Stereo Tool

I'm going to try to explain this the best I can: is it possible to make the AGC behave like a track has been Replay Gain scanned?

I mean, instead of trying to maintain a constant audio level during the whole track, take into account the differences in level and only maintain the average.

I know it must sound strange but I'm in a situation where I don't know when a non-scanned track is going to appear and no matter how hard I try, I can always notice the AGC effect. I tried with longer attack and release times but I'm still not there...

Is this something doable or am I asking something that the AGC was not designed to do?

Thanks!

In any real-time processing that is just simply not possible.

_________________
control point
control point2

Well, the ST AGC is especially bad at this because it doesn't use past samples, only current values, and moves too fast for loud input. In short it's made to compensate momentary changes.

With a long term averager one can get similar results to Replay Gain iff the beginning of the track gives a close approximation to the loudness of the complete track. Then one may only hear one correction in the beginning.

However that's obviously not always working.

Which settings to adjust if I don't want songs' quiet / fading in beginnings to be as loud as the other parts of the songs?



top channel: original, extremely brickwalled, but the beginning part is quieter

bottom channel: after stereo tool. much better sound, with wider dynamic range, but: the first part is as loud as the other

You can try AGC ratio and/or stronger gates.
But really, This software is Real-Time processing Tool(!) Meaning that any filter in it does not look back and forward more then few samples/milliseconds. That cannot be clasiffied as "bad".

_________________
control point
control point2

You're right Bojcha, I see that it's just a case of me wanting to use a tool that wasn't designed to do what I want to do. Same as trying to peel a potato with a spoon (lol).

Thank you all for your answers.