PlanetZ forums - scopeusers.com

okay most recent compressor design are based on the classic layout, with treshold, attack and release, this seems to be the ordinary and classic way to achieve more loudness from a fixed headroom ...

Yesterday I though about why there are no new approaches to reach that goal...
Especially with digital information....

All compressors I know so far work like shown in the following picture:



But in general I think it has 3 main bad components:

(1) pumping at heavy compression setting and lower treshold levels....
(2) signals below the treshold get the strongest "relativ" boost
(3) headroom loss due to heavy compression and normalising process in digital world leads to quality loss (okay not big)

after normalsing the compressed signal should have a curve like this:



right?


I thought some time about this and came to the conclusion that it should be possible to do a more satisfieing compression algorithm design for an even smoother sound result:



I think this would have some main advantages:

(1) smoother compression over the whole dynamic range, not only in the upper range
(2) no more need for attack and release parameters, as there is no hard cutting treshold line
(3) no headroom loss if correct programmed

I`m not sure, this may be compete bullshit, but im interested in some other opinions...

That looks like you´ve described the knee caracteristic, to make a compression sounds more musically. Soft Knee Hard Knee.

The soft knee puts ,allow me to say, a slope on the compression curve, like a knee looks.

There is a work around to achieve a more musically compression.

For example, if you want to get a compression of 8:1, try to use more than one compressor.
You can use three, each with a 2:1 compression.
This work around is called multi knee compression.

Test it!

just googled and that soft-knee looks like what I`m looking for...

thanks Chris...

so you would take 3 compressors with 2:1ratio to approximate a smooth slope by three linear lines when the tresholds are set at different levels....?

sounds like a good workaround...

by the way, we live in the same area, I`m also in Bavaria....
hehe cheers

FMR Really Nice Compressor does this kind of multiple cascade compression.
It sounds really good!!!

erminardi wrote:FMR Really Nice Compressor does this kind of multiple cascade compression.
It sounds really good!!!

ahhh I understand...

the really nice is really nice...

I am pretty sure this method is used in most compressors that have any "character" to their sound. The knee can have many different shapes.

what about this kind of dynamic processor device?
does it exist?
what is its name??

and, most important of all, would it be useful??

Looks like some compressor that turns into an expander?

I guess it could have some use for audio tweakers.

this wouldn`t work because you can`t transfer levels at 0 to higher ratios... your red line always ahs to end at 0.

hifiboom wrote:this wouldn`t work because you can`t transfer levels at 0 to higher ratios... your red line always ahs to end at 0.

then something like this... that goes to 0 throgh a smooth curve
(sorry for the drawing, was freehand!)

bosone, I think thats what the SPL Transient Designer is doing when you raise the sustain....

bosone wrote:what about this kind of dynamic processor device?
does it exist?
what is its name??

and, most important of all, would it be useful??

You could get something like this, if you compressed the whole thing at ie 2:1 with the lowest threshold possible. That gives you a more horizontal line like the first part of the red line. Then to make the last part come back to normal, you would have to use an expander with the equivalent ratio and the threshold set to the point where you want to line to go up again. Setting attack and release times may be challenging though, and I wouldn't the upper "normal" part to stay totally normal.

I guess it could be usefull, if you want to keep the "somewhat natural" dynamics of a track in the louder range of a track - and at the same time have the soft parts more pressent in the mix.

Another aproach wich will probably keep the dynamics more natural in the louder part would be to use parallel processing. Double the track. Channel one goes straight to the mixer with all the natural dynamics. Channel two goes thru a compressor with a low/medium ratio and very low threshold. You may want to put a slight delay in front of the uncompressed channel to keep the two channels in sync. However, there might still be some phase issues. I am not sure about this. But try it out and see, if the result is of any use/fun.