS detector, which is very hard to do without the matching inherent to monolithic IC processes. I strongly doubt you'll be able to get the hybrid module used in the Orban any easier than the THAT chips.

JDB.
[and while rebuilding such a module is possible in theory, it most likely requires matched transistors which are - again - at least as hard to get and as expensive as the THAT chips, with inferior drift and distortion performance. There's a reason why nobody rolls their own any more]
 
Sorry, edited now..I can find that IC locally. But what i'm looking is a de-esser without VCA ic,  as far i know it's work like narrow band audio compressor. So many compressor circuit out there, but only a few de-esser. Never mind, maybe i should try that Behringer again :-\ can I convert my 1176 into de esser?
 
If you have a compressor with a sidechain input and an equalizer, you have a de-esser.
You mult your vocal track, one signal goes through the compressor as normal, the other
signal goes through your eq and then feeds the side chain input.  It seems counter-intuitive
but you then use the eq to BOOST the frequency that is "essing."  Since the side chain input
is telling the compressor when to compress the boosted frequency will make the compressor
respond quickest to the ess sound. 
 
simonsez said:
as far i know it's work like narrow band audio compressor.

The best de-essers (including the 902) don't work that way. What they do is to compare the level of the ess-band signal energy to the total signal energy, and start compressing when the ratio exceeds a certain amount. For this you need reasonably accurate detectors, a hard task for discrete circuits.

The old trick of using an EQ-ed portion of the main signal into a sidechain input only works well for (relatively) constant input levels; once your dynamic range gets too large that scheme breaks down.

JDB.
 
Thanks guys, So...it's more complex task than just ordinary compressor and no simple circuit can make a good de-esser. I tried that orban once, and it much2 better than my de-esser plugins.
 
Nah... de-essing is not that mysterious.

It's basically a fast attack/fast release limiter, but responsive to HF content.

The Orban circuit appears to be concealing some of it's "secret sauce" to prevent easy copying (apparently a good idea).  :eek:

The FET shunt is the full range gain element, with some distortion reducing (fraction of signal fed into gate) circuitry added.

Their frequency shaping (HPF or BPF) is inside their black box... It looks like they may be half wave rectifying before the shaping, I'd be inclined to filter before I rectify, but different strokes for different folks.

The nature of De-essing, only being active to squash transient sibilance events means you can probably tolerate a lower performance gain element (like FET) that is not in  circuit most of the time. That said I would still invest in the distortion reduction, since de-essers are operating on the full range signal so you might hear distortion on LF content present during de-essing events. I have used FETs like that before in gates or peak limiters since the audio path is very clean when they are not active, and when they are active the FET nonlinearity is not dominating the sound quality. 

There are several published descriptions of distortion reduction with simple FET element (old National Applications for one), and DBX has published papers about the basics of de-essing (in AES preprints). 

Have fun, this looks like a good basic design project to get started on... Perhaps not worth literally copying the incomplete Orban schematic, but an OK general plan.

JR

 
What? :eek: me? I think it's beyond my capability, maybe in my next life (if any ;D).. :-[ :-[ :-[. I even cannot remember the ohm law....
 
The oldest de-essers were a standard limiter with a treble boost in the sidechain. Peak 6KHz "should" be 6-12dB down from peak broadband levels. If not, it is probably an SSS!. And while an SSS! is happening, you can't hear anything behind it. So boost the sidechain path 6dB-12dB at 6KHz (Q=1 bandpass is appropriate) and let it slam the whole track.

This won't do anything for soft "sss". In fact the original motivation was not so much musical, as to avoid bad treble distortion in optical (film) sound recording systems. Soft "sss" is just bad acting; consult your voice coach. Loud SSS! slams the light-ribbon, then the film development and transfer process schmears the tight wiggles on the sound track: yuck!
 
The best analog deesser currently available is the Maselec. It's truly amazing. It only has one knob and it always does the right thing. The side chain is not a simple EQ boost. It looks at both the envelope and the frequency content of the program and varies the compression characteristics accordingly. It also works great as a disk cutting high frequency limiter.

A DBX 900 works well on single tracks but the Maselec is work of art by a designer who knows how it will be used.

Sorta OT but I have been ruined for DIY deessers because I know nothing will be as good. Plenty of good DIY EQ's and compressors though.
 
any expirience with Publison Fullmost reliefe enlarger .. Deesser
and ADR Scamp Dual Deesser s25

 

Attachments

  • s25 Scamp Deesser.jpg
    s25 Scamp Deesser.jpg
    86.3 KB · Views: 210
[quote author=Gold]
The best analog deesser currently available is the Maselec.
[/quote]

Oh yes.

Yes INDEED.

[quote author=Gold]
It's truly amazing.
[/quote]

Unquestionably.

[quote author=Gold]
it always does the right thing.
[/quote]

It's bloody uncanny!

[quote author=Gold]
...It also works great as a disk cutting high frequency limiter.
[/quote]

Note the VERY important difference.

PRR's description is invariably accurate for vocal work. An 'unvoiced sibilant' ALWAYS contains massive amounts of HF energy, whereas a 'voiced sibilant' cannot contain anything LIKE the same amount, so usually needs NO compression. Because of the lack of voicing (the voicing is the chest/nose sound that makes a "Sssss" into a "Zzzzz") a broad band (full-range) control can be used on VOCALS and SPEECH/DIALOG, but punches annoying, DEEP holes in music, making the whole track 'duck' when a cymbal is hit, for example... USELESS for full-band programme/music processing.

The 902 has two options; full-band ducking (excellent for voice) and HF-only ducking. You switch-select the type you want to try. The other genius of the dbx 902 is the split-band energy comparison.

The Maselec takes this further. with varying 'band-spread' and extremely 'transparent' (by which I mean to say that any artifacts it may produce are generally either unnoticeable or largely euphonic) process-band gain reduction.

Leif Mases -it has to be said- is a fabulous designer, and a delightful genius.

[quote author=Gold]
Sorta OT but I have been ruined for DIY deessers because I know nothing will be as good. Plenty of good DIY EQ's and compressors though.
[/quote]

Bang on-topic, I think, since it raises the fact that there are DIFFERENT requirements from a De-Esser. -I actually think that the 902 is MORE useful as a 'solo-vocal-floating-within-heavily-compressed-music-bed' processor than the Maselec... but for analog full-track De-Essing, "Move over dbx... the boss just walked in".

For digital full-range program, the Weiss is pretty damned good, but NOTHING in the analog realm approaches the stunning capability and intelligence of the Maselec.

Keith
 
+1, (wideband) De-essing needs to be performed on the raw problem tracks, to prevent cross modulating LF content.

Does DBX even have a stereo link? A de-esser designed for cleaning up final mixes (like broadcast and after the fact processing) would probably benefit from band specific gain modulations, and smarter control crunching. That is a somewhat different class of processor with sundry examples out there.

For project studio work a simple de-esser should still have some utility when properly applied.

JR


 
I've been told that the Maselec feeds reverse polarity audio back into the signal path for gain reduction. I popped the top once and didn't see a VCA so it's possible.

I only work on stereo program so the Maselec is the ticket.
 
Back
Top