A DIY de-esser

[chris p]

I've built myself a Gyraf SSL clone compressor, complete with a sidechain insert point. My main motivation for the insert point was de-essing, but I'm short an external EQ to achieve this.

So I was wondering whether any of you DIYers and tech expects have any comments or pointers on the following, ....

For de-essing, you are really only looking at a single band pass filter, say from 9kHz to about 14kHz (?). At it simplest and most passive, then, we're looking at a capacitor (the HPF) in series with an inductor (the LPF). Being the sidechain, we're not too concerned about how badly we mangle the sound provided the transients in the target band trigger the compressor's VCAs.

Can you indeed make a passive "plug-in" jack for the insert socket that does not more than this, and will it do the job? If so, I imagine you could have a number of them with different component values that slightly jiggle the cutoff frequencies.

Without getting ahead of myself, can anyone suggest a website to look up the values for the components? I'm guessing around 3.3uF for the caps (eg Jaycar RY6953) and about 47uH for the inductor (eg Jaycar LF1530?).

[Kurt]

Doesn't the Jaycar site have a table for working this stuff out?

[PeterR]

the Q value of the LC filter would be reasonably importand I imagine?

you cant go with a terrible quality inductor cause otherwise you'll be boosting (cutting everything else really) a broader range of frequencies which will possibly trigger the de-essing when you dont want it.

you need a decent (higher) Q value, which requires a good inductor. I had a spreadsheet done up for all this stuff but I can't find it. The DC resistance of the inductor is the key for getting a good (high) Q. Low DC resistance = high Q (narrow notch) Unfortunately the higher frequency you wish to effect the lower the DC resistance must be (the required inductance is lower though so this helps). Ill look at work tomorrow for my spreadsheet (gave resonant frequency, reactance and Q for L and C values) also worked backwards.

with a passive LC bandpass you loose voltage too so you have to consider if the signal will still trigger your compressor.

Its been too long since I investigated this stuff. The best place to look is groupdiy and check out the NYD and pultec eqs under passive eqs. The NYD eq requires a super low source impedance and a 600ohm load so it probably wouldn't suit.

http://www.prodigy-pro.com/forum/viewtopic.php?t=253

[JustinS]

There are different ways you can implement it - a band pass filter or a simple high pass filter... this thread at the lab:

http://www.prodigy-pro.com/forum/viewtopic.php?t=11433&highlight=high+pass+ssl

Has filter boards for sale specifically designed to put a Highpass filter in the sidechain of of the SSL. You'll have to tweak the caps to getter higher turn-over points - 200Hz is a wee bit low for a de-esser. Here's a website that goes through passive eq theory to work it out:

http://home.online.no/~jaeioluf/sound/eqpassiv.htm

One thing to think about, you don't have to go the passive inductor route, Fred Forsell has a nice white paper on simulating inductors with op-amps:

http://www.forsselltech.com/Evolution%20of%20an%20EQ%20Design2.pdf

Justin.

[PeterR]

spreadsheet is gone sorry.

try here

http://my.integritynet.com.au/purdic/electronic-basics.htm

scroll down there is a calculator there for LC at given frequency.

Im pretty sure reactance is important to how effective the band pass is. Higher reactance = more voltage difference between resonant and unfiltered signal. The formula for reactance only needs to be done on one of your values (L or C) as they have the same reactance at their resonant frequency.

these are some starting values from that NYD eq

5kHz, 18mH, 0.047uF, Max DCR of L = 20Ohms

some of the formulas use - pF and uH etc.