Stereo Tool

That's what I meant when I wrote:
"Edit: What would happen if band 1 does not respond to Celine Dion? So the volume below 200 Hz wouldn't drop if she starts to sing? Would that improve things? Just like band 2 is locked to never get lower than band 1, I could add a band where I only process bass and lock band 1 to that..."
(a few posts back).

But right now I think SuperH's solution is better (maybe combined with this at a later point).


By the way, what I called 'a bug in the AGC' isn't really a bug - the 1st processing step outputs already compressed auto at around the specified volume, so the 2nd AGC only responds slightly to it. But, not yet compressed songs still have large peaks, that do cause a big effect in the AGC. Which results in a bigger volume drop, instead of a smaller volume drop that occurs when I increase the input volume of the 2nd AGC.

somehow i miss that. Ok, carry on...

I haven't posted any updates anymore since you asked! Next one will be BETA2 again

OOPS! Michi, I wanted to reply to your post but I accidentally hit 'Edit' instead of 'Reply' - and killed your whole post (except for this single line...)

You were asking for people to post their AGC tweaks to help you with configuring it... Sorry!!!

With slower down speed (around .0001) things are much better .. but for sudden loud burst not good for few seconds.
Here another idea.. for this slow speeds.. to make some hold time for "peek removing" .. i guess hold time must match the down speed. After agc is down "peek removing" is not active or just sometimes when is needed.

Bojcha,
1. Is COMPLETELY solved by the 2nd AGC idea - if I can fix the behavior for low AGC input levels.
2. That's true - currently it's adapted dynamically, but that works best for settings around .0003. Anyway, with the 2 AGC's solution, it doesn't really matter anymore...

Waitinig to hear that ..

@ Hans

No problem at all.
Where is big brother google and its cache when you need it ?!

Next time: Just write "Shut up !"
Remember Hans it was your idea (not my) to put your photograph into Stereo Tool's about window.
Though months ago I suggested you to spend some bytes more for this photo to have a higher quality, the quality is good enough to identify a person.
Hans I know your face !
Keep your eyes open for black limousines that stop infront of your home and men with sunglasses and hats !
"Hans, you always were like a son of mine. Get into the car. Let's drive and talk a little bit."

But Hans today is your lucky day.
I had some (not all) fragments of my original post in a text file.
Besides the naming issue it had other information and questions.
I will rebuild these fragments in another additional post later !
There are days you build more than two BETAs !
That's why I asked for explicit BETA naming (using numbers from 001 to 999).
Something like:

dsp_stereo_tool_installer_V5_BETA_xyz.exe

or (is the word installer not obsolete, because every child knows what an Exe does ?)

dsp_stereotool_V5_BETA_xyz.exe

or (starting with "stereotool" would keep all different stereotool versions together in every download folder !)

Stereotool_DSP_V5_BETA_xyz.exe

Stereotool_Standalone_V5_BETA_xyz.exe

Stereotool_CLI_V5_BETA_xyz.exe

Stereotool_VST_V5_BETA_xyz.dll

Stereotool_VST_noSSE2_V5_BETA_xyz.dll

Especially for the VST dll version "Stereotool" should be at the beginning of the name (word).
No doubt everyone who already uses the VST version renames the dll, because
vst_stereo_tool_BETA.dll is an absolute no-go !
A VST dll should start with its name !

In the end it is your decision (you are the mastermind).
I just think it would help all users to know exactly which BETA builds they actually use.
And this would help you Hans, because it would open the door for more user feedback.
The danger that you will drown in too much feedback does not exist.

And indeed:
It would be great if someone could post (or upload) his special AGC settings in a .sts preset.
In combination with the suggested explicit BETA naming it would help to be sure that people are not talking about apples and oranges.

Will take some time...

I'm still configuring things. Unfortunately, after fixing the issue that the RMS volume is lower when loud spikes are present when the (average) input level is close to the configured output level, the 'loud start' problem at the start of a new track has returned... On the other hand, it also improved the Celine Dion song further.

An advantage is that the 1st AGC can be configured MUCH more relaxed, with a much lower Up Speed as well. THAT does greatly improve the 'loud start of tracks' issue - and it can be combined with gating to improve things further.

It'll always be problem with down speed ... that's why i say, some compressing (or call it whatever you want) to fast push loud bursts till the agc comes down. What i read so far that's only way for this problem!

We already know that slow AGC is always better solution for processing, but that sudden audio bursts searches solution.

That's exactly what's happening with these 2 steps