Quote: I noticed there are sections where it repeats 1, 2(2,3), 3(3,2), 4, 5. Would you know how to write that as a script so I can run the whole clip though it and see how much it fixes? Or would each frame need to be done manually?

Right yes, as I suspected the cartoon is not 24fps, it's 18fps. Here is the "correct" way to get it back to 17.982fps:


Srestore is intelligent, so it should choose the frames that don't have chroma-bleeding most of the time. You can of course just use tdecimate and go straight to 17.982fps like this: TDecimate(cycler=2,cycle=5), but I did a test and srestore definitely performed a better job of selecting better frames. You can manually decimate the frames instead, but you would need to do that shot per shot!

(2017-02-11, 05:33 AM)Valeyard Wrote:

Quote: I noticed there are sections where it repeats 1, 2(2,3), 3(3,2), 4, 5. Would you know how to write that as a script so I can run the whole clip though it and see how much it fixes? Or would each frame need to be done manually?

Right yes, as I suspected the cartoon is not 24fps, it's 18fps. Here is the "correct" way to get it back to 17.982fps:

Code:

inputfile="thundercats.d2v"
mpeg2source(inputfile)
assumebff()
TFM(order=0,d2v=inputfile)
TDecimate()
AssumeFPS(24)
srestore(frate=18,speed=-25)
AssumeFPS(18000,1001)

Srestore is intelligent, so it should choose the frames that don't have chroma-bleeding most of the time. You can of course just use tdecimate and go straight to 17.982fps like this: TDecimate(cycler=2,cycle=5), but I did a test and srestore definitely performed a better job of selecting better frames. You can manually decimate the frames instead, but you would need to do that shot per shot!

Thanks very much for taking the time to do that script and having a look.

That does look better overall, though some of it does still have the same issue. I don't think I have the patience or time to do it manually, so I'm going to try plan B and track down the PAL disc.

The PAL disc is unlikely to solve the issue. The cartoon is 18fps, NTSC is 30fps, and PAL is 25fps. Even if they did a modern standards conversion, it'll still be inferior to the NTSC disc.

Also, I haven't seen the rest of the episode, so I can't say for certain the whole of the episode is 18fps, but that section is. It's not unusual with old cartoons to see a mixture of frame rates. For example, the opening credits could be 15fps, 24fps, 30fps while the rest of the episode is 18fps, or there could be parts done at 18fps and action sequences done at 24fps or 30fps, etc.

I just had a look to see if the episode is online anywhere and it's on dailymotion. http://www.dailymotion.com/video/x4vdfr8

The episode has none of the issues of my DVD as far as I can see, and when downloaded the fps is 25. This leads me to think the PAL DVD could be OK. The opening titles to the previous episode, which I had to replace due the issues with them, look fine here. http://www.dailymotion.com/video/x4vdfr7...th-time_tv

Yes it does. I guess if they IVTC'd it then there's less chroma crosstalk since it builds progressive frames.

On topic of chroma-bleeding... (note: not trying to hijack the thread!)

I was curious whether this kind of chroma-bleeding that is present here is similar to when colors are slightly out of place to what's onscreen (for example, during some TV broadcasts, especially older analog broadcasts) where it's not a matter of chroma-bleed from one frame/field to another but that the centering is off...

Would it be just a matter of "moving" the chroma layer to align with the rest of the screen or is there more to it than that?

I can show examples if necessary (and start separate thread, if needed)

Feel free to post here.