What controls the high roll-off on the "flatline"

[rosssurf]

I am loosing just a touch of high frequencies on my Flatline when it is engaged. Can I change a value of something to adjust this for more or less Highs to pass through while the compressor is on?

I have done the  stripboard layout by John Hollis. the only thing I have changed from the orriginal is to use a 100K resistor in place of the called for 220k across the LDR to keep the output more inline with my bypassed signal. I am using an acoustic guitar so my input is rather low.

[petemoore]

I am loosing just a touch of high frequencies on my Flatline when it is engaged. Can I change a value of something to adjust this for more or less Highs to pass through while the compressor is on?
  >>Can't tell, maybe someone else knows the schematic offhand.
  I have done the  stripboard layout by John Hollis.
  >>Doubtful..I've never seen a stripboard layout done by John Hollis. Circuit design is different than a layout design, perhaps you've chosen something from the layouts gallery?
  the only thing I have changed from the orriginal is to use a 100K resistor in place of the called for 220k across the LDR to keep the output more inline with my bypassed signal. I am using an acoustic guitar so my input is rather low.

>>Guessepe doesn't know but suspects the 100k/220k substitution is'nt the problem, and suggests a 220k/100k substitution to test this scenario.

[Mark Hammer]

Unless you employ frequency-splitting and multi-band compression, almost ALL single-band stompbox compressors will eat up some of the treble and sparkle.  I've never understood how that happens in a deep enough way to be able to explain to myself or others, but it happens.

Boosting the treble beforehand will not necessarily alleviate it (nor will fancy-schmancy input buffers).  You probably don't want that treble boost when you bypass anyways, and for what a compressor will do with that boosted hiss when you stop playing....you don't wanna know.

Probably the simplest thing to do is to mimic what I've suggested for the Dynacomp and Ross units.  On those, the 50k output pot is preceded by a 10k fixed resistor.  If you place a small-ish capacitor in parallel with that fixed resistor, it behaves similarly to having the "brite" switch enabled on a standard Fender amp, by providing a lower-impedance path for the high end than for the rest of the spectrum.  The price paid is that the 10k+50k behave like a 60k pot that can never be turned up full.

You can do the same thing on your Flatline, by adding an 18k or 22k fixed resistor between the 1uf output cap and the 100k pot.  Place a 1000 to 1500pf cap in parallel and tell us if you like what it does.  You should notice a) a small but noticeable drop in maximum output level,  b) a small, but noticeable boost above 1500hz, and c) a small but noticeable boost in hiss level during silent parts.

[johngreene]

I can see where a opamp-based compressor would lose high-frequency response if the LDR is in the feedback loop of the amplifier. As the gain of the amplifier is increased the bandwidth response is reduced. So the amplifier has the lowest bandwidth at the highest gain setting. Compensation is difficult because of the fact the response changes depending on the gain. One way of eliminating this is to configure the opamp for constant gain and use the LDR to control a voltage divider on the input. This will at least retain a constant bandwidth over the operating range of the compressor making compensation possible (or at least easier).

--john

[Mark Hammer]

Not so sure about that take on it.  Most compressors work by reducing gain, such that bandwidth ought to increase when compression occurs, following your logic, and of course that doesn't happen.  The other thing that doesn't make a lot of sense, to me anyways, is that while the gain/bandwidth product is a reality, the gain has to be particularly high for bandwidth restriction that might affect the brightness of a guitar to take place.  Most of the ones I'm looking at now have gain-bandwidth products in the range of several megahertz, implying that to drop the bandwidth below 10khz, you'd need some serious gain applied, and generally more than what any compressor would require (at least in the audio path).

The other thing is that it doesn't seem to matter what means of gain/level control is used.  It could be an LDR in an op-amp feedback loop in Anderton's EPFM compressor, an Orange Squeezer using a FET/resistor composite input attenuator, or a current-controlled VCA as in the Dynacomp.  They all do it, it drives me nuts, and I wish I understood why it happens.

[johngreene]

I was merely saying that I could see how it could happen with op-amp feedback based design. But as you say, the amount of gain-bandwidth may not be reduced enough to affect the audible frequencies. I can't say having never experimented with it.

However, with regard to your comment about the four compressors you mention all having this 'feature' the only common thing amongst them I can really see is they all use diode type rectifiers and therefore are more likely to be affected by peak signal. It would be interesting to take a compressor that shows this high-frequency reduction and use a RMS detector and see if the problem is affected at all.

--john

[goosonique]

In did a buffered JH Flatline ...was to sparkely bright for me ... removed it later sounded so fake ...
I believe what Mark shared will give a better deal... but a compressor has to do what a compressor has to...what the he** just PLAY IT LOUD ...