Appropriate IVTC method ?

[varunb]

I have a NTSC telecined DVD movie & i want to encode it to 23.976 progressive avi format. My basic problem is that which ivtc method should i use to remove all the interlaced artifacts? I was going to use the 'forced film' method in DGIndex but then i stumbled upon the ivtc tutorials in one of the sticky threads. It said that even the forced film method is not perfect for those dvd movies which are not 100% film & to use the telecide plugin in avisynth. Does this telecide filter produces the same output video as the Telecide filter (found at neuron's site) made for virtualdub ? If not, then which IVTC method should i use that will remove all the interlaced artifacts ? Should i do forced film first & then deinterlace it in virtualdub using the smart deinterlace filter at the frame rate of 23.976 ? I just want my inverse telecined video to b lossless in quality with non-jerky motion (basically for tv playback)

[neuron2]

There's no way we can tell you the appropriate method to use without knowing more about your video. I can say that as long as you are using Avisynth anyway to serve the video, you should prefer the Avisynth filters to VirtualDub ones.

What is the film/video percentage reported by DGIndex?

You can try these methods and decide which is better:

1. Force Film in DGIndex alone.

2. Force Film in DGIndex with a postprocessing deinterlace step such as FieldDeinterlace(full=false) in your script.

3. Honor Pulldown Flags in DGIndex followed by IVTC in your script, using Decomb (Telecide + Decimate).

[varunb]

Well dgindex says that the Video Type = FILM 99%. So what do u think, Neuron 2 ? Now which method should I use ?

[neuron2]

Use method 1 of my post above. If the result has any objectionable combing, use method 2.

[varunb]

I think i should prefer the decomb filter & do ivtc rather than forced film. there's one more thing. since my movie is 99% film should i use the syntaxes for normal ivtc or syntazes for hybrid movie in the avs file when using the decomb plugin ?

[neuron2]

I think i should prefer the decomb filter & do ivtc rather than forced film. Why?

there's one more thing. since my movie is 99% film should i use the syntaxes for normal ivtc or syntazes for hybrid movie in the avs file when using the decomb plugin ? Since you're ignoring my advice without citing a reason, why should I bother to reply?

[varunb]

well its given in the tut that the decomb filter is the default choice for performing true ivtc as forced film leaves certain artifacts. I am also not sure whether the 2nd method u had posted will remove all the artifacts as I have never encoded any movie using the avs script as frameserver. Now can u tell me finally what i should i do ?

[neuron2]

I already told you.

Why don't you *try* the methods and see what you get? :)

From the guide: "Most of the time Force Film will work fine at those high percentages". The only way you'll know for sure is to try it. When Force Film can work, it is better, because it is faster and more reliable than heuristic field matching.

At 99% film it would be pointless to process as hybrid video. That 1% is probably the starting logo or the credits.

[manono]

Since I wrote that portion of the tutorial, maybe I can help (and maybe not). Force Film usually works with those high percentage but less than 100% Film D2Vs. Often the Interlaced/Video portions are at the very beginning or between scenes, and fall during some black frames where there won't be any artifacts.

Naturally enough, you don't want to get burned later by finding a few interlaced frames, which often really stick out, and are really annoying. Personally, I often IVTC that kind, just to be on the safe side. Or you can go Tools->Parse D2V in DGIndex and see if you can find where the video portions are, and see if they are on black frames or during some active video. Sometimes it's just at the very beginning of the film.

Or you can take neuron2's advice and add FieldDeinterlace(Full=False) or TDeint(Full=False) to a Forced Film D2V to clean up any stray interlacing that might sneak through. Or, if as he suggests, you find the video part at an opening logo or end credits, you can add the conditional deinterlacer just for that part, and leave the rest alone.

[StephenChow]

I used Force Film for Western movies (like The Soul Plane), DVD source is PROGRESSIVE, percentage it does not show, it show FILM instead. The result is the D2V movie turn into 23.976fps (I mean pulldown progress ) but still contain A LITTLE interlaced frames, about 5 minutes, there are 2 interlaced frames.
As a result of this, I used Decomb IVTC or TIVTC all the time.
However, I'm thinking about it since Neuron2 said Forced Film is better if it works. Thank you for inform me.

To varunb:
after DGIndex generates D2V files, you should make .AVS file that point to this D2V and then put it AVS file in VirtualDub and check every frames to ensure that it works fine. Finally encode this AVS.