THE ENGINEER'S THUMB... At last, a better compressor!  [documentation]

Original effects with schematics, layouts and instructions, freely contributed by members or found in publications. Cannot be used for commercial purposes without the consent of the owners of the copyright.
User avatar
mictester
Old Solderhand
Information
Posts: 2923
Joined: 11 Sep 2008, 20:29
my favorite amplifier: Mesa Boogie, Roost Sessionmaster, AC30
Completed builds: Hundreds! Mostly originals, a few clones and lots of modifications.
Location: Somewhat closer to Amsterdam than before!
Has thanked: 32 times
Been thanked: 844 times
Contact:

Post by mictester »

I'm not claiming originality for the concept (though I was doing "switched mode" gain control in the 70s). The real issue is getting a sufficiently linear envelope-to-PWM conversion. It doesn't matter for things like the MXR Autowah (which is quite a clever use of CMOS but is a bit too component-dependent for my taste), but when you're trying to do accurate compression, that conversion law is crucial.

Just for the sake of amusement, I lashed up your gain cell on a solderless breadboard and found that I get an almost reciprocal voltage to gain curve. If this could be made accurately logarithmic, the synthesiser chaps will think you're wonderful! I didn't try your feedforward level sensing - I used a take-off point after the gain cell (in the more conventional feedback mode) and the subjective results were actually very good. A little tweaking of the bias components for the 13700, and the noise and distortion improve usefully, but at the expense of more components and a higher supply voltage (it works better at 15V). A little more work on the sidechain, and we're close to the Morley circuit!

With the brief experiment I tried, I think the circuit shows some promise, but I'm still not convinced by the feedforward topology. We really do need that delay line in there!
"Why is it humming?" "Because it doesn't know the words!"

User avatar
Guilherme
Information
Posts: 47
Joined: 11 Jul 2007, 16:15
Has thanked: 5 times
Been thanked: 4 times

Post by Guilherme »

merlinb wrote: Almost a nice idea, but unfortunately there's rather a lot of "prior art":
Hey, that's mine :) It's from 1990. There's an error on the drawing, swap inputs +/- on U9A.

I only tested it on bass, I liked it.

BTW later I learned there was a (supposedly very nice) MXR studio limiter that used a switching VCA, the MXR 136 Dual Limiter.
Hard to find schematic:
http://www.gearslutz.com/board/attachme ... ualsch.pdf

User avatar
merlinb
Solder Soldier
Information
Posts: 176
Joined: 18 Nov 2010, 17:39
Been thanked: 160 times

Post by merlinb »

mictester wrote:If, however, it overshoots and sounds as peculiar as I suspect it will,
I set up a test sine wave and played it back through the my compressor. The upper image is the original audio and consists of of a 30mVp-p 200Hz signal interrupted by a one second burst at 100mVp-p.
Image

Here is a zoomed in picture of the attack phase- note the complete lack of overshoot!!!
Image

User avatar
mictester
Old Solderhand
Information
Posts: 2923
Joined: 11 Sep 2008, 20:29
my favorite amplifier: Mesa Boogie, Roost Sessionmaster, AC30
Completed builds: Hundreds! Mostly originals, a few clones and lots of modifications.
Location: Somewhat closer to Amsterdam than before!
Has thanked: 32 times
Been thanked: 844 times
Contact:

Post by mictester »

With the transient at the beginning of every guitar note, you're almost certain to overshoot! Try a more realistic signal.
"Why is it humming?" "Because it doesn't know the words!"

User avatar
merlinb
Solder Soldier
Information
Posts: 176
Joined: 18 Nov 2010, 17:39
Been thanked: 160 times

Post by merlinb »

mictester wrote: Small hint to the initiated - look at "A Voltage-Controlled Low-Distortion Audio Gain Cell" in the JAES in 1975!
This reference appears to be in error- I can find no such paper.

User avatar
Guilherme
Information
Posts: 47
Joined: 11 Jul 2007, 16:15
Has thanked: 5 times
Been thanked: 4 times

Post by Guilherme »

If I had to guess how ANY compressor with a fast attack and a relatively slow release would behave with a signal that has a short, large transient, a fast decay and then a much lower sustain level, such as aguitar note, I'd bet it would completely lack overshoot, like above, BUT it would have kind of an "undershoot" after the attack, so just after the attack there would be a "swell" while the gain slowly rises to bring the sustain part of the note to the same level as the attack. Or, in other words, "pumping". Like this:

Image

Here's a paper from the IRE that helped a lot when I was designing the PWM thingie above:
"A Review of Automatic Gain Control Theory"
D. V. Mercy
The Radio and Electronics Engineer, Vol. 51, No. 11/12, pp. 579-590, November/December 1981
http://localhostr.com/file/oGJqWiK/mercy.zip

User avatar
merlinb
Solder Soldier
Information
Posts: 176
Joined: 18 Nov 2010, 17:39
Been thanked: 160 times

Post by merlinb »

Guilherme wrote:If I had to guess how ANY compressor with a fast attack and a relatively slow release would behave with a signal that has a short, large transient, a fast decay and then a much lower sustain level, such as aguitar note, I'd bet it would completely lack overshoot, like above, BUT it would have kind of an "undershoot" after the attack, so just after the attack there would be a "swell" while the gain slowly rises to bring the sustain part of the note to the same level as the attack. Or, in other words, "pumping".
That's true, but it is normal behaviour for any music compressor. In essence, your example is exactly the same as the image I posted, but with the loud passage made shorter; you have shifted the release phase over to the left; the release phase now occurs striaght after the attack phase. Mictester seems to be confusing music compressors with broadcast compressors, which are not allowed to behave this way.

Thats for the article! :D

User avatar
mictester
Old Solderhand
Information
Posts: 2923
Joined: 11 Sep 2008, 20:29
my favorite amplifier: Mesa Boogie, Roost Sessionmaster, AC30
Completed builds: Hundreds! Mostly originals, a few clones and lots of modifications.
Location: Somewhat closer to Amsterdam than before!
Has thanked: 32 times
Been thanked: 844 times
Contact:

Post by mictester »

Broadcast compressor / limires are frequently used in recording studios. The reason is that they provide a vice-like grip on over-levels and back in the days of analogue tape (like I still record on) it was a great way of staying out of saturation.

There is no practical difference between a broadcast limiter and a competently designed guitar effect - broadcast units usually employ delay lines to "see the peaks coming" and thereby eliminate pumping and overshoot issues. There is no reason why this approach cannot be applied to a guitar effect. I have a compressor / limiter for guitar here that uses an attack time of 8ms and a delay of just slightly more than that. The delay is imperceptible to the un-aided ear (you can clearly see it on a 'scope), but there's no overshoot, no pumping and little distortion or noise. I'm working on simplifying the circuit somewhat, and thereby making it cheap and simple for the home constructor.... watch an adjacent thread for details coming soon (when I have a little time off).
"Why is it humming?" "Because it doesn't know the words!"

User avatar
merlinb
Solder Soldier
Information
Posts: 176
Joined: 18 Nov 2010, 17:39
Been thanked: 160 times

Post by merlinb »

mictester wrote: There is no reason why this approach cannot be applied to a guitar effect.
Yes, but when was the last time you saw a feedforward+delay line guitar compressor pedal?? Never, that's when.

User avatar
uncleboko
Cap Cooler
Information
Posts: 509
Joined: 25 Mar 2008, 20:19
Has thanked: 58 times
Been thanked: 41 times

Post by uncleboko »

Surely -

Most folks' idea of a guitar compressor is that it should prolong the note being played, without too much noise being introduced and the listener WILL hear the effect.

A broadcast compressor/limiter on the other hand (una parva distinta de pescados) operates as a brick wall so that the audio level will not rise above a certain preset amount, and its action will preferably not be heard the listener.

User avatar
Guilherme
Information
Posts: 47
Joined: 11 Jul 2007, 16:15
Has thanked: 5 times
Been thanked: 4 times

Post by Guilherme »

A lot of people complain about digital fx, recording interfaces or guitar synths that have latencies around 8-10 ms, or even less.

Studio gear, if you're not doing "software monitoring", can have a much longer latency with no problem.

For a guitar compressor, no latency is way more important than an instant attack that I won't want to use anyway (you need some attack left or it will not sound like a guitar and it will disappear in a mix).

User avatar
joe_pineapple
Breadboard Brother
Information
Posts: 74
Joined: 07 Sep 2009, 22:31
my favorite amplifier: eden highwayman
Completed builds: woolish mammoth, a variety of germanium and silicon fuzz factories, machine, parallel universe, random # generator, dirty sanchez, swamp thang, SHO, DLS, peach fuzz, cream puff, prunes and custard, red llamalike,
Location: san francisco
Has thanked: 105 times
Been thanked: 2 times
Contact:

Post by joe_pineapple »

Guilherme wrote:A lot of people complain about digital fx, recording interfaces or guitar synths that have latencies around 8-10 ms, or even less.

Studio gear, if you're not doing "software monitoring", can have a much longer latency with no problem.

For a guitar compressor, no latency is way more important than an instant attack that I won't want to use anyway (you need some attack left or it will not sound like a guitar and it will disappear in a mix).
yeah, i think that's really important.
instant attack w/small latency for studio playback, sound editing and broadcast are all great. but it's no fun to play an instrument with latency in yr signal.

the instant attack does sound good tho... as an effect
"..the finished product is just the garbage left over from the work." R.G. Keen

User avatar
Duckman
Opamp Operator
Information
Posts: 1492
Joined: 20 May 2009, 01:45
Has thanked: 320 times
Been thanked: 134 times

Post by Duckman »

:lol: :thumbsup

User avatar
Guilherme
Information
Posts: 47
Joined: 11 Jul 2007, 16:15
Has thanked: 5 times
Been thanked: 4 times

Post by Guilherme »

merlinb wrote: Thats for the article! :D
Here some more joyful reading on this fascinating subject (at least for a nerd like me):

"Audio Dynamic Range Compression for Minimum Perceived Distortion"
Barry A. Blesser
IEEE Transactions on Audio and Electroacoustics, vol. AU-17, No. 1, March 1969

"On Transient Distortion in Hearing Aids with Volume Compression"
Igor V. Nabelek
IEEE Transactions on Audio and Electroacoustics, vol. AU-21, No. 3, June 1973

The PWM thing above was published here:

RIBEIRO, G. F. ; NIGRI, A. M. . Compressor/Limitador de Áudio usando PWM. In: IV CRICTE - UFRJ, 1990, Rio de Janeiro, RJ, 1990.

User avatar
pedalgrinder
Cap Cooler
Information
Posts: 548
Joined: 29 Aug 2011, 06:05
my favorite amplifier: Ampeg SVT
Completed builds: pedals with pots switches and everything
Location: South Australia Adelaide
Has thanked: 198 times
Been thanked: 172 times

Post by pedalgrinder »

I also built the elector tda7052a compressor and found it to be quite good all i heard was to much distortion and not clean enough but on a bass i love it. Why the hell would anyone worry about distortion and fuzz on this forum every second entry is a fuzz or distortion. Hahhahahah i've been burnt on the stake a few times for mentioning about to much distortion pedals. Hhahahahahaha. Hey merlinb i reckon you've done great job. I just reckon your release could be a longer time constant and your attack could use a seperate buffer. like in this schematic i found in a magazine for a tv volume controller
tv volume compressor.jpg
Hey Mictester i see you've reposted the schematic i sent you via email of the mxr 143 limiter but i never heard back from you. Where did you get to? something doesn't seem right with that schematic although i did build it on vero and vero can be a nightmare for mistakes but i just didn't want to play ball. But anyhow check out the attachment i reckon this may give some good ideas for a sidechain that can really has some variation.
What's the best thing about fat chicks and scooters? There both fun to ride around until your mates find out!

User avatar
HEAD
Resistor Ronker
Information
Posts: 442
Joined: 16 Dec 2008, 00:13
Location: Leipzig, Germany
Has thanked: 24 times
Been thanked: 49 times

Post by HEAD »

I've ended up building the tda7025a version kindly provided by Pjotr(?) from HEXE effects: http://www.hexeguitar.com/diy-tadek_e
Works pretty nicely and without any noticable distortion - at least for my deaf ears.

User avatar
jstbrowsin
Grease Monkey
Information
Posts: 108
Joined: 30 Aug 2009, 00:53
Has thanked: 65 times
Been thanked: 4 times

Post by jstbrowsin »

um that seems to be a dead link :(

User avatar
pedalgrinder
Cap Cooler
Information
Posts: 548
Joined: 29 Aug 2011, 06:05
my favorite amplifier: Ampeg SVT
Completed builds: pedals with pots switches and everything
Location: South Australia Adelaide
Has thanked: 198 times
Been thanked: 172 times

Post by pedalgrinder »

yeah i've had a look at his design looks great does a really nice job! Just as a bass player not what iam after. I wasn't entirely happy with his sidechain for me! Iam not knocking his design in anyway iam only saying about taste entirely different thing. I just found the elector version had a lot of room to move as in adjusting the sidechain how wanted similar to the tv volume compressor sidechain i posted earlier. As much as thats not for guitar i really like it's sidechain. Mainly as you can fully control via having a different cap time constant and a buffer between for attack and release. Iam sure theres other designs that take this approach i just haven't come across a schematic that does it that way. But yeah i reckon for a guitar the hexe design would be right up most guitarists alley more sustain orientated.
What's the best thing about fat chicks and scooters? There both fun to ride around until your mates find out!

User avatar
HEAD
Resistor Ronker
Information
Posts: 442
Joined: 16 Dec 2008, 00:13
Location: Leipzig, Germany
Has thanked: 24 times
Been thanked: 49 times

Post by HEAD »

For me it does work. Try searching for "hexe effects", then browse to diy section "tadek compressor".

User avatar
pedalgrinder
Cap Cooler
Information
Posts: 548
Joined: 29 Aug 2011, 06:05
my favorite amplifier: Ampeg SVT
Completed builds: pedals with pots switches and everything
Location: South Australia Adelaide
Has thanked: 198 times
Been thanked: 172 times

Post by pedalgrinder »

yeah thats the one iam refering to the tadek really nice unit. I imagine most guitarist's would love it!
What's the best thing about fat chicks and scooters? There both fun to ride around until your mates find out!

Post Reply