Partially Field Blended NTSC DVD info

[Booshman]

I inverse telecined some DVD's and I have come across some field blended footage. I'm curious if anyone could tell me why some footage would contain blended fields when the th rest is fine?

To give an example. I IVTC'd a disc with six episodes of a cartoon. There was one episode with the intro full of blended fields. The same episode had just a few frames that were blended in one scene. Another episode has several minutes of blended fields in chunks thoughout.

How would this happen duing the mastering of the disc? Has anyone run into this issue before?

I'm going to try and get the PAL version of the disc to see if the same issue is present, but it's a bit frustrating to have to go to the extra effort.

[spoRv]

Are they blended fields, or frames? If they are fields, no way to recover them; if they are frames, it's possible to fix them using the correct fields - often the fields order change during a telecined program, don't ask me why, let's wait for the experts!

[Booshman]

I've checked the fields separately using Separatefields(). Looks like this:



So the footage is buggered as I thought.

I'm quite interested to find out what would lead to this happening, when the rest of the footage is fine.

[spoRv]

No solution but find another source... PAL version could be one.

[Booshman]

Yeah I realise that. I'm just very curious to find out the reason for this to have happened. What has gone on in the mastering process to have a few minutes out of 3 hours on a disc with blended fields.

[Chewtobacca]

I inverse telecined some DVD's and I have come across some field blended footage. I'm curious if anyone could tell me why some footage would contain blended fields when the the rest is fine?

How would this happen during the mastering of the disc? Has anyone run into this issue before?

Of course. This happens through indifference or incompetence or both.  I'd guess that someone ran their source files through some sort of automated convertor and didn't care much about what the result looked like.  It's quite common for video to contain a mix of different kinds of footage (e.g. soft telecined and hard telecined), each of which should be handled in a distinct way; use one approach to handle the whole video and the result is the kind of mess that you have just come across.

[Valeyard]

There are solutions to that problem. The main question I have is how does it look prior to IVTC? Cartoons are not always 24fps, if it's a US cartoon it could be 30fps.

[Booshman]

There are solutions to that problem. The main question I have is how does it look prior to IVTC? Cartoons are not always 24fps, if it's a US cartoon it could be 30fps.

It's from Thundercats a US cartoon.The pic is without IVTC applied.

[Valeyard]

Right, so it's in the source. I'd suggest using a manual algorithm to decimate fields and seeing how long before it breaks. Essentially what you have is not this:

TBTBTBTBTBTB -> TTBTBTBBTBTBBTB ... but something like: TXBTBTBXTBTBXTB ... where X is a blended field. That's the best case of course, you could have frame-blending which will be harder to remove. Can you post a decent length sample at 59.94i (~5-10 mins, sound not required)?

[Booshman]

How do I get it to 59.94i? Assume FPS(59.94)?