Query re type and AR of a DVD

[pooksahib]

Hi. I wonder if someone would be kind enough to give me a second opinion on the DVD I have here. I've made this clip using DVDDecrypter and then mpeg2cut2.

First question – is it truly progressive (I think so but have a niggling doubt)?
Question two - what is the proper aspect ratio? The DVD box says 4:3 but it looks fine to me in 16:9 and when indexed with megui, megui says “Input DAR 455:192 (2.370)”

I'm a tad confused... Many thanks.

[jagabo]

The video is encoded interlaced but the frames are progressive. But the two fields are identical -- made that way by a duplicate field deinterlacer (throw away one field, replacing it with a duplicate of the remaining field).

The picture is obviously 16:9 DAR. The MPEG 2 stream is marked 16:9 DAR, but there is a sequence_display_extension that indicates that DAR is in a 540x576 portion of the frame. That's why MediaInfo reports it as, and some players try to play it as, 2.37:1, or rounded to 2.40:1.

[LMotlow]

I don't understand why people refer to any old MPG as "DVD". This is just re-encoded b.s., re-processed for no improvement whatever. The growing frequency of this kind of garbage is depressing.

2.37:1 ???

[pooksahib]

Hi jagabo and thanks for that comprehensive answer - kind of you to take the time. It's good to know I don't have to treat this vid in any way (apart from cropping the sides maybe) - my last DVD needed QTGMC, took absolutely ages.

Out of curiosity, what were you using to find that 16:9 DAR mark and the sequence_display_extension?

[jagabo]

If you want better image quality get rid of the duplicate field and use nnedi3 to fix the remaining field:

Crop(8,0,-8,-0)
SeparateFields()
SelectEven()
nnedi3_rpow2(2, cshift="Spline36Resize", fwidth=1024, fheight=576)

Or if you want to keep it anamorphic use fwidth=704, fheight=576.

[pooksahib]

Thanks, never heard of nnedi3 before. I had a look at the wiki but don't quite understand: what's wrong with the remaining field that we need to fix it?

Out of curiosity (again), you seem to prefer Spline36Resize over Spline64. Any reason for that? And, if you wouldn't mind, there's my other curiosity question in post #4.

Thanks again.

[jagabo]

The second field in the frame is just a copy of the first. If you watch the video full screen you will see obvious aliasing artifacts. needi3 will clean up a lot of that aliasing.

I picked Spline36Resize is mostly out of habit. If I'm serious about a video I'll try different resizers to see what works best. Sharper isn't always better.

GSpot showed the sequence_display_extension in the MPG file.

[pooksahib]

Thanks. The process took a while but the result is good. I went with 1024x576 - would there be any material difference to the PQ or filesize if I kept it anamorphic, do you think?

[jagabo]

In general, a larger frame size requires more bitrate to maintain quality. With CRF encoding you will get a larger file. With bitrate based encoding you will get lower quality.

And each resize causes a loss of quality. Though with low res material like this you probably won't notice much of a difference going from 704 to 1024 wide.

[hello_hello]

The resizing filter can have a slight effect on sharpness, or lack of, compared to anamorphic encoding. I generally prefer to resize to square pixels with a sharpish resizer. The video from the DVD I used for these screenshots had already been sharpened a little too much, but that made the difference easier to see.
https://forum.videohelp.com/threads/378501-Resolution-Bitrate-and-Quality#post2444902

[pooksahib]

Thanks guys.