Quote:
Originally Posted by mikelj But surely, for some given content, if the ABT2010 performs 2:3 IVTC to 60p without issues, then there should be no issues with IVTC to 24p for the same content, since the same 24fps are being re-constituted? |
I guess the difference is between the "should" and "does".
There was a discussion around this subject during beta testing some ABT product based on the ABT2010. Here's a comment from ABT following a change they made from using "film bias" to "forced 3:2" mode for 24(29.976)hz output.
"The problem with any deinterlacing mode such as film bias which dynamically switches between film and video modes (aka, weave and motion-adaptive) is that it will sometimes lose lock on even a perfect source. This can happen for a number of reasons, such as very little motion in the image or a transition between very different scenes. When this happens, and particularly when the deinterlacer reacquires lock, it can sometimes be out of phase with what the frame rate conversion hardware is doing during the period of no lock on the 3:2 cadence. This will typically cause a stutter as the frame rate conversion logic adapts to the new situation.
The forced 3:2 deinterlacing mode is different. Once it initially acquires lock, it will never lose lock or produce a change in the 3:2 phase as long as the source maintains a steady 3:2 cadence. Consequently, for a 'perfect' source such as BD, you should get a perfect conversion from a 60 Hz source with 3:2 pulldown to the 24 Hz output. However, if the incoming 3:2 pulldown sequence is disturbed or momentarily disrupted, the forced 3:2 mode can take some time to switch to the post-disturbance situation. During that period you will likely see combing in the image rather than stutter. With a 'perfect' source this sort of disturbance only happens when the user alters playback, such as pausing or chapter-skipping. Many sources are less than perfect, however, and will yield problems with any deinterlacing and frame rate conversion method (short of, perhaps, motion compensation). These include many DVDs (and even some 'perfect' DVDs with a less than perfect player), movie trailers (which are normally very heavily edited), and nearly all broadcast material which was originally recorded on film, as the vast majority of that material is edited in the video domain without regard to the 3:2 cadence.
So the advantage of the new scheme is that it gives very good results for a very good quality source. The disadvantage is that with a less than perfect source you will see some combing at transitions (and maybe a little stutter, which is unavoidable). However, since a less than perfect source cannot be converted to a 24 Hz output without some type of disruption, you'll never get a perfect conversion anyway. In my opinion, you shouldn't convert any thing but a source with a perfect 3:2 cadence to a 24 Hz output. Otherwise, you're pretty much guaranteed to have problems of one sort or another."
Avi