Reproduced your test, seeing the same thing. Not sure if the shift is the same, but there's a shift.
I bet that if you are not Fast Importing, that QuickTime is doing the decode, and feeds an incorrect signal to the Avid DNxHD encoder.
If you Fast Import, it seems you bypass QT decode. Transcoding after that preserves levels just fine.
J
On 27 nov. 2011, at 02:54, johnrobmoore wrote:
> Played around on my mbp system os 10.6.8. I created a 1080i 59.94 RGB 709. I imported HD bars tiff to prores 444 and then exported a QT same as source. If I import that back in and preserve resolution levels on the internal scope seem right on and vector dots are good. If I toggle to YCbCr and import and preserve things are still fine. If I choose to convert to DNxHD 145 there is a definite level shift on import. Going back to 709RGB and importing converting the 1:1 10b yields a seemingly identical level shift. All imports are set to 601/709. These shifts are very similar to what I've seen in earlier versions like 4.0.x when exporting to prores 422 and reimporting. Seems like somehow the QT gamma issues are happening but I'm just comparing from memory not a current empirical comparison. This is almost as fun as checking frame accuracy on FCP before FCP_X. ;-)
Sunday, November 27, 2011
Re: [Avid-L2] Avid ProRes
__._,_.___
Search the official Complete Avid-L archives at: http://archives.bengrosser.com/avid/
MARKETPLACE
.
__,_._,___
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment