Stereo Tool

Will add that. (Already planned to do so)
Ok, now building with no up-speed (just immediate), will see how that behaves.
That's odd - it starts at 4000 Hz so below that nothing should happen??? (Although it is possible that Loudness is unable to 'hide' distortion in existing lows which can lead to some small volume drops).

And that mean what? whren it's down = no UP, or Up Speed 0ms ?

That's "realise" Down speed should be 0ms and up speed fast as much as possible.

Yup, next version has 'as fast as possible' and it does indeed sound better (I had expected the opposite). Building now.

('as fast as possible' here means that I'm splitting the audio into 256 sample (6 ms) chunks, and calculate the required level for each chunk based on its RMS level.)

I have good news for myself

i can make lows from 0-350Hz fixed to around 95%, and to have punch in same time
it goes from 90-96%
Ant the best thing it stays like that with loudness from x1.01 to x4.00 + It goes down with more loudness

Here's beta 60:
- Loudness Highs AGC: Very short up/down speeds (volume changes within 6 ms)
- Loudness: Removed bug that caused mid frequencies to be pushed down by loud 'S' sounds.

Winamp DSP plugin: http://www.stereotool.com/download/dsp_ ... 510-60.exe
Stand alone version: http://www.stereotool.com/download/ster ... 510-60.exe
VST version: http://www.stereotool.com/download/vst_ ... 510-60.dll
VST version (No SSE2): http://www.stereotool.com/download/vst_ ... 510-60.dll
Command line version: http://www.stereotool.com/download/ster ... 510-60.exe
Linux command line version: http://www.stereotool.com/download/ster ... BETA510-60 (not available)
Linux GUI version: http://www.stereotool.com/download/ster ... BETA510-60 (not available)

Correct setting for 'De-esser: Limit highs below' still needs to be determined, I found ~ 71000 (-1.76 dB) to be the maximum value to avoid gaps in the mid/lows in the Eminem song (which appears to be one of the most difficult tracks in this regard). I've left the default at 65536 (-2.46 dB) for now, feedback is - as always - welcome.

Hans pls make variables saved... we have all the time to set the sliders on any new Beta or after restart..

_________________
Sexy Girls in your city

Please tell me:
- WHICH SLIDERS DO YOU CHANGE?

Because I want to clean up all the sliders that are not used/useful, and keep (and save) the others.

Thanks...

on every restart
Igrnore if Below
Clip Lowest bass
Then rise to (strict)
Dirty bass
Accounts for highs
High vs lows
to
Clip Highs feedback
Punch clip lowest
Max Highs burst

_________________
Sexy Girls in your city

Maybe I should add: AND TO WHICH VALUES?

I think for most of these sliders there's probably 1 value that's best for everyone. If that's the case, I can remove the slider (and keep the interface simpler). 'Dirty bass' and 'Punch: Clip lowest bass at' are definitely not one of those. And maybe 'Clip lowest bass at' and 'Then rize to' should also remain there.


If I'm right that for most sliders a single value is best for everyone, then the fact that you need to change the settings means that my defaults are not good (which is very well possible, I focussed on the algorithms behind them, not so much on the settings - not yet anyway).

Chris: Why do you change 'Account for highs'? I was actually thinking about removing that altogether, because my impression was that it only has bad effects on the sound...


I'll make a list of all the sliders here:

Sensitivity boost x1.00 - set higher to respond stronger to predicted distortion
Ignore if below 50% - Instead of mapping measured distortion level 0%..100% at bass reduction 0%..100% we map 50%..100% at 0%..100% (so 50% distortion = 0% bass reduction). Note that Loudness will repair some of the distortion, and that low distortion levels aren't noticeable.
Reduce sensitivity for non-constant sounds x1.00 - Idea is: Voice distortion is most noticeable for constant tones. Higher = less sensitive during changing tones.

Clip lowest bass at (dynamic) 90%
from 0 Hz upto 150 Hz
Then rize to (static) 100%
Strict ENABLED (I will change the default!) - This currently works on the whole low frequency range (upto 400-800 Hz), should I change that? Or make it strict for both ranges?
at 300 Hz

Dirty bass 0% - Higher = more dirty
Account for highs 0% - If more highs are present, bass level is reduced. This might protect against some distorted bass sounds, but also results in bass level drops when loud other (often 'S') sounds are present.

Highs vs lows protection runs from 400 Hz
to 800 Hz - Tones above this are protected against loud tones below it and vice versa

Clip highs feedback strength 100% - How strong do we reduce the bass when voice distortion is predicted?
Maximum feedback 50% - Never reduce the bass more than this

Punch: Clip lowest bass at (dynamic) 95% - During punch, keep the LOWEST bass below this level. This time it's NOT the whole area upto 400-800 Hz, which is odd and probably wrong...

Ignore if below 10%
Clip lowest bass at (dynamic) 60-70%
from 0 Hz upto 150 Hz
Then rize to (static) 90%
Account for highs 10%

my values... i am not sure they are all right..!

_________________
Sexy Girls in your city