Nate Weaver |
January 14th, 2006 07:50 PM |
Quote:
Originally Posted by Les Dit
The encoded video is only 5 megabit , so there are plenty of compression artifacts due to that. 9 megabits/sec works better for 720p.
The edge enhancements are far too evident for a film like 35mm transfer. For example, look at the edge of the buildings in the train footage. The ringing ( overshoot) due to the video edge enhancements are a disappointment, as I know this can be turned down in the camera.
I come from a film background, and I'm starting to think that folks coming from a video background just aren't tuned into these defects.
If the test was supposed to see how close the hd100 can come to a film look, then re shoot with the sharpening turned down. Film *never* has overshoot or ringing.
-Les
|
I learned a lot about this from Adam Wilt during the course of the 4 camera shootout last week. I got to see on a scope just how too much image enhancement destroys the signal.
I had been working with -3 or -4 detail on my HD100 prior to the tests, now I'm doing more of my own testing with -9 or detail off completely. As Jay Nemeth said to me...you can always add detail in post, but if you shoot with it, you can never take it away. I knew that before, of course, but adding an unsharp mask in post always seemed like it'd be a pain.
Since color-correction products like DaVinci do their own edge-enhancement when transferring film, but the film itself of course has none...I've decided maybe the best approach is to treat edge enhancement like I do color-correction. Necessary, but something best done in post where I can fine tune it for the intended distribution (small screen NTSC, large screen HD, or even NOT adding it for a filmout).
|