Stereo Tool
https://forums.stereotool.com/

Better AGC sudden spike handling
https://forums.stereotool.com/viewtopic.php?t=4076		 Page 3 of 6
Author:  Brian [ Wed Feb 01, 2012 6:37 pm ]
Post subject:  Re: Better AGC sudden spike handling
Quote:
note @all: test using slower "Down Speed" (3.000- 5.000), you'll better notice new mechanism, and also should bring better consistency, but just a bit more compress_like_sound.
Try my beta 5 preset with the increased gating as a brake for the AGC. If the levels don't get out of control, the spike reduction only kicks in when it really needs to, at least I think / what I've tested so far...
Author:  Bojcha [ Wed Feb 01, 2012 7:09 pm ]
Post subject:  Re: Better AGC sudden spike handling
with new beta ?
Author:  hvz [ Wed Feb 01, 2012 7:29 pm ]
Post subject:  Re: Better AGC sudden spike handling
New version should be read, but i'm not home right now so i cannot upload it now...
Author:  Brian [ Wed Feb 01, 2012 7:55 pm ]
Post subject:  Re: Better AGC sudden spike handling
Quote:
with new beta ?
Yes. I had the idea before Hans posted this. It works with or without this change. I'm still working on the up speed and push values though.

Edit:

One big misconception about the AGC settings is "But never raise band 2 more than xxx% above band 1 output level"

When I first looked at that, I thought that 100% meant doubling. Everyone else had 250% or 300% or something crazy, but "everyone else was doing it", so I did too. 100% though means "never raise band 2 more than band 1". The 250-300% settings that some people have cause these major spikes. As I mentioned with respect to the spike in "The Pretender", it's not the bass (Band 1), but the attack area of percussion (Band 2, or upper Band 2 / Lower Band 3). When set to 250-300%, combined with a higher up speed and not enough gating, situations like what happens in "The Pretender" (loud to soft to suddenly loud) causes the sudden spike.

I could be wrong (people love to tell me I'm wrong, so...), but it is possible that adjusting settings may reduce the need for the new spike protection, or remove the need completely.
Author:  hvz [ Wed Feb 01, 2012 11:52 pm ]
Post subject:  Re: Better AGC sudden spike handling
I found a bug that I had to fix first. Now building again, will upload the new version in about an hour.
Author:  hvz [ Thu Feb 02, 2012 1:04 am ]
Post subject:  Re: Better AGC sudden spike handling
Bug that Bojcha mentioned is fixed now:

Winamp DSP plugin: http://www.stereotool.com/download/dsp_ ... 22-003.exe
Stand alone version: http://www.stereotool.com/download/ster ... 22-003.exe
VST version: http://www.stereotool.com/download/vst_ ... 22-003.dll
VST version (No SSE2): http://www.stereotool.com/download/vst_ ... 22-003.dll
Command line version: http://www.stereotool.com/download/ster ... 22-003.exe

Also, the BS412 implementation is improved - it now also works if you don't enable FM output. But there is still an issue which sometimes causes the level to drop suddenly (I think why, but I need to investigate it further).
Author:  hvz [ Thu Feb 02, 2012 1:42 pm ]
Post subject:  Re: Better AGC sudden spike handling
@Brian: Those settings might help to reduce spikes, but they'll never completely remove them - even if you use only 1 band you can still have spikes.
Author:  gpagliaroli [ Thu Feb 02, 2012 1:54 pm ]
Post subject:  Re: Better AGC sudden spike handling
Quote:
Hans one indicator of gain reduction in the new Compressor Singleband would help to see when and how does the compressor, to make a finer adjustment. If possible it would be good, later going to try the new version.
It´s possible?
Quote:
Something that occurred to me as an idea and throw it to work:

- From the AGC information is available the change of level, at least for the time of lookahead.

- If you have this information, my idea is that you can handle the threshold and compression ratio Singleband Compressor (not sure if both or one of the two) automatically.

- At times it is not necessary to use the Singleband compressor, automatically adjust the threshold to a high level and compression ratio to 1:1, so that would not affect the signal.

- In the case is detected from the AGC a significant level jump (you could set this), the AGC threshold and automatically adjust compression ratio that makes the peak will not be able to handle the AGC, is controlled or absorbed by the compressor while accommodating AGC gain levels.

- I think the condition of automatic adjustment of the threshold and the compression ratio should be, first lower the threshold to the lowest level to maintain the compression ratio as low as possible. This to make it less noticeable as possible.

- From this last show that you could pass the threshold to a fixed value and only needed to calculate the peak absorb excendente, for simplicity.

I hope you get the idea, and I hear opinions?

:geek:
any opinion?
Author:  hvz [ Thu Feb 02, 2012 4:29 pm ]
Post subject:  Re: Better AGC sudden spike handling
gpagliaroli: What I made is actually close to your idea - which would also have worked. So, yes, it's a very good idea. I don't think there's a need for it anymore though since the AGC is now handling it in a similar manner (basically, I run 2 AGC's, one with a very fast drop speed if the volume jump is much bigger than the AGC can handle. Then, continuously, I choose the best match between those two AGC levels. After a few seconds the two should be identical again, and there's no difference anymore between the old and new behavior.

The separate compressor would have required more configuration values though, which would make it more difficult to configure. The current solution is automatically configured correctly, if the threshold levels are set correctly.
Author:  gpagliaroli [ Thu Feb 02, 2012 4:58 pm ]
Post subject:  Re: Better AGC sudden spike handling
I tried the new beta version and I've noticed that has significantly improved levels abrupt increases, good job! ;)

I suggest a track for testing, "Earth Song" by Michael Jackson (especially in 2:37 minutes, at least the version that I have generated a lot of work to AGC).
Page 3 of 6 All times are UTC+02:00
Powered by phpBB® Forum Software © phpBB Limited
https://www.phpbb.com/