Stereo Tool

I want to share this paper, I hope can help ...
http://www.elec.qmul.ac.uk/people/josh/ ... ES2012.pdf

Edit:
Add 2 links more
http://www.eecs.qmul.ac.uk/~dimitrios/A ... design.pdf
http://www.elec.qmul.ac.uk/digitalmusic ... index.html

_________________
by GAP
"Less is More" (Bob Katz)

Knee is working now (next beta). You have the 'Dynamic release to 0 dB' set to 0? If it's set to 1 you'll often see an abrupt end in rise.

With that to 1 release is always flat.
Also thresholds under -8 (-12, -15..) gives flat reelase, no metter what.
But we'll see with next beta then.

Didn't have time to build everything but here's the stand alone version with proper knee (and also far more efficient).
http://www.stereotool.com/download/ster ... 04-016.exe

Will look into release behavior now. If it's flat it will still be flat with this knee, because the knee doesn't affect the release.

Bojcha, can you tell me which settings you are using? With 'Dynamic release to 0 dB' set to 0 I'm not getting a flat release. Tried with thresholds between -40 and -6, with and without knee, different ratios, feedforward and feedback. It all looks smooth. Note: I'm using "Smart RMS" detection. With peak you *do* get a nearly straight line. If you want somewhat peak-like behavior without that try the new 'Smart RRMSS' mode.

Default settings in beta16 - http://i.imgur.com/T3IBha3.png
More threshold (~-18) - http://i.imgur.com/Lv2VXSc.png

testfile 0dB sine, then -8dB.

But default settings is with the Dynamic release to 0 dB set to 1!

Yes, and also to 0 is just bit fixed and with lower threshold again more flat.
And btw, knee is still not working.

@Bojcha:
Knee:
- Input 1000 Hz tone, rising level.
- Loaded defaults
- Threshold to -16 dB
- Ratio to 1000:1
- Feedforward (so feedback to 0%)
- Top: Knee 0 dB. Bottom: Knee 10 dB. Output is normalized.



Looks fine to me?

0 dB & -8 dB, default setting, set "Dynamically release to 0 dB" to 0, threshold to -18 dB:



Also looks good to me?

I've also tested -54 dB and it looks nearly identical.