k0r0n4
27th November 2004, 09:02
Ok, I apologize if the topic is confusing but I didn't know how to sum this problem up in a few words.
Pretty much it's like this. I have a music video made by someone else that I'm trying to repair. From what I can tell, this is what he did.
Grabbed the source footage (NTSC)
IVTC
Did ***NOT*** Decimate to remove duplicate frames. (This is the source of the problem. But wait, there's more!)
Timed the video to audio by grabbing different parts of the footage
I'm assuming the source footage after being IVTCed did not have an abnormal pattern of duplicate frames (meaning it could have been fixed with the standard Decimate() filter and not SmartDecimate). Thus, the final footage looks a lot like this:
A B C D D E F G H H I J K L L M*N O O P Q R S S T U V
where each letter is a frame and not a field (remember, it's now progressive) and a * denotes a scene change. Thus the pattern may (and probably will) become offset at a scene change, as you can see in the previous example.
Now, my questions are:
A) Is there a way to get rid of the duplicate frames WITHOUT loosing synch with the audio?
B) And, if so, how would I go about it?
C) Also, would I need to use a variable frame rate?
Also, on a related note, would it be better to use a filter such as "Dup Filter: Replace near duplicates with exact duplicates" and then send the clip through SD with exact frame matching or instead use SD's "Diff NUM: WHAT DIFF" option which reads:
indicates the difference between to source field NUM and NUM+1. WHAT is the classification of the difference with 0 meaning the same, 1 similar, and 2 different. DIFF is the numerical difference.
Or is "Dup Filter" just meant for clearing up noise?
Pretty much it's like this. I have a music video made by someone else that I'm trying to repair. From what I can tell, this is what he did.
Grabbed the source footage (NTSC)
IVTC
Did ***NOT*** Decimate to remove duplicate frames. (This is the source of the problem. But wait, there's more!)
Timed the video to audio by grabbing different parts of the footage
I'm assuming the source footage after being IVTCed did not have an abnormal pattern of duplicate frames (meaning it could have been fixed with the standard Decimate() filter and not SmartDecimate). Thus, the final footage looks a lot like this:
A B C D D E F G H H I J K L L M*N O O P Q R S S T U V
where each letter is a frame and not a field (remember, it's now progressive) and a * denotes a scene change. Thus the pattern may (and probably will) become offset at a scene change, as you can see in the previous example.
Now, my questions are:
A) Is there a way to get rid of the duplicate frames WITHOUT loosing synch with the audio?
B) And, if so, how would I go about it?
C) Also, would I need to use a variable frame rate?
Also, on a related note, would it be better to use a filter such as "Dup Filter: Replace near duplicates with exact duplicates" and then send the clip through SD with exact frame matching or instead use SD's "Diff NUM: WHAT DIFF" option which reads:
indicates the difference between to source field NUM and NUM+1. WHAT is the classification of the difference with 0 meaning the same, 1 similar, and 2 different. DIFF is the numerical difference.
Or is "Dup Filter" just meant for clearing up noise?