13 Oct
2017
13 Oct
'17
11:19 p.m.
On Mon, 2 Oct 2017, Carsten Kurz via DCPomatic wrote:
2.11.24 looks good with Dolby TrueHD streams, also
7.1. Although it
starts recalculating the audio when I switch from 5.1 to 7.1 (12ch) DCP
to see the BSL/BSR channels. Maybe it's better to include all available
audio channels in the input file for analysis right from the start? Or
maybe DOM does that already, but is false triggered towards a new
analysis if the DCP audio arrangement is changed? No big deal for short
pieces, but for full length features, it is annoying if another 20min
analysis run is started, especially since R128 proves to be very useful
now and I would advise everyone to do perform the full audio analysis.
I think this is a little tricky; the audio analysis shows how the audio
will look in the DCP, not just how it is in the content. Part of this is
the audio mapping between content and DCP which changes when you go from
5.1 to 7.1.
I'm sure there are ways to optimise the 5.1 -> 7.1 case but they might be
rather fiddly.
It also appears as if calculating audio is now a lot
faster than before
(2.11.22) - at least for the 7.1 THD full length feature, it now only
takes around 13min, compared to around 45min for the 5.1 ac3 version
yesterday. Yes, the 5.1 ac3 analysis also computes much faster now with
.24 - a mere 6min!
Good...
I notice that the more audio streams in the input
file, the longer it
takes - probably approx. the same time for each stream, or every single
audio channel adds up. So it is certainly wise to only include the
necessary streams in the MKV, or to extract unnecessary streams before
with an MKV splitter.
OK. I haven't benchmarked this stuff lately so I'm not sure on this
either way.
While I understand it's busy at DOM-HQ, may I hint
again at:
http://dcpomatic.com/mantis/view.php?id=835
because it is probably very easy to implement coloring the graph legend ;-)
Done ;)
Cheers,
Carl