LEQ(m) boost AFTER DCP creation

[Kaufman]

Hi all,

I hope you can help. I have a source wav sound file with the following spec when analyzed. I'm happy with the LEQ(m) being at around 83db

Before.jpg

But the problem is, that once I create the DCP and bring it in to review it... the LEQ(m) is now up to above 96dB!

After.jpg

It's the only value that's changing... and kinda ruining things.

Does anyone have any ideas?

Many thanks!
Kaufman

[Carsten]

Must be some bug in LEQ calculation. Maybe you have a different number of DCP audio channels than source audio channels, and that tricks LEQ(m) calculation towards a bad result. Can you show us a screenshot of your audio matrix?


I think your DCP is as you intend to - as peak and LUFS remain the same between source and DCP, and the graphs are identical, the DCP LEQ(m) must be wrong.

BUT - 83.56 LEQ(m) - if correct, and -14.46 LUFS is VERY loud. Is this meant to be so loud? At reference playback level, this will raise complaints, I am sure, at least within a normal cinema crowd. If I'd have these LUFS or LEQ(m) values, I would reduce gain by at least 3dB - create the DCP, and do a test run at our cinema. At about 15min, I assume this is a short?

[Kaufman]

Many thanks Carston!!

Yes, this is a short.

Yeah, my sound mixer was taking his cue from 82-85dB for commercials and a DCP discussion with these guys at -- https://www.thearchetypecompany.com/dcp -- who wanted it at 85dB

Screenshot 2023-11-06 at 11.41.45.png

Good to know from your experience this is too high though. Is my mixer's intel wrong? I can look into reducing in DCP-o-matic.

I'm not going to get a chance to test at a cinema in the short term, although I can start looking into it. Mainly, this will tour festivals in a few countries and I'm at the mercy of each cinema/projectionist.

A previous DCP has gone out at these levels, and I'm travelling to see it this week. Will the projectionist, modify volumes across the shorts programme do you think? I guess I can ask to do a sound check ahead of the screening?

re LEQ(m) boost: here are the matrixes.

Matrix from files.png

Matrix from resulting DCP.png

Thanks very much for your help!
K

[Carsten]

>Ref level = -20dBFS | Output level = 85dBc


That doesn't mean 85dB LEQ(m)!

LEQ(m) is loudness, and it 'means' that 85dB LEQ(m) is the max tolerable level, which practically means, it is a COMPLAINT level. This is the max allowed LEQ(m) value for trailers, which are usually mixed at insane levels, but are usually only short pieces, 1-3min. Many cinemas play preshow/trailers already at reduced playback level to reduce complaints. If your's is a short and meant to be played back at normal feature level (may be all over the place at festivals), I would suggest to adjust the level of your short so that a LUFS value around -17 to -20 is hit. At a festival, I guess there is always someone with their hand on the volume knob, but you can still make his life easier. Also, sometimes when the first seconds of a piece come out too loud, volume is decreased too much in an overreaction, and your audio will suffer.

A max allowed loudness of 85dB LEQ(m) is NOT a recommended or 'target' level. It is a max tolerable level. That's a difference. Now, for ads - 82dB LEQ(m), and trailers - 85 dB LEQ(m), these MAY be desirable loudness values for mastering companies to deliver what their client wants in order for their ad or trailer to not appear weak during preshow. But that is not a requirement for your short feature.

Your matrix screenshots clearly depict the LEQ(m) bug - you probably created a SMPTE DCP and followed DCP-o-matics advice (or were forced by it) to create a 16ch audio file. The LEQ(m) value of 83.56dB is true for your 6ch input file, but it miscalculates for the DCP where only 6 out of 16ch are populated with audio. It's clearly a bug. The 83.56 LEQ(m) value also matches a -14 to -15 LUFS value.

[DamienB]

Hi,

I have a similar thing.

My DCP was rejected because it's too loud.
They told me I'm about 90Leq(m)

The requests are
85Leq(m)
-3db True peak

The graph says I'm good.

It's the created DCP reloaded in O-Matic,
It's an upmix, stereo to 5.1 done with Audible, brought to Pr exported as multi session PRO RES HQ
imported to DCP-O-Mati. All Channels are well recognized and in the correct order in DCP-O-Matic

Is the messure of O-matic not in same time range than the one of the cinema?

Any Ideas?

[Kaufman]

>Ref level = -20dBFS | Output level = 85dBc


That doesn't mean 85dB LEQ(m)!

LEQ(m) is loudness, and it 'means' that 85dB LEQ(m) is the max tolerable level, which practically means, it is a COMPLAINT level. This is the max allowed LEQ(m) value for trailers, which are usually mixed at insane levels, but are usually only short pieces, 1-3min. Many cinemas play preshow/trailers already at reduced playback level to reduce complaints. If your's is a short and meant to be played back at normal feature level (may be all over the place at festivals), I would suggest to adjust the level of your short so that a LUFS value around -17 to -20 is hit. At a festival, I guess there is always someone with their hand on the volume knob, but you can still make his life easier. Also, sometimes when the first seconds of a piece come out too loud, volume is decreased too much in an overreaction, and your audio will suffer.

A max allowed loudness of 85dB LEQ(m) is NOT a recommended or 'target' level. It is a max tolerable level. That's a difference. Now, for ads - 82dB LEQ(m), and trailers - 85 dB LEQ(m), these MAY be desirable loudness values for mastering companies to deliver what their client wants in order for their ad or trailer to not appear weak during preshow. But that is not a requirement for your short feature.

Your matrix screenshots clearly depict the LEQ(m) bug - you probably created a SMPTE DCP and followed DCP-o-matics advice (or were forced by it) to create a 16ch audio file. The LEQ(m) value of 83.56dB is true for your 6ch input file, but it miscalculates for the DCP where only 6 out of 16ch are populated with audio. It's clearly a bug. The 83.56 LEQ(m) value also matches a -14 to -15 LUFS value.

Thanks so much Carsten. Your answer is so helpful in every way. I appreciate your time / expertise! K

[Carsten]

Hi,

I have a similar thing.

My DCP was rejected because it's too loud.
They told me I'm about 90Leq(m)

They are wrong. They use the wrong tool or use it with bad parameters. I'm pretty sure your audio is okay.

Are you able to check the channel count of your DCP audio mxf? E.g. by loading it into DCP-o-matic again?

Gruß - Carsten

[carl]

I'd be interested to know if you get more consistent LEQ(m) results with this build.

[Carsten]

(edited)

I can reproduce the issue in plain 2.16.66. Took me a while to get to it.

Will try the test build.

[Carsten]

Looks as if there's a small glitch in 2647 - I can only perform audio analysis when input and DCP channels are both set to 16ch. Or possibly just the same number of channels. If I do set it like this - the LEQ calculation error seems to be gone