PWM compressor

[JohnRoberts]

I have never quite understood the attraction of paralel compression.

Cheers

Ian

[veer feed]  ;D ;D ;D  If there was some there there, you could make a dedicated compressor to do the same thing as parallel compression.  [/veer]

This discussion needs to appreciate that the VCA or gain element is not the hard part about dynamic processor design. Less now than ever. 

A PWM gain element needs to be good enough that it doesn't introduce a sound of it's own... (I probably already said this, but I'm old and forget.) If you are expecting some magic from PWM that may be a dry hole. I've used PWM to made sine waves for two generations of my drum tuners...  Not the purest sine waves but cheap (free inside the cheap  processors I used) and clean  enough to wiggle drumheads without exciting false notes.

JR

 

[ruffrecords]

A PWM gain element needs to be good enough that it doesn't introduce a sound of it's own... (I probably already said this, but I'm old and forget.) If you are expecting some magic from PWM that may be a dry hole.

JR

All I said at the start of this thread is that it looks like you could do a PWM compressor using a micro these days. Amazing how these topics take on a life of their own.

Cheers

Ian

 

[zamproject]

All I said at the start of this thread is that it looks like you could do a PWM compressor using a micro these days. Amazing how these topics take on a life of their own.

Cheers

Ian

Hi

Maybe it's because ppl (like me?!?) reading Ian asking "what about a uControled PWM compressor ? " think that you have a plan to work on it  :-X

best
Zam

 

[desol]

Where there is a will there is a way....

 

[abbey road d enfer]

. When the compressor works, the gain reduction is simply limited to a maximum of 6dB.

Which is huge according to mastering standards where 1 or 2 dB compression seems to be nirvana, even when it needs to repeated on 3 or 4 different flavours (plug-ins are so cheap), of course followed by brickwall-limiting before actually being released...

 

[Andy Peters]

I am probably missing something but it seems to me that parallel compression is actually rather limited. If it involves mixing an original signal with a compressed one, then, if the two signals are equal when there is no compression the overall signal level is just 6dB higher. When the compressor works, the gain reduction is simply limited to a maximum of 6dB.

Take one channel, compress it to hell, push the makeup gain up so you're back at unity, and then add in the original signal. voila! 6 dB gain. OF COURSE IT SOUNDS BIGGER.

 

[JohnRoberts]

Take one channel, compress it to hell, push the makeup gain up so you're back at unity, and then add in the original signal. voila! 6 dB gain. OF COURSE IT SOUNDS BIGGER.

Double patching an input without compression gives you +6 dB.  ;D ;D ;D

JR

 

[abbey road d enfer]

Double patching an input without compression gives you +6 dB.  ;D ;D ;D

JR

Spoilsport!  ;D

 

[Andy Peters]

Double patching an input without compression gives you +6 dB.  ;D ;D ;D

Ssssssh, don't burst their bubble!!

 

[Andy Peters]

LOL.  It freaks out the younger live engineers that have never used an analog desk when I can get "that" drum sound live.  They can't figure out how to do it on their SC48.

I get "that" drum sound live on an SC48 by having an excellent drummer who knows how to tune his instrument, and by using decent mics on the kit. I am also old enough to have used and abused analog desks, and I say this: gimme the SC48.

 

[Gold]

I never did get parallel compression. Especially as a 'more transparent' option. Crush the crap out of the program to the point of obvious distortion and then mix it back in. How does that help?

I'd love to see a PWM compressor project.

 

[kambo]

I never did get parallel compression. Especially as a 'more transparent' option. Crush the crap out of the program to the point of obvious distortion and then mix it back in. How does that help?

I'd love to see a PWM compressor project.

not in album music to listen, or itunes release...
but in SFX/movie/trailer music i use it all the time (unless its going to be released as music to buy)

 

[kambo]

I get "that" drum sound live on an SC48 by having an excellent drummer who knows how to tune his instrument, and by using decent mics on the kit. I am also old enough to have used and abused analog desks, and I say this: gimme the SC48.

haha, thats very typical engineer comment, when they cant decent sound out of a drum kit..

 

[desol]

The strangest thing(for me) about this Pye copy...is that unlike any of my other units(apart from possibly the la2a...sorta) is that no matter how much signal I cram into the stupid thing, it just keeps getting better?? It doesn't cave like other units(brand name x). One can just keep pushing signal into it, and it's just keeps getting better and better and better....instead of worse? Kinda like the energizer bunny.

Counter-intuitive, to say the least. I can see what the big deal is now about this...type of unit.

I can also see this type of thing being formidable, switched up, with digitally controlled mastering clothes on. Doh.

 

[usekgb]

I've been doing a little bit of thinking on this.  What if we were to multiplex two PWM pins together to achieve a higher PWM resolution with a given frequency?  The PWM output waveform would not be smooth, but the overall duty cycle would have a resolution equal to the sum of both PWM resolutions.  Thoughts?

 

[abbey road d enfer]

I've been doing a little bit of thinking on this.  What if we were to multiplex two PWM pins together to achieve a higher PWM resolution with a given frequency?  The PWM output waveform would not be smooth, but the overall duty cycle would have a resolution equal to the sum of both PWM resolutions.  Thoughts?

What do you mean "multiplex two PWM pins"? Can you draw a sketch of your idea? And explain the math behind "resolution equal to the sum of both PWM resolution"?

 

[JohnRoberts]

I've been doing a little bit of thinking on this.  What if we were to multiplex two PWM pins together to achieve a higher PWM resolution with a given frequency?  The PWM output waveform would not be smooth, but the overall duty cycle would have a resolution equal to the sum of both PWM resolutions.  Thoughts?

I have actually done this to make sinewaves from a cheap 8 bit micro. I scaled one PWM output down a factor of 8 or 16. I don't remember which.  I think I tried 16 then fell back to 8. A lot of work for only three more bits. It works but for lowest distortion the precision between the scaling resistors matters.

As I've shared before with compression the side chain matters more than the gain element, but PWM promises to be an unwanted/unneeded sound variable.  Any useful gain element needs to sound like nothing at all, a straight wire with cut.

JR

PS: For my later generation design I just used two 16b PWM outputs with equal weighting (to drive a class D amp IC differentially). The sine-wave was more than clean enough to wiggle just the correct notes in the drumheads. If too dirty you can get unintended resonances from harmonic distortion. Not clean enough for a studio audio path.  :

 

[emrr]

Parallel compression.  :  Stopped doing that years ago as I got better compressors and (likely) also got better at setting them up. 

Nowadays having a parallel mix knob on a unit is generally the make or break part of the decision to buy for many of the younger generation. 

 

[dmp]

Trying parallel compression was a huge "ah ha" moment for me with compressors. It's not for transparently controlling the dynamics; you're using the compressor as an effect. We usually do a compressor working hard - blended in just until you're liking what you hear. I don't think you can get the same thing at all using a compressor directly.

 

[Rocinante]

I agree I always used it (albeit rarely) as an effect.  For instance; It can blend bass and kick nicely within certain genres of music (hip hop, industrial, etc...) and acts as a blend knob.