Filtering help needed

[Selur]

okay (not sure I'll remember but good to know atm. )

[x264]

How do i spot the 30p frames in a hard telecined DVD?

[jagabo]

How do i spot the 30p frames in a hard telecined DVD?

Each pair of fields comes from the same frame.

[x264]

so the 30p frames will look non-interlaced?. The DVD i'm having shows the 3:2 pattern.

[jagabo]

so the 30p frames will look non-interlaced?.

It depends on whether the pair of fields in each encoded frame come from the same source frame. It's possible for the field pairs to come from adjacent source frames, in which case the frames will have comb artifacts.

[x264]

It depends on whether the pair of fields in each encoded frame come from the same source frame. It's possible for the field pairs to come from adjacent source frames, in which case the frames will have comb artifacts.

Is there a software for finding these details from a DVD?

[jagabo]

It depends on whether the pair of fields in each encoded frame come from the same source frame. It's possible for the field pairs to come from adjacent source frames, in which case the frames will have comb artifacts.

Is there a software for finding these details from a DVD?

I use DgIndex in Honor Pulldown Flags mode, open via Mpeg2Source() in AviSynth, Then step through the video with VirtualDub with the Bob Doubler filter. Or just open the MPG/VOB file with VirtualDub (with the MPEG source filter) and use Bob Doubler.

[x264]

What are the mistakes that an IVTC filter like TIVTC will make besides field match failure?

[jagabo]

What are the mistakes that an IVTC filter like TIVTC will make besides field match failure?

Post processing in TFM() can mess up and accidentally deinterlace progressive frames. Pulldown can be used to convert any frame rate from 19.98 to 29.97 to 59.94 fields per second. So TDecimate() may not get the frame rate right unless you specify mode and rate. TFM().TDecimate() cannot account for field blended PAL/NTSC conversions.

[x264]

Do i need to change anything here other than the field order?

[Selur]

These settings are for deinterlacing and not for ivtc,... you might want to read up to understand what fields are, what interlacing, pulldown and telecine is.
If you understand the basics about these things you might start to understand what the different deinterlacing technics are,..

[jagabo]

Do i need to change anything here other than the field order?

You have to change the field order to match your source. Use whichever algorithm you want -- except double up frames. The point is you need to see each field seprately.

[x264]

These settings are for deinterlacing and not for ivtc,...

It depends on whether the pair of fields in each encoded frame come from the same source frame. It's possible for the field pairs to come from adjacent source frames, in which case the frames will have comb artifacts.

Is there a software for finding these details from a DVD?

I use DgIndex in Honor Pulldown Flags mode, open via Mpeg2Source() in AviSynth, Then step through the video with VirtualDub with the Bob Doubler filter. Or just open the MPG/VOB file with VirtualDub (with the MPEG source filter) and use Bob Doubler.

This is what i'm talking about.

[Selur]

than stay with bob()

[x264]

You have to change the field order to match your source. Use whichever algorithm you want.

What do i do after this?

[x264]

you might want to read up to understand what fields are, what interlacing, pulldown and telecine is.
If you understand the basics about these things you might start to understand what the different deinterlacing technics are,..

I know the basics.

[x264]

after loading the avs script in virtual dub there are two screens, the one on the left has more detail than the one on the right.

[Selur]

the avisynth script you are opening should only contain the Mpeg2Source(...) call no filtering,...

The left screen shows the input, the right screen shows the content after applying Virtual Dubs own filter

[x264]

It depends on whether the pair of fields in each encoded frame come from the same source frame. It's possible for the field pairs to come from adjacent source frames, in which case the frames will have comb artifacts.

So if there are no comb artifacts then it means that the pair of fields in each frame comes from the same source frame.

[manono]

So if there are no comb artifacts then it means that the pair of fields in each frame comes from the same source frame.

First, make sure there's movement. Otherwise even real interlaced sources will look progressive. Your statement (it means that the pair of fields in each frame comes from the same source frame) is correct, yes, assuming there's movement between frames.

Personally I think you're making this harder than it really is and in spite of what you said - I know the basics - I don't think you do. Either that or you're incredibly lazy. Otherwise we wouldn't have a thread nine pages long filled with mostly basic questions.

[x264]

Ok.

[x264]

I didn't understand this part, could any of you explain this part?

[jagabo]

There are no frames in interlaced analog video. Interlaced analog video is an alternating sequence of top and bottom fields. The top field consists of scanlines 0, 2, 4, 6... 478, the bottom field consists of scan lines 1, 3, 5, 7... 479. When you watch interlaced TV you see one field at a time, never an entire frame. Interlaced NTSC video is 60 fields per second (really 59.94 but we'll use 60 for simplicity), again, always alternating between top and bottom fields. How do you get get 24 frame per second film into 60 fields per second? First you duplicate frames to change 24 frames into 60 frames. To do that you need to multiply the number of frames by 2.5. Since you don't want to have half frames you alternate between duplicating frames 2 and 3 times, on average 2.5 times (24 * 2.5 = 60), hence a 3:2 duplication pattern:

24 fps frames: A B C D E F
60 fps frames: A A B B B C C D D D E E F F F

So 6 frames has become 15 frames (or, multiplying both by 4, 24 frames has become 60 frames) *

Now you pull fields out of those frames, alternating between top and bottom fields (t=top field, b=bottom field):

60 fields per second: At Ab Bt Bb Bt Cb Ct Db Dt Db Et Eb Ft Fb Ft

That's what you see on an interlaced TV screen. One field at a time. Alternating between top and bottom fields. Some of the original film frames are seen over 3 fields, some over 2 fields.

When you digitize and capture those fields it's customary to pack pairs of fields into frames, creating 30 frames per second:

captured frames: AtAb BtBb BtCb CtDb DtDb EtEb FtFb...

If you look closely you'll see that some of the frames contain fields from the same film frame (AtAb, for example), while others contain fields from two different film frames (BtCb, for example). When the two fields come from the same film frame the video frame looks progressive. When the two fields come from different film frames the video frame looks interlaced (if there was motion between the two film frames).



* What if the source wasn't 24 fps film? Webcams typically run at 30 fps. 8mm film ran at 18 fps. A PAL video might come to you as 25 frames per second. Instead of multiplying frames by 2.5 you multiply by a different number. 30 fps is multiplied by 2 (ie, every frame is displayed twice). 25 fps needs to be multiplied by 2.4 (frames are repeated in a 3:2:3:2:2 pattern). 18 fps needs to be multiplied by 3.3333 (repeat in a 3:3:4 pattern).

[x264]

How do i fix this blending which is present in a small section of the soft telecined DVD Episode?

[Selur]

-> https://forum.videohelp.com/threads/342729-applying-different-filters-to-different-sect...ia-avisynth%29

[x264]

Which filter should i use for fixing this?

[Selur]

there are multiple ways to do it, for details read the other thread

[x264]

One of my encodes which was running for about 24 hrs got cancelled after encoding 18000 frames. Can i encode the rest of the frames and join it with the other encoded file?

[manono]

Can you play the part that was encoded? What format is it in? Anyway, I join different Lagarith AVIs all the time, but it depends at least partly on what you got from the first encode. When I'm doing long and complex filtering on a source, I encode as lossless.

[x264]

The encoded mkv file plays fine in both VLC and MPC. The cancelled encode contains frames 0-19562.