Page 1 of 3

Any opinions about Nonlinear-Compressor XI.ny ?

Posted: Fri Jul 18, 2014 8:20 pm
by Paul L
Someone pointed me to this plug-in that is supposed to be a compressor, limiter, or expander depending on the settings. It is a few years old but I can't find discussion of it anywhere. I think the controls are weird, and the code is a bit weird. (And inefficiently written to implement the underexplained mathematical formulas.)

http://audacity.238276.n2.nabble.com/No ... 69449.html

Re: Any opinions about Nonlinear-Compressor XI.ny ?

Posted: Fri Jul 18, 2014 10:20 pm
by kozikowski
The description sounds like Chris's Compressor. Non-linear compander is how off-air broadcast sound chains work. Companding is determined by the show itself.

"These are the legal specifications and requirements for the air show and it doesn't matter what the original show was."

They're difficult to fudge. I once worked for a station that was given a mild warning/suggestion for loud overmodulation. The station was so terrified that they later got an actual broadcast standards violation citation for undermodulation. Obviously this was before the Loudness Wars.

I should have kept a copy of it. Nobody would believe me.

Koz

Re: Any opinions about Nonlinear-Compressor XI.ny ?

Posted: Fri Jul 18, 2014 10:27 pm
by Paul L
I have read the code for both. This other is very unlike Chris's Compressor. I can't detect that it has any reputation good or bad.

The fancy part of Chris's is defining the envelope just so, but then a simple gain curve is applied. The fancy part of this one is all in the gain curve -- but strangely it does nothing to define an evelope, not even a simple snd-avg. I think that sounds like a recipe for distorting the timbre of everything.

Re: Any opinions about Nonlinear-Compressor XI.ny ?

Posted: Fri Jul 18, 2014 11:07 pm
by kozikowski
It might be. The process is more difficult than it seems. Every few years, someone would come up with a different way to process the broadcast sound chain. CBS Labs, Dorough, etc. The last one (I forget the name) built the companding into the electronics for the broadcast transmitter. That was a very unusual step because it required replacing a portion of the certified, expensive transmitter with a separate product instead of just buying a little box with flashing lights on it. However, they did such a good job acoustically that they mopped the floor with everybody else.

That's the goal. The show mostly sounds the same as before, except it now conforms to some desirable standard.

That sounds simple, doesn't it?

Koz

Re: Any opinions about Nonlinear-Compressor XI.ny ?

Posted: Sat Jul 19, 2014 9:28 am
by Robert J. H.
If I remember correctly, the effect is a fancy and complicated way to do some wave shaping. It is nearer to a guitar distortion/compression effect than a multi-tool compressor.
By the way, have you looked at RBD's compressor? The calculation of the non-linear 2-fold compression/expansion curve is quite intriguing.
It's in the library section of Nyquist.

Re: Any opinions about Nonlinear-Compressor XI.ny ?

Posted: Sat Jul 19, 2014 2:34 pm
by Paul L
Can you give me a more exact pointer? Is it undocumented? Is it in C Or Lisp?

Re: Any opinions about Nonlinear-Compressor XI.ny ?

Posted: Sat Jul 19, 2014 2:41 pm
by Robert J. H.
Paul L wrote:Can you give me a more exact pointer? Is it undocumented? Is it in C Or Lisp?
compress.lsp
(13.3 KiB) Downloaded 61 times

Re: Any opinions about Nonlinear-Compressor XI.ny ?

Posted: Sat Jul 19, 2014 3:16 pm
by Paul L
I found it in Nyquist 3.08 source. So nobody adapted it into a .ny for Audacity? But that should be easy enough for me to try.

Re: Any opinions about Nonlinear-Compressor XI.ny ?

Posted: Sat Jul 19, 2014 3:18 pm
by steve
Igor likes his "magic formulae".

Re: Any opinions about Nonlinear-Compressor XI.ny ?

Posted: Sat Jul 19, 2014 3:31 pm
by Paul L
steve wrote:Igor likes his "magic formulae".
Robert is right, what I am looking at is billed as a compressor/expander but without any smoothing at all of the envelope, wave shaper is the better name.

And he could have written it more efficiently with only 5 calls to mult instead of 12.