Page 1 of 1

[Update] CineForm Loader/Saver 1.4

Posted: Tue Feb 05, 2019 4:11 am
by cinewrangler
New in this version:

Loader:
  • timecode in DPX files now follows SMPTE 12M specifications
  • reading drop-frame timecode flag (and adding it to metadata)
  • fixed: incorrect gamma curve conversions for files that don't have the encoding curve tag set in the CineForm metadata, i.e. all files written by Premiere Pro or After Effects
  • fixed: bugs in the CineForm SDK that would cause alpha channel values to be encoded incorrectly
Saver:
  • added "Segment File" option to write large QuickTime MOVs as multiple smaller files
  • timecode in DPX files now follows SMPTE 12M specifications
  • improved encoding speed when writing QuickTime MOVs
  • improved encoding speed when image format is float16
  • fixed: compatibility issues caused by the way the QuickTime MOVs have the properties of the timecode track set
  • fixed: bugs in the CineForm SDK that would cause alpha channel values to be decoded incorrectly

Re: [Update] CineForm Loader/Saver 1.4  [SUBMITTED]

Posted: Wed Feb 06, 2019 5:44 pm
by SecondMan
Terrific update, @cinewrangler! Thank you for sharing.

Submitted! :)

Re: [Update] CineForm Loader/Saver 1.4

Posted: Tue Feb 19, 2019 2:04 pm
by RBemendo
In this version, installed from Reactor, I'm unable to use CineForm loaders in Fusion 9.0.2. I've tried CineForm files from Premiere and After Effects, as well as from Resolve. I've even created a Cineform file in Fusion and tried to bring that in as a loader, and FU crashes. I'm experiencing this on multiple computers. Any thoughts?

Re: [Update] CineForm Loader/Saver 1.4

Posted: Wed Feb 20, 2019 4:58 am
by cinewrangler
RBemendo wrote:
Tue Feb 19, 2019 2:04 pm
I'm experiencing this on multiple computers. Any thoughts?
Your "multiple computers" all run the same basic OS (i.e. Win, Linux or Mac)? If so, which one?

Edit: And can you please try to write DPX-C with Fusion and then see whether the Loader also crashes on those?

Re: [Update] CineForm Loader/Saver 1.4

Posted: Wed Feb 20, 2019 7:32 am
by RBemendo
The multiple computers are all on Windows 10 64bit.

When writing a DPX-C with Fusion, the loader is seen with correct trim in and trim out loader settings. But it still crashes.
When importing CineForm .mov files, the trim in and trim out loader settings are not set and are not read as an actual loader.

Re: [Update] CineForm Loader/Saver 1.4

Posted: Wed Feb 20, 2019 10:36 am
by cinewrangler
Hhm, I wasn't expecting this to be happening on Win10. I'm doing all my testing on an ancient Win7 here (since I don't want an OS that decides in its own when it is time to install updates).

Edit: Found the problem. Also get crashes here with YUV 4:2:2 CineForm files. I'm mostly using RGB files which don't crash. Will try to fix it and post an updated version soon.

Re: [Update] CineForm Loader/Saver 1.4

Posted: Wed Feb 20, 2019 11:29 am
by RBemendo
Great to hear you found it! Thank you! This plugin has been a great workflow helper in Fusion.
Cheers.

Re: [Update] CineForm Loader/Saver 1.4

Posted: Wed Feb 20, 2019 11:31 am
by SecondMan
Some fine troubleshooting going on here :)

Thanks both!

Re: [Update] CineForm Loader/Saver 1.4

Posted: Thu Feb 21, 2019 2:00 am
by cinewrangler
The bugfix story continues here.
RBemendo wrote:
Wed Feb 20, 2019 11:29 am
This plugin has been a great workflow helper in Fusion.
Actually if you use CineForm as an intermediate codec between PP/AE and Fusion I'd recommend using the RGB version over YUV. It has 12bit instead of 10bit, can carry alpha and best of all it is RGB. Most (nowadays maybe even all?) rendering is done in RGB anyway, so I simply don't understand why people still insist on converting to YUV when writing their image data to file. Compared with RGB, YUV has very bad codeword utilization (only about 20% of the available code-words in 10bit actually represent an RGB color, making 10bit YUV roughly equivalent with 9bit RGB in terms of bit-depth). If you think you can preserve superwhites with YUV legal range, just use a proper log transfer curve for RGB instead.