Did something change with audio analysis?

Anything and everything to do with DCP-o-matic.
dillonjamescole
Posts: 7
Joined: Fri Nov 29, 2019 11:07 pm

Did something change with audio analysis?

Post by dillonjamescole »

I have used DCP-o-matic to convert dozens of .mkv files ripped from Blu-rays over the years. Anytime in the past I've analyzed the audio from one of these files, I never had a the True Peak line be red to warn me the audio was too loud.

But, today I'm trying to demonstrate to a local filmmaker where his audio levels should be in a film he's working on. The film is in 5.1. I'm using DCP-o-matic to analyze the audio from several different .mkv files ripped from Blu-rays to give him a visual example of what professionally mastered levels look like before his next test screening and they're all showing the true peak as being a lot higher than I thought they would be - true peaks louder than -3dB. I've looked at the audio from Blu-ray rips of The Lighthouse, Coco, Groundhog Day, Labyrinth, and a couple others to make sure I had good variation in genre. But after analysis, the true peak is always louder than -3dB. I don't remember it being this way using previous versions of DCP-o-matic.

Does that seem right? Maybe I'm just misremembering, but I could've sworn that most of the center track would hang out around -10dB for most of the feature.

I'm using program version 2.14.23. 2018 15" MacBook Pro - 2.4GHz i7 6-core processor.

Thanks for any help!
Carsten
Posts: 2804
Joined: Tue Apr 15, 2014 9:11 pm
Location: Germany

Re: Did something change with audio analysis?

Post by Carsten »

Hmm. What would be wrong with TP hitting -3dB (or even 0dB)? I understand you haven't seen this before with your selection of content, but there is no legal requirement for leveling this kind of material. I have seen TP slightly above the 0dB line, and even that is perfectly legal.
Keep in mind, neither sample peak nor true peak is a an actual indicator for program loudness. RMS or LUFS is a much better indicator.
The red color in DCP-o-matic does not indicate an actual problem. Only clipped signals are an issue. The red color only indicates that one should probably have a look before you start conversion.

That said, I don't know if Carl changed something in the code that would result in different TP levels between earlier and current versions. I can easily check it, though, as I have many different previous versions installed on my machine. I do remember though that Carl and I negotiated about the threshold levels where the colors yellow and red would appear, but that would not have changed anything about the actual numbers.
https://dcpomatic.com/mantis/view.php?id=1054&history=1


This is the analysis of the Bluray 'Cheyenne - This Must Be The Place', which is not actually known as a 'loud' movie.
As you can see, the overall LUFS value of -28dB indicates rather low loudness. Yet, TP is at -1.65dB. DCP-o-matic 2.14.25


Bildschirmfoto 2020-02-07 um 01.27.54.png
This second screenshot is from a dialog-driven arthouse movie, again, Bluray. Note TP hitting -0.04dB, close to max.

Bildschirmfoto 2020-02-07 um 01.39.48.png
I tested both files with DCP-o-matic 2.10.5 and 2.14.25, and the values are identical.

- Carsten
You do not have the required permissions to view the files attached to this post.
dillonjamescole
Posts: 7
Joined: Fri Nov 29, 2019 11:07 pm

Re: Did something change with audio analysis?

Post by dillonjamescole »

Carsten,

Your reply has provided excellent insight. Thank you very much!

Best,
Dillon