Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Matt]

Media Center 16.0.93 adds the ability to normalize videos during playback.  We discussed the idea initially here: http://yabb.jriver.com/interact/index.php?topic=64157.0.

The option, which is on by default, is available here:
Options > Video > Audio for Video, DVD, & Television > Normalize volume.

There were a few design goals to the option:

• Work in one pass so that it works with live sources like television.
• Leave dynamic range intact, so that it's not a compromise when used with a high-end system.  Any change should be a simple volume adjustment (i.e. a linear gain).
• If the volume is turned down, the volume should not restore until a new file is played so the system is rarely audible.  The only possible audible change should be a linear (and permanent) lowering of the volume on the first loud noise.
• The system should react instantly so that it's safe to use with home videos or other clips that use 100% of the volume range.
• Any volume gain should ramp up over the first few seconds of the video so that clips that use 100% of the volume range right away don't get altered.
• Adjustments should happen as late as possible in the DSP chain so that any headroom gained from using Room Correction, etc. can be utilized.

From my testing, the new option meets these goals reasonably well.

Enjoy.

[jmone]

FYI - When I started Red October testing I turned this off on the first track as I was getting an odd warbling (sound pulsing in and out).

[Hendrik]

How would this behave with a scene setup like this:

Quiet Dialog, maybe going to 50% of the volume range, then a scene with massive action, explosions, all ranging up to nearly 100% of volume, then going back to the quiet dialog.

Most volume normalizer i have tested will most likely increase the volume of the first dialog, then turn off the correction for the action scene, and keep the second dialog quiet (or in a even worse case, slowly increase its volume again) - causing a massive difference to the first dialog.

I don't think this can be properly dealt with in a 1-pass approach, just wondering how your would deal with it.

[jmone]

I'm a bit skeptical on this one.  It sounds like you want to compensate with a particular issue in the FFDSHOW's a52 AC3 decoder implementation.  I too hear a large sound pressure difference between a52 and other decoders such as avcodec and LAVAudio (but I don't get any honking issue on seeks).  Do you also get any issues with LAVAudio's implementation with seeking?  If not you are better to dump the poor decoder for a "better" one IMHO.

[jmone]

Also - tried "Normalise" with DTV Playback and got a bad "pulsing" of the audio levels on both AC3 and MP2 channels.

Note: I get consistent sound pressure levels switching between the AC3 and MP2 channels with LAVAudio and no "honk"

[Matt]

How would this behave with a scene setup like this:

Quiet Dialog, maybe going to 50% of the volume range, then a scene with massive action, explosions, all ranging up to nearly 100% of volume, then going back to the quiet dialog.

The massive action will turn down the volume for the rest of playback of that file.  The majority of movies show their cards with regards to the volume levels they use reasonably early, but there will be exceptions.  In these cases, the one-time volume adjustment is the trade-off.  Many videos, and especially HDTV, will benefit from a volume boost and never be loud enough to turn the boost down.

Most volume normalizer i have tested will most likely increase the volume of the first dialog, then turn off the correction for the action scene, and keep the second dialog quiet (or in a even worse case, slowly increase its volume again) - causing a massive difference to the first dialog.

If the volume gets reduced (or more accurately the gain reduced), it will stay reduced for the rest of the video.

I do not want this to be a dynamic range compression feature, so growing the gain again is not desirable.  The feature is instead designed to claim the huge amount of unused headroom in the signal of many videos.

I don't think this can be properly dealt with in a 1-pass approach, just wondering how your would deal with it.

2-pass removes the compromise of a possible volume decrease, but it won't work with live sources.

I considered databasing the peak level encountered so that it could be used on the next playback.  However, that gets complicated by this design goal:
Adjustments should happen as late as possible in the DSP chain so that any headroom gained from using Room Correction, etc. can be utilized.

In other words, changing other DSP Settings like the level of a particular speaker will change the signal the normalizer sees since it's late in the chain.

[Matt]

It sounds like you want to compensate with a particular issue in the FFDSHOW's a52 AC3 decoder implementation.

Regardless of the decoder, I've yet to find an HDTV program that goes over about 15% on the level.  In real-world terms, the volume in my shed needs to be about triple watching HDTV compared to listening to music.

[jmone]

Here are two typical graphs for AC3 and MP2 HDTV playback for me.  I also run through a bunch of other clips and they all look around the same (note: Tested on my Main PC).

Are you sure it is not something else in your shed doing this?

[Hendrik]

Peaks of 20% and 34% isn't really using the full dynamic range. :p
But yeah, they usually over-compress broadcast formats, which causes this.

[jmone]

Interesting - I had a look at my music and interestingly they all had a replay gain of between -18 and -6db applied to them so that when Volume Leveling is ON in the DSP (Trackbased, Fixed) my music level is dropped to the same apparent sound pressure levels as my Video Playback (with peak leaves around 20%).  When I turn Volume Leveling OFF or the Adjustment "Automatic based on current playback then peak levels jump to 100% and 90% respectively and sound pressure level is correspondingly higher.  I found similar results on both my MainPC (mixed to 2CH) and the HTPC (mixed to 5.1).

I'd say that this is an area that I now have more interest in as currently and I now can hear a sound pressure difference between music and video sources pending the DSP Volume Leveling settings.  So in my musings on this topic:
1) Most Important:  I want the same apparent "Volume Level" across all my playback (Music, Video, TV)
2) Next Most Important:  DSP Settings that use the dynamic range available

At present with DSP Vol Leveling ON / Fixed, I have the same apparent "Volume Levels" but nothing much peaks above 40% so it is way to "quite" (and I thought it was just crappy little speakers on my Main PC).

Matt - you have me convinced that something needs to be done but please make it a harmonious system approach for all sources!

Thanks
Nathan

[Matt]

You might try the video normalization in 16.0.94 and see how it works for you.

Like nevcairiel said, any one-pass normalization will involve some compromises.  But I believe the good outweighs the bad in the current implementation.  I need to live with it for a while in the real-world now.

[Matt]

When I turn Volume Leveling OFF or the Adjustment "Automatic based on current playback then peak levels jump to 100% and 90% respectively and sound pressure level is correspondingly higher.  I found similar results on both my MainPC (mixed to 2CH) and the HTPC (mixed to 5.1).

The 'Automatic' replay gain feature was invented for exactly that reason.  There's no reason to turn every song in a playlist down a lot if you're just listening to that playlist.  Why turn every Metallica song down 10 dB if you're only listening to Metallica?

With automatic enabled, the volume within a playlist will be the same.  Between playlists the volume can change.

You get the maximum volume that respects the replay gain value and avoids any clipping (as specified by the peak level).

The feature is hard to explain, but really good in my opinion.

[jmone]

Had a play with the following settings (let me know if I've got them set correctly):
* Tools --> Options --> Video --> Normalize CHECKED (note: no more pulsing with Video Clock also checked)
* Tools --> Options --> Video --> Playback Device = Same as audio
* DSP Settings --> Volume Leveling --> Mode = Track Based, Adjustment = Current Playlist, Overflow Handling = Clip Protection

Results:
- Sound pressures appear the same across most content (Audio and Video) with the exception of TV
- Good use of the full dynamic range with eg action movie went from <10% in quiet scenes to >95% for short action burst
- Sound pressures on DVB-T (both AC3 and MP2) are significantly lower with Peak Levels under 30%
- Some Video Files were hitting 100% (mostly Home Video and some Music Video Content)

[glynor]

Pet peeve.  Ignore me. 

went from <10% in quite scenes to >95% for short action burst

http://www.dailywritingtips.com/quiet-or-quite/

[jmone]

Ignore me. 

OK then!

[glynor]

OK then!

Ha ha ha!  Sorry.  I forgot I did that.

That particular day I encountered that same mistake what seemed like 15 times (but was probably only more like three).  The latent grammar nerd in me couldn't take it and had to let it out, and figured I could get away with it on you without starting some kind of idiotic troll-war.

What's funny is that I'm often incredibly careless when I post, which leads me to obsessively double check them for messing up things like it's and its and stupid stuff like that.  But I never actually use the preview function because I never really intend to self-edit.  I always do, and then while I'm there I always tweak this thing or that thing and fiddle, and complain to myself that I should have used the preview button, but now it was too late.  It would be so much better if I could just let it go, and leave "quite" as "quiet" because then I wouldn't add sentences (like I just did this one).

And so, as always, I'm just going to go ahead and click...

[jmone]

Mate - no probs at all from my perspective (you should see how bad my posts are prior to hitting the "Spell Check" button on the posts!  ).  I can do with all the help I can get on spelling and grammar or is it grammer....I think grammar (spell check said grammar so it must be right...or write maybe....anyway glad I'm in IT not a teacher!)

[MrC]

Neither can my quiet Gramma quite write right either.

[JustinChase]

My personal favorite that no one can seem to get right...

Their taking there stuff over they're.

Wait, that's not right 

sorry, yes, volume normalization is appreciated, good work.

[audunth]

Is the Normalize volume setting supposed to be global (not zone specific)?