Am 24.02.2016 um 02:13 schrieb Carl Hetherington via DCPomatic:
This analysis (using FFmpeg's filter) is already
in the testing version of
DoM, although currently only on some platforms as it requires a small
patch to FFmpeg. Remind me: what OS are you running DoM on?
Carl, I assume this filter creates summed output of all channels in order to create a
somehow realistic rendering of auditorium loudness according to EBU broadcast-specs. I
have watched the numbers while playing around with audio levels a bit, but currently I am
not sure how to use it.
I recently assisted a short film project and it occured to me that common information on
the net is somewhat shorthanded with respect to audio levels - most amateurs or semi-pros
ask 'Which level should I aim for in cinema, give me a dB number'. Now, on forums,
very often they get the answer:
'Must be 82dB LEQ(M) for advertising, and 85dB LEQ(M) for features'
But that is wrong - these are the maximum allowed levels. It doesn't mean that once
you have leveled your audio track as such, it has the proper listening volume. These
numbers are there to avoid customer complaints and warrant adherence to local standards
like SAWA. They are maximum numbers that shouldn't be exceeded.
An intimate dialog scene leveled to 85dB LEQ(M) would be much to loud.
There IS NO number that expresses the 'proper' volume for your specific feature,
short, or ad. The only way to approach the proper audio levels in a cinema is to listen to
your soundtrack in a controlled listening environment that has ben calibrated using common
pink-noise techniques like 85dB SPL(C) for screen channels and 82dB SPL(C) for surround
channels. Even that leaves some effects out, but it is the only way to come close to an
international standard for cinema audio levels.
It's basically the same as with brightness - there is no standard on how white a snow
scene has to be or how orange a sunset. It is your artistic decision. Same with audio
levels. The only way to get this right is to use a solid calibrated viewing and listening
environment and apply your own choice.
- Carsten