Improving my compressor design

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
abbey road d enfer said:
In case of a sudden blank, voltage at D11's anode becomes low enough for the diode to be less conductive, which reduces the shunting effect on the rectified voltage thus increasing ratio. It results in downward expansion, thus reducing noises that usually appear when compression ceases.
Fascinating. Almost like a 'GR hold ' then !?
I'm wondering whether D11 also serve to  sort off 'buffer' the Ratio pot, preventing it from interferring with other resistances in the vicinity -- such as the release time pot, which is also shunted to ground? Maybe this is nonsense...
 
Script said:
Fascinating. Almost like a 'GR hold ' then !?
that's correct.

I'm wondering whether D11 also serve to  sort off 'buffer' the Ratio pot, preventing it from interferring with other resistances in the vicinity -- such as the release time pot, which is also shunted to ground? Maybe this is nonsense...
I wouldn't follow you in this direction...
 
I killed way too many brain cells last century refining side chains for dynamics processing. The majority of my effort was for companding noise reductions mostly used with limited dynamic range audio paths. A BBD delay line or cassette tape could exhibit a marginal to inadequate dynamic range... expanding that 2x with a 2:1 compressor before the limited path, and 1:2 expander after, could remap a -45 dBu noise floor to -90dBu while doubling the head room.

=======TMI about vinyl NR you can skip this====

Vinyl did not escape untouched, CBS records attempted an ill fated 1:2 playback expander, called CX, compressing the recording before cutting to vinyl. I won't list all the mistakes CBS made, but I got approached through Popular Electronics with the offer of a cover article, and free CX license. Hard to quantify the promotional value of a cover article on a 500k circulation magazine, but I did not say no. One perhaps amusing anecdote, I only had a few weeks to turn the design and I was provided pro forma circuit designs in the CX licensee package. Not heavy lifting after all my NR experience. I cranked out my version of CX and corrected a mistake I found in one time constant of the pro forma design. I delivered my article to Popular Electronics on time and notified CBS of the mistake in their design advice. My cover article was scheduled to run in the Christmas- December issue (not too shabby). At the New York AES show that year, literally days away from the issue printing I visited the URIE booth, since they made the 2:1 encoder for CBS. After introducing myself, they said "so you're the guy who found the mistake."  :eek:  CBS told URIE to change the time constant in their encoder to agree with the mistake because two other licensees had manufactured thousands of decoders with the wrong time constant mistake. Of course CBS did not bother to tell me, so I was about to publish the only "wrong" CX decoder design since they changed the specification over night. I frantically contacted the editor at Popular Electronics and made my design agree with the changed time constant in the nick of time.  I still have a stack of CX records CBS gave me as a licensee, but screw them....  :mad: 
======

Of course to compand an audio path well you want to do it as transparently as possible and minimize audible artifacts from the tens of dB gain manipulations. 

For a somewhat longer list of side chain features (still incomplete) beginning with the obvious;
  -attack time : you want this fast enough to pass sudden transients cleanly, but not so fast that you can hear brief transients modulate the noise floor.
  -release time: you want this slow enough that the side chain does not try to release during a single cycle of a LF waveform.
  -hold: Ideally you hold the side chain at a stable gain so there is no signal modulation occurring.
  -adaptive or signal dependent time constants: Ideally you want faster attack for large signal level changes, and very slow for small level changes. There are too many ways to accomplish this to list. 
  -HF pre/de-emphasis: widely used in multiple NR systems this takes advantage of the relatively modest amounts of HF content in natural music. Boosting HF before and cutting it again after provides some "free" NR, but just like the real world there is no free lunch, so HF pre/de-emphasis reduces HF headroom.

Within this short list of basic parameters there are subtle psycho-acoustic relationships in human audition that can be capitalized on for transparent NR. For just a couple examples, in human audition loud sounds can mask quiet sounds making them less objectionable unless playing by alone. Another human hearing trick is that our ears sensitivity changes slowly after a loud sound stops, so this provides a brief window where larger gain changes can be concealed.  There are other tricks like this.

====TMI about tape NR you can skip this too =====
We are all (some of us are) aware of DBX and Dolby tape NR, I sold two different Tape NR kits and this is my best later design.

p-522.jpg


This design uses even more tricks than I listed. I have shared this schematic here before so there are more details in the past discussions. To better understand this you might want to look up a NE572 compander chip... that IC contains a fair gain element and two rectifiers per that can be used separately. In fact back in the 80s I used a 572 inside the side chain for a commercial compressor I designed (LOFT-Phoenix Audio Lab).  I just did a search and see somebody new is using the "Phoenix" name for audio products.... the new Phoenix  like the legend arose from the ashes (but somebody else's ashes. )

JR
 
This is the kind of discussion that's super interesting but also almost completely disheartening.

I just don't have the resources or ability to independently get some of this stuff. It's tribal knowledge that is being lost, I'm afraid.
 
dogears said:
This is the kind of discussion that's super interesting but also almost completely disheartening.
Sorry I invested decades on the bench acquiring this tool set. Read AES papers on the psychoacoustics (Diana Deutsch comes to mind http://deutsch.ucsd.edu/psychology/pages.php?i=101 but i see she has over 200 publications), and dbx white papers that generously described various techniques. Read schematics but don't automatically assume that the old ways were high art. In a number of cases it was serendipity of the only component choices available that happened to play very nice together.
I just don't have the resources or ability to independently get some of this stuff. It's tribal knowledge that is being lost, I'm afraid.
If it still had value it wouldn't be lost, it would continue being used. The compelling demand for tape NR, went away with magnetic tape as the dominant recording medium. Using modern digital there is not need to squeeze the audio to extend dynamic range.  Nowadays dynamic processing is more of a subjective stylistic mix related exercise where being transparent may not be the ultimate object. FWIW pretty much anything we could do in analog can be done in digital (including adding back in euphonious distortion, but I don't roll that way).

JR

 
 
Don't get me wrong, I know there's no shortcuts. Lots of respect for the expertise of you and other posters on here for that. There's just no real way to do "apprenticeship" type work on these things.

I get you on the digital side, but even then it seems to me (maybe wrong) that understanding the analog domain would make manipulating audio in the digital side easier.
 
dogears said:
Don't get me wrong, I know there's no shortcuts. Lots of respect for the expertise of you and other posters on here for that. There's just no real way to do "apprenticeship" type work on these things.
I don't really agree. Most of us old designers learned by reading the very few schematics we could put our hands on (no google then, can you believe that?  ;) ), and putting whatever knowledge we acquired at school  in practice on the bench, with modest resources. One was lucky if he had a multimeter then, and I put my hands on a 'scope for the first time in my 20th year.
As I wrote earlier, study teh schemos of all the great (or not so great) classics.

I get you on the digital side, but even then it seems to me (maybe wrong) that understanding the analog domain would make manipulating audio in the digital side easier.
That is definitely true. Failure to observe that can result in disasters (did I tell about the first VST?)  :eek:
Apart from delay, most processes in digital equipment is an emulation of analog circuits.
 
dogears said:
Don't get me wrong, I know there's no shortcuts. Lots of respect for the expertise of you and other posters on here for that. There's just no real way to do "apprenticeship" type work on these things.
some of these markets have always been niche (small) markets, piggy backing on consumer audio or general electronics technology.

I expect these jobs still exist just using modern technology. Being a niche market, not many jobs. Back in the day some of these designs were done by studio techs who were busy trying to keep the less than reliable analog gear operational. Modern digital gear is almost disposable, not very fixable.
I get you on the digital side, but even then it seems to me (maybe wrong) that understanding the analog domain would make manipulating audio in the digital side easier.
It is important to understand the analog receiver (human audition/perception). Diana Deutsch and her peers have accumulated a massive body of work... Not simple connect the dots design advice, but numerous pearls in the content, that requires some digging.

JR
 
I know in my experience as a young engineer design standards, internal guidelines / best practices, and previous designs to reference are basically solid gold for learning. ***Especially*** if the engineer that did the earlier design is around to say - hey , what the heck is going with this part?  :p

Without that you spend as much time reinvinting the wheel as designing - when a set of rules of thumb would propel your progress tenfold.
 
dogears said:
Without that you spend as much time reinvinting the wheel as designing - when a set of rules of thumb would propel your progress tenfold.
Don't negate the virtues of mistakes. they leave a solid imprint of how your mindset can put you in an erroneous path, and they force you to drift from the main direction and consider avenues you wouldn't have thought of.
 
abbey road d enfer said:
Don't negate the virtues of mistakes. they leave a solid imprint of how your mindset can put you in an erroneous path, and they force you to drift from the main direction and consider avenues you wouldn't have thought of.

A wise man once said "The only people not making mistakes are those doing nothing"

Cheers

Ian
 
I think this is a great thread, although ive had to read more than a few posts several times to try and get a better comprehension. I have a decent understanding of amps, power, eq, but I too have a hard time wrapping my head around the circuits used to make compression behave as it does. I'll just keep reading, thank you.
 
Rocinante said:
I think this is a great thread, although ive had to read more than a few posts several times to try and get a better comprehension. I have a decent understanding of amps, power, eq, but I too have a hard time wrapping my head around the circuits used to make compression behave as it does. I'll just keep reading, thank you.
Think of a compressor as something that does what you do when you ride vocals. Too loud, you back off the fader, too soft you crank it.
The time it takes you to realize it's too loud and tell you hand is the attack time, the time it takes you to figure out it's becoming too soft and move up the fader is the release time. Now there are many things that can interfere with those timings, like how smooth is the fader, how alert you are. It's the same with a compressor, VCA's react almost instantly, but rms detectors are not instant, opto cells are intrinsically slow. Anyway, instant reaction is not welcome because it would create an abrupt level change that the ear would consider as noise.
If you're familiar with the notion of feedback in psychology, you get it.
 
abbey road d enfer said:
Think of a compressor as something that does what you do when you ride vocals. Too loud, you back off the fader, too soft you crank it.
The time it takes you to realize it's too loud and tell you hand is the attack time, the time it takes you to figure out it's becoming too soft and move up the fader is the release time. Now there are many things that can interfere with those timings, like how smooth is the fader, how alert you are. It's the same with a compressor, VCA's react almost instantly, but rms detectors are not instant, opto cells are intrinsically slow. Anyway, instant reaction is not welcome because it would create an abrupt level change that the ear would consider as noise.
If you're familiar with the notion of feedback in psychology, you get it.
a little simplistic but true.

A similar explanation for tape noise reduction involves riding two faders, one feeding the input to the tape path, and the second on the output coming back. When the program input gets quiet you boost it up higher, upon playback you lower it back to get the original level signal return, but with the tape noise floor reduced a like amount.

Of course the side chain magic involves automatically decoding the encoded, squeezed audio for accurate playback tracking.

JR
 
The Trident again:
abbey road d enfer said:
I wouldn't follow you in this direction...
So the other direction then, I deduce.
Does the description below nail it ? I don't know ANYthing about FET compressors.

Once CV voltage is too low and the diode cuts off the Ratio pot (max of 1Meg), the sidechain is shunted to ground via what looks like a 1 Meg resistor and the Gain Reduction trimmer (450K), sending max ratio voltage (i.e. 20:1 or more with max of 1M45 possible) to the FET gate (meaning max reduction thete).

Once the audio gets louder again, CV rises and the diode starts conducting again, making CV go back to the ratio dialled in on the faceplate ratio knob (btw. 1:1 to 20:1).

In general, the unit should be less 'noisy' when no audio is coming in on the input.
 
Script said:
The Trident again:So the other direction then, I deduce.
Does the description below nail it ? I don't know ANYthing about FET compressors.

Once CV voltage is too low and the diode cuts off the Ratio pot (max of 1Meg), the sidechain is shunted to ground via what looks like a 1 Meg resistor and the Gain Reduction trimmer (450K), sending max ratio voltage (i.e. 20:1 or more with max of 1M45 possible) to the FET gate (meaning max reduction thete).

Once the audio gets louder again, CV rises and the diode starts conducting again, making CV go back to the ratio dialled in on the faceplate ratio knob (btw. 1:1 to 20:1).
That's correct.
Very often the actual ratio is not the ratio that's dialled. Obviously going from under threshold to over threshol happens continuously, as nature hates discontinuities, so the ratio goes from zero to nominal over a range that depends on how the side-chain is designed. Just the same, when dealing with very high levels, the side-chain may run out of gas, so the ratio would fall behind nominal. On top of that, many compressors have a rising ratio, whether the designer intended it or not. All feedback compressors do that (in fact a UREI 1176 has an actual ratio that is not excatly what's printed on the front panel). VCA+RMS compressors intrinsically do not exhibit this behaviour, but many designers intendedly use a non-linear side-chain for providing the well-known soft knee characteristic.

In general, the unit should be less 'noisy' when no audio is coming in on the input.
This not often the case; most compressors actually do pump the "silence". It can be ignored or dealt with with a noise-gate or digita editing.
 
I've got a really basic question.

How is the sidechain gain reduction law derived for various VCA styles?

I assume for a fet you use the Id/Vgs curve along with the series resistor to determine the equivalent voltage divider per volt applied as control voltage?

How is it done for a diode bridge, like the 2254?
 
dogears said:
I've got a really basic question.
basic but not easy...
How is the sidechain gain reduction law derived for various VCA styles?
VCA literally means voltage controlled amplifier and the most common example (like dbx/THAT corp VCAs) have a gain law where X mV of control voltage results in Y dB of gain change.  That was easy, too easy.  There are many different variant gain control technologies often lumped together with VCAs.  Some VCAs use two LTP (long tail pairs) with the first in a feedback loop and the second not. The ratio of current between the two LTPs controls gain linearly, as opposed to log controlled VCAs (like THAT/dbx). 

Then we have OTAs (operational transconductance amps) that are current controlled, multipliers. An AC input voltage modulates an AC output current, that scales linearly with control current.

Don't forget (opto) light controlled resistance elements, in a voltage divider like JFET limiters.

There are more...
I assume for a fet you use the Id/Vgs curve along with the series resistor to determine the equivalent voltage divider per volt applied as control voltage?
most FET limiters I did were pretty crude feedback topologies, where output rising above some threshold, fed back and increased (or decreased) some gate voltage to modulate the effective resistance of a JFET in a voltage divider configuration.

I shared this story before, back in the 70's I figured out how to linearize the transfer function of a JFET limiter using a dual JFET, the first was inside a NF loop, the second tracked a linear control voltage reasonably well, but I abandoned this approach for the VCAs available back then that weren't nearly as good as modern VCAs.
How is it done for a diode bridge, like the 2254?
sorry I don't recognize what that is and am too lazy to search.

JR
 
Thanks John. I was using VCA in the general sense - any voltage controlled amp.

I sort of get the THAT VCAs, and anyway however the mV / dB gain reduction is derived it's there on the datasheet.

I'm curious in how you would derive that relationship for a FET or diode bridge (the 2254  being the Neve diode bridge). 
 
dogears said:
Thanks John. I was using VCA in the general sense - any voltage controlled amp.

I sort of get the THAT VCAs, and anyway however the mV / dB gain reduction is derived it's there on the datasheet.

I'm curious in how you would derive that relationship for a FET or diode bridge (the 2254  being the Neve diode bridge).

I have only used JFET limiters in NF configuration... slope control is general more/less not some strict math relationship like dbx VCAs.

JR
 
Back
Top