hi
i enjoy using Buzzroom compressors, and one of them is called a "feedback compressor"...
who knows what's that mean ??
it might be obvious for english native users, but i don't understand the meaning here...

does anyone has an explanation ?
thanxx

Found in the thread about the Compadre compressor, I couldn't explain it more clearly myself:

rune_lh wrote:

A "feed backward" compressor uses it's own output to drive the gain reduction and results in a smoothed compression sound. Many vintage opto-style compressors work this way, and it's good for vocals etc. The "feed forward" architecture found in most modern compressors offers a better control and accuracy than a feedback architecture, and simply sounds better on drums.

thanxx !
it's very clear

BTW, I should say that Voxengo Voxformer's compressor according to this specification is a feed-backward compressor.

Feedback compressors where very normal in the old days and was mostly used on cheaper equipment. Problem was that the electronics was not to accurate and the feed forward architecture needed allot of manual trimming and better circuitries and was therefore rather expensive. You also used it in limiters because it works very fast. Bad site is that you get ringing in the compression. You can get analog compression chips for very little money now that makes it possible to make a accurate feed forward compressor very cheap so it is the preferred technique now.

Torben

Torben, thanks for telling the history.. What I was using in limiters? Who prefers feed forward technique?

Aleksey, Feed forward compression are the preferred method in modern compressors first of all because you do not have the ringing effect and now with the new log converter IC's and VCA's it is not just pro products that can afford it. I am just talking technique here; people must make up there own mind regarding the sound.

Torben

OK, what is that ringing effect you are talking about? Also, please do not make quick judgement about Elephant limiter as it uses different algorithms.

Aleksey, there will always be ringing when you have feedback as you get complex poles. It will mean that if you have a sudden change in compression the compression will swing until it settles down on the correct value. I would never judge a plugin or hardware based on its technique alone.

I do not hope you got the impression that my post was meant as an attack on your products? I was just answering sinkmusic’s question. I highly respect your products.

Torben

Forever,




Kim.

Torben, OK, then just say that such ringing can be very pleasant at times, depending on the implementation.

But honestly I can't understand what's wrong with that ringing and how I can hear it. Compressor's VCA is not the same as filter as it produces non-linear distortion and colors the sound by its means (i.e. it's not fully a filter to which 'ringing' (resonance?), 'poles', 'zeros' terms are applicable).

So all I asking you for is to share with me your knowledge because I have never heard about such thing as 'ringing sound of the compressor'. Maybe ringing is not the correct word? Right now I understand it as 'pre-ringing' of linear-phase filters, but can't grab your concept applied to a non-lookahead compressor which never knows what it will compress next.

Aleksey, I am just talking technique, I prefer to let each individual user decide what sounds good or pleasant. Taste cannot be discussed or analysed. Lets just take an obvious example, reverb. Technical the best reverb must be the real thing, where the music is played in a room or hall. But in reality many people prefer the sound of high-end artificial reverbs like Lexicon because they are used to that sound from records. As for compressors the technical point of view has always been that feed forward was the best but in may cases it was difficult to make due to high tolerances and maybe drift in the components used so feed back was used as a necessarily trade of. But still, many people might find that feed back compressors gives the effect they want or even sound better than feed forward compressors. Also think about engineers has worked hard to minimize noise, distortion and pumping effect and now some people prefer these things in a compressor. As for the complex poles, you always got them when you have a feedback. In a delay with feedback you got an infinite numbers of complex pole pairs.

Sinkmusic, sorry for stealing your thread with boring technical stuff

Torben

Quote:

sorry for stealing your thread with boring technical stuff

no, please, keep going since you're already at it.

Aleksey Vaneev wrote:

OK, what is that ringing effect you are talking about? Also, please do not make quick judgement about Elephant limiter as it uses different algorithms.

Hi Aleksey,

Interesting discussion - please continue

Just to insert a native english speaker's translation ...

When Torben says "You" I think it is clear he means "A person" or "One" not Aleksey Vaneev. He is not actually commenting on the code in Elephant.

Regards,
Eg

egbert wrote:

Aleksey Vaneev wrote: OK, what is that ringing effect you are talking about? Also, please do not make quick judgement about Elephant limiter as it uses different algorithms. Hi Aleksey, Interesting discussion - please continue Just to insert a native english speaker's translation ... When Torben says "You" I think it is clear he means "A person" or "One" not Aleksey Vaneev. He is not actually commenting on the code in Elephant. Regards, Eg

Yes that's right. OK now I understand why Aleksey seemed to take it personally, I must be careful with the word "you"

Thanks
Torben

Thanks for letting me know.

Torben wrote:

Yes that's right. OK now I understand why Aleksey seemed to take it personally, I must be careful with the word "you"

Actually using "You" like that is perfectly OK colloquial english - it is just that it makes "You" more likely to be misunderstood. The use of the relatively unambiguous word "one" makes "one" sound like a member from the British royal family on a grouse hunt or something - you rarely hear it in colloquial english.

The main thing is that you two - international treasures of DSP coding that you are - don't start a blood feud over nothing

Regards,
Eg

Torben wrote:

Aleksey, there will always be ringing when you have feedback as you get complex poles. It will mean that if you have a sudden change in compression the compression will swing until it settles down on the correct value....

hiya Torben, Aleksey! hope all is well in your respective countries. (btw, could i crash on your couch for the next 4 years? )

this is an interesting little discussion. i knew about the "ringing" effect of a FB topology, but i had never heard it expressed that way. i always thought of it as "compression overshoot". i had never heard it interpreted in the context of filter theory. so... interesting.

FWIW, i've been told that this ringing (or overshoot) is influenced by the sample rate and the attack time constant. with lower Fs and shorter attack times, more ringing. i was messing around with an opto design months ago, and used an Excel spreadsheet to graph the compression over time, so as to find the shortest possible attack time that wouldn't result in ringing. the other solution to this "problem" (although some might not see it as a problem at all) is to use upsampling, which as a side-effect would help to prevent aliasing at high ratios (i.e. limiting). but you guys already knew that...

take care everyone.
== chunk