can someone explain feedforward vs feedback compression?
-
- KVRist
- Topic Starter
- 49 posts since 22 Sep, 2012
Ok so feedforward makes a lot of sense to me. 1) The compressor reads the peak level of the incoming signal or possibly the rms. 2) The compressor applies volume reduction according to the ratio, attack and release.
Feedback doesn't make as much sense. As far as I understand the compressor is reading the output signal - right? Surely this creates a nightmarish cause and effect loop? I don't see why this would be desirable. It seems that a feedback compressor would be much harder for the user to control.
Can someone shed any light on this?
Feedback doesn't make as much sense. As far as I understand the compressor is reading the output signal - right? Surely this creates a nightmarish cause and effect loop? I don't see why this would be desirable. It seems that a feedback compressor would be much harder for the user to control.
Can someone shed any light on this?
-
- KVRAF
- 3477 posts since 27 Dec, 2002 from North East England
You are getting a cause and effect loop, but not in the nightmarish runaway feedback sense because the threshold value is essentially being raised when compression is taking place, not lowered. It is perhaps more difficult to control instantaneous dynamic changes with the extreme and predictable precision possible with a feedforward compressor, but on the flipside it's perhaps more difficult to make feedback compression sound bad because this 'adaptive' threshold behaviour tends to result in compression that 'rides' the signal very smoothly. It's also the reason why it's difficult to get large amounts of gain reduction out of a feedback compressor. There's only so much the signal can be squashed before the output is below the threshold at which compression begins. I find feedback compressors to be very forgiving and better able to tolerate 'non-optimal' settings without creating obvious gain reduction artifacts, so they're usually my go-to choice when small amounts of clean gain reduction are required.
-
- KVRian
- 1290 posts since 13 Mar, 2007
"Feedback uses the output of the
compressor as the detector input and then feeds
that level information back to the compression circuit
to determine gain reduction. This can allow initial
transients to leak through, but can also yield a very
different smooth compression that is highly desirable for
some instruments."
It is important to remember the evolution of dynamics processing' It started with limiters where not at all brickwall !! They were really more like "levelers" or something .. They brought the bottom up whilst they lowered the top. There is a very good history over at the RANE audio site . here
http://www.rane.com/pdf/ranenotes/Dynam ... essors.pdf
Look at the last page... upward and downward expansion sort of a "hinge or tilt".
So it's just another flavor of dynamics control that uses a pretty light touch ,,,,, Engineers reach for different comps because they all have different strenghts , here's a good read on this ...
http://www.soundonsound.com/sos/sep09/a ... essors.htm
Cheers
compressor as the detector input and then feeds
that level information back to the compression circuit
to determine gain reduction. This can allow initial
transients to leak through, but can also yield a very
different smooth compression that is highly desirable for
some instruments."
It is important to remember the evolution of dynamics processing' It started with limiters where not at all brickwall !! They were really more like "levelers" or something .. They brought the bottom up whilst they lowered the top. There is a very good history over at the RANE audio site . here
http://www.rane.com/pdf/ranenotes/Dynam ... essors.pdf
Look at the last page... upward and downward expansion sort of a "hinge or tilt".
So it's just another flavor of dynamics control that uses a pretty light touch ,,,,, Engineers reach for different comps because they all have different strenghts , here's a good read on this ...
http://www.soundonsound.com/sos/sep09/a ... essors.htm
Cheers
Financial solvency and KVR Mix as well as oil and water.
-
- KVRist
- Topic Starter
- 49 posts since 22 Sep, 2012
Ok, thanks for the responses that somewhat confirms what I suspected. I can't imagine ever wanting to use feedback compression but it's nice to know what it is. Also thanks for the links and the ideas.
Still interested in what anyone else has to say!
My next question is about lookahead times but I'll start another thread.
Still interested in what anyone else has to say!
My next question is about lookahead times but I'll start another thread.
-
- KVRer
- 12 posts since 19 Dec, 2013 from Los Angeles, CA
Read the user manual to this:
http://www.tokyodawn.net/tdr-feedback-compressor-2/
If it weren't for the latency, I would start with a feedback compressor quite often. In practical use, I replace a zero latency compressor on various buses with a feedback compressor during mixing.
http://www.tokyodawn.net/tdr-feedback-compressor-2/
If it weren't for the latency, I would start with a feedback compressor quite often. In practical use, I replace a zero latency compressor on various buses with a feedback compressor during mixing.