Video formats for 4K DCPs

[maurotto]

what is the better setting of the video input for DCP 4k?
is better one video Quick time not compress o jp2000 ?
what format do you seggest?
thank

[carl]

I would say the best input is the one with the highest resolution and least compression; I'm not sure if the actual format will make much difference. What are your options for the input file?

[maurotto]

I did some tests to see how the video format dcp o matic converts to 4k at higher speeds.
With a H264 video H@5.1 to 240 Mbit converts to 0.9 to 1 fps (4K)
With a video JP2000 conversion rises to 1.2 fps
If the video is made ​​up of many static images, such as a slide show of photographs, the output format of Premiere, Quick time set in Photo Jpeg is very fast, about 14 fps - 20 fps (4K) depends on the number of transitions of the photos.
So popular video formats to Fps variable are well managed by DCP o matic.

[maurotto]

I forgot to point out that the Quick time uncompressed format is handled well by DCP o Matic, but it is too heavy, about 440 Mbytes / sec.
That's why I decided not to use it.

[maurotto]

there is a problem that continues to occur. Does not convert the entire file, but often ends before the conversion. Looking at the XML file is read duration values ​​different between audio and video. The problem seems to me to present with video formats in which DPC o matic can not analyze the sound.

[carl]

When you create a film with one of these problematic files, DCP-o-matic should create a file called "ffprobe.log" next to the DCP. Can you send that to me (carl@dcpomatic.com) and I will take a look.

[maurotto]

I did several tests with different formats and these are the results.
DCP o matic converts video formats that have a structure variable-fps without error ,but misinterpreted the duration of the video, and then the conversion ends sooner than expected.
There is no problem with video formats with fps constant
I suggest to check the possibility of using variable fps, because I've seen the DCP o matic convert so much faster, despite the incorrect result.
If you require clarification are available.

Again congratulations.

Mauro