Stereo Tool

I've been playing with some Loudness settings a bit and now I have a more stable sound when loud highs are present - at the cost of loosing some clarity in the highs when they get really loud. I'm still working on it, not sure yet if this is the right direction, but I'll post something here to get feedback...

(The goal of this is: If loud 'S' sounds are present, other sounds shouldn't be pushed down. Or at least not as much as they are now - even though it's not much, the Orban seems to do a better job at it and I can't stand that )

lol...
Yes that's problem with loud "s"..
But if you listen better orban files you can hear that they are much smudged.
But true .. some fast and good HF filter they have

I have just checked Beta 510-19 VST version.

It still crashes (I use the free Acon FX chainer) !
Obvious still something wrong with VST.

Earlier versions (for sure Beta 3) worked without any problems.

By the way talking about de-essing:
My favourite de-essing tool is Spitfish (freeware !):
http://www.digitalfishphones.com/main.p ... &subItem=5

I have made some changes, now getting the same 'S' level with even MORE mids than the Orban - about 3-4 dB more than before! The bass is still dropping a bit more than in the Orban though. The 'S' should sound just as clean as before - but there's far less need to reduce 'Allow louder highs' - you can even set it above 0 now without too many ill effects.

Building now...


(Basically, if there are loud highs present and they are also loud relative to the lows, I boost some of the Loudness settings for frequencies below 6000 Hz, which allows some more Loudness distortion - only at those frequencies - which in turn results in a louder output.

Winamp DSP plugin: http://www.stereotool.com/download/dsp_ ... 510-20.exe
Stand alone version: http://www.stereotool.com/download/ster ... 510-20.exe
VST version: http://www.stereotool.com/download/vst_ ... 510-20.dll
VST version (No SSE2): http://www.stereotool.com/download/vst_ ... 510-20.dll
Command line version: http://www.stereotool.com/download/ster ... 510-20.exe
Linux command line version: http://www.stereotool.com/download/ster ... BETA510-20 (Build error - server was offline)

Note: I haven't yet had time to compare this with version 5.19, it should be 100% identical except when loud highs are present. But I still need to check if I didn't make any mistakes - I did make (and undo) quite a lot of changes during editing.

I'm getting occasional graphics glitches with beta20 that I didn't get with beta15.

Please see attached:

I am getting the same graphics anomalies as superH. Haven't had enough time to evaluate the sound yet.

edit:The missing controls come back if you hover the mouse pointer over them. They seem to disappear if you open any other controls or presets.

Ye .. same here ..

Anyway, Hans try "S" with this track...
http://www47.zippyshare.com/v/58273648/file.html
Problem with this track is that loud "S" drops sound in multuband, and even in AGC ..
I tested with loudnes only and result is good .. but not good if Pre-Emphasise is used over 10us

Another problem is when FM Mode is used. Level Drop (with composite limitter ON or OFF). Once i asked for FM bypass-mode. I know it's not possible but with BBP is, and works perfect! To remind .. That Orban files are from FM de-emphasised.

Does this mean that BETA510-19 was better? I'm trying to improve things here, but 'perfect' isn't possible - there's always some tradeoff between clarity, output volume and constant output level.

About AGC: Are you suggesting that I ignore high frequencies in the AGC? (Or, for example, that I use some ITU-1770 like processing, but opposite, eg. reducing the highs before performing the AGC? As long as there's no side-band equalizer...)

About Multiband: This should only affect the high frequencies, unless you're using the rightmost sliders.

About Loudness not good for pre-emphasis over 10 us: Was it better before (eg. BETA510-19)?

About Bypass mode: You can manually select the filters that you (don't) need to get the same sound. Output should indeed sound nearly identical to the input, especially since the spikes cause by pre-emphasis will be relatively low. I want to add this at some point, but haven't had the time for it yet...



I'll try the track you've sent tonight.

No, That's not smart, That volume drop with very loud highs should be always covered by multiband.
This AGC works better then ever. But, some EQ would be even better (parametric )

I did not try "only" Loudness in b19 .. this in b20 seem has no problem with loud 'S'.
I am just pointing to pre-emphasis, Hard limit and mostly on Composite Limiter.