Stereo Tool

I can't hear difference when 1770 is ON ?

SKIP Select RMS / Reduced bass / ITU-1770? - What that mean ecxaclty?

Difference isn't big for most tracks - but try the 2 tracks you sent me before and then the difference is pretty big. (With ITU-1770 on, the ITU-1770 based volume difference, measured in an external tool, is 0.5 dB, without it it's 2 dB).

Try that "Gimme Gimme Gimme" track, that has very little highs and hence responds much more (relatively) to ITU-1770.

SKIP Select RMS / Reduced bass / ITU-1770? - What that mean ecxaclty?[/quote]
Initially I wanted to allow a selection "RMS", "RMS with ITU-1770 below 200 Hz (reduced bass)", "ITU-1770".

Anyway, I think that RMS isn't really useful, so I removed that one.

What remains is default (ITU-1770 below 200 Hz) or full ITU-1770.


"SKIP" = I'm not going to do it

New BETA uploaded.
- Changed ITU button position
- New settings are saved to disk now
- AGC output level corrected (not lower anymore when RMS smoothing is used)
- Fixed bug in band 3 behavior (other bands as well, actually)
- Changed band 3: Restored original functionality (before BETA2 versions) because they sound better (more natural, less different from original). Default settings are now IDENTICAL to those of band 1!
- Changed default RMS smoothing time from 10 to 20 seconds because I could still hear the volume go up and down. With 20 seconds it's far less noticeable.

I've done a comparison against version 4.22 and 5.00.

Compared to version 5.00, the average AGC output is nearly identical in this BETA version, except that for already compressed songs the volume got lower and for non-compressed songs it got higher. (Which is good). The bass level appears to be 0.4 dB lower than in version 5.00; highs are identical (averaged over a 12 minute body of test sounds).

Compared to version 4.22: If you turn 3 bands on in the BETA version, the frequency content is nearly identical (in version 4.22 the average output level of the AGC is higher). For some tracks, there are bigger differences.

I just received a report from DUB<^>STEP that the new version sounds harsh, when using the Webradio / Hard Rock preset.

Cause is a bug: The AGC doesn't do much in this case, but I'm actually boosting the output level by 40%! Will look into this tomorrow.

(Basically, this bug occurs when the AGC output bars are (nearly) completely filled. For most presets, with loud input levels and much lower AGC settings, this problem does not occur. First step in this case is AGC, which doesn't do much. Second step is 'boost when the level is too low', and that DOES occur and boosts - default setting - by 40%).

A fix for this issue has been uploaded.

NOTE: In cases where the AGC does very little, the output can still be louder than before (but not louder anymore than what Pre Amp specifies). So for certain presets, such as this Hard Rock preset, the AGC target level might need to be lowered slightly.

One thing!

This setting, "Responce time" is to big to be 20sec!

The Other Parametar is Speed of drop.. whis is really slower at 20 sec then in 5sec. 5 Sec should be enough or max 10sec, but speed of drop should be slower.
Currently Speed of drop is good at 20sec, and that speed should be at 10 or even 5 sec.

I think you're right...

I think I also need to lower the 'maximum volume boost' setting as much as possible. 150% (3.5 dB) is quite a lot - but at least for some tracks (like Gimme Gimme Gimme) it's needed.

To lower that setting, increasing the up speed and lowering down speed helps, but it does lead to other effects (for example, lowering 'down speed' makes the start of a new track louder, and the 'remove remaining peaks' slider must be set higher).


Edit: To my really, really, really big surprise:
- For 1 band (which is what I've been testing so far), 150% is needed
- But if I select 2 bands, 120% already seems to be enough! (And even 115% isn't bad)
- That means that, for the same drop speed, 20/50th of 20 seconds = 8 seconds can be selected
Aside from having a faster response time, the maximum boost is now reduced by more than half, which makes the output level far more constant. (Smoothing filter is less noticeable, 'harsh' sounds (DUB<^>STEP) are greatly reduced, etc.)



Still, having a separate drop speed might improve things further. But use these settings instead of the defaults, they really work a lot better.

This is probably caused by the fact that in tracks with a high dynamic range, the bass is very strong - and in the 2nd channel the bass below 200 Hz is completely ignored.

(Note: This does indicate that Bojcha might be right about stronger ignoring the bass in the RMS measurements... Still, that would deviate from the ITU-1770 spec).

I know that because i measured "processor's" AGC all frequencies .. and yes it deviate much from ITU.BS-1770 specs, but in basic that's it. Actually that is configurable there! That's why i ask for Detecting-EQ in AGC, or something.

Maybe you'll solve one song i have witch is so thin (at start) .. but i think that's more because multiband.

slower drop speed for "responce time" 5sec(scan) + 2sec(drop)

That doesn't necessarily mean anything... For example, in Stereo Tool, because I'm using 2 bands, tones below 200 Hz have no effect on tones above 200 Hz. So if you measure the output level for a 50 Hz tone, it might seem that it still counts at -5.3 dB (that's the ITU spec), but for tones above 200 Hz it has no effect.

(In fact, in Stereo Tool, using 2 bands, ANYTHING below 200 Hz is ignored for frequencies above 200 Hz. So what you want is already partially there...)

That cannot be true.. watch how much 1st band drops when bass at Beyonce's song kicks at ~30Hz.
it should drop, but not that much! but again, multiband is droping that bass alot. That's true too.

hmm maybe first two bands in multiband need same thing

don't BAN it !