-
Quality loss exporting to QT Ref
Hi,
I have found heaps of comments on how exporting QT reference files do not impair the quality of the video, however…
… At work I’m burning subtitles in the video with FCP. We use a piece of software that exports an XML file and creates a TIFF picture per subtitle block. We import the XML in FCP, relink the files and we’re done importing subtitles!
Once I’ve layered the video underneath, I export the sequence as a QT Ref. I end up with perfectly sharp subtitles over slightly pixelised video. The pixelisation translates into awful aliasing when compressing to SD MPEG2 from Compressor.
When I send the sequence to Compressor straight from FCP, without creating a QT Ref, the compressed SD video does not show such artifacts.
My conclusion is that something happens around the QT Ref conversion. Is it really lossless? Does it take bandwidth from the video to burn in the subtitles?
FWIW, the original video material in this case is 50Mb/s XDCAM and does not show pixelisation. It seems that XDCAM does not support more than a 50Mb/s rate which is what got me thinking that the subtitles were using too much of the bandwidth. Is that even possible?
I would understand if the whole video would be a little less sharp once the subtitles get burnt-in, but the difference between the sharpness of the subtitles’ text and the pixelisation of the video is quite surprising.
Any idea on the matter would be greatly appreciated. I’ll particularly appreciate little hints like a link to “video compression explained” or more specific articles such as “how FCP creates new video from several sources”.
Thanks for your time!
