02:31 | anuejn | okay so here is a little progress update from my side
| |
02:31 | anuejn | the axiom-recorder now has a rewritten DNG writer that is based on dng-rs that was created specifically for this reason
| |
02:32 | anuejn | s/reason/purpose/
| |
02:32 | anuejn | we can now specify dcp profiles as yaml files (default used is https://github.com/apertus-open-source-cinema/axiom-recorder/blob/ffd0efaaf4da6e810a694772436b2ff8fd06f2d6/src/nodes_io/default_dcp.yml)
| |
02:33 | anuejn | this allows us to experiment with the color science stuff more easily and write 32bit float dngs properly
| |
02:34 | anuejn | in addition to this, we now also have a CinemaDngReader node that can read the CinemaDngFiles the writer has written
| |
02:35 | anuejn | this allows us to replace raw12 for some applications with directly writing CinemaDng which is quite convenient as it can carry metadata such as width and height with it and can be properly openend in davinci
| |
02:35 | anuejn | it can not yet read arbitrary dng files (because it does not support compression, multiple strips or tiling) but it is probably good enough for now and should be decently fast
| |
02:36 | anuejn | (not much difference compared to using raw12)
| |
02:36 | anuejn | please try it out and report any bugs :D
| |
07:08 | balrog_ | joined the channel | |
07:09 | se6astian_ | joined the channel | |
07:14 | balrog | left the channel | |
07:14 | se6astian | left the channel | |
11:18 | se6astian_ | cool anuejn, very good!
| |
11:19 | se6astian_ | could you add a few more examples to the github readme like an up to date record command to the new format
| |
11:19 | se6astian_ | and maybe playback of recorded clip
| |
11:19 | se6astian_ | anything else that would be commonly used?
| |
11:41 | se6astian_ | writing cinemadng directly how does this affect fpn correction: is it done before writing or can the fpn compensation data be embedded in the cinemadng file?
| |
12:22 | aombk | joined the channel | |
12:22 | aombk2 | left the channel | |
12:31 | anuejn | se6astian_: afaik there is no way of how to embed the fpn compensation data into the cinemaDNG
| |
12:31 | anuejn | you could do it before (depends on your pipeline) or not
| |
12:31 | se6astian_ | right
| |
12:32 | anuejn | there is no real difference from cinemaDNG and raw12 apart from the one having a standardized header
| |
12:32 | se6astian_ | is it MXF or DNG ?
| |
12:32 | anuejn | what is MXF?
| |
12:32 | se6astian_ | one frame per file or container with multiple frames
| |
12:33 | anuejn | one frame per file
| |
12:33 | se6astian_ | CinemaDNG specifies directory structures containing one or more video clips, and specifies requirements and constraints for the open format files, (DNG, TIFF, XMP, and/or MXF), within those directories, that contain the content of those clips.
| |
12:33 | se6astian_ | https://en.wikipedia.org/wiki/CinemaDNG
| |
12:33 | anuejn | ah interesting
| |
12:34 | anuejn | I am currently really writing CinemaDNG but it might be interesting to support mxf
| |
12:35 | anuejn | ah well but cimemaDNG in mxf is unfortunately not supported by the free version of davinci resolve :(
| |
12:36 | se6astian_ | one file per frame is perfectly fine IMO
| |
12:36 | anuejn | yeah at least for the moment
| |
12:37 | se6astian_ | its also fine for not just this moment :P
| |
12:39 | anuejn | I mean I can kinda see why it would be handy to have not a gazilion files laying around but \_( )_/
| |
12:45 | se6astian_ | but its also handy that you can quickly pick a single image from a sequence
| |
12:45 | se6astian_ | so I wouldn't see it as a showblocker
| |
19:35 | aombk | left the channel | |
19:37 | aombk | joined the channel | |
20:01 | aombk2 | joined the channel | |
20:01 | aombk | left the channel |