do i need to deinterlace when editing a dvd back to dvd?

[kopmjj]

Hi quick question i have some dvds i want to edit , but is it necessary to deinterlace if im gonna end up making another dvd?
say for example i want to reduce noise from a dvd and change color settings and sharpen and put it back to dvd again.

and if i dont need to deinterlace then can i edit vob files direct or do i have to convert to mpeg2 interlaced do the editing and encode back to dvd?

[aedipuss]

vobs contain the mpeg. you can extract it with vob2mpg. use the mpeg to edit.

[redwudz]

Generally, don't deinterlace if the end format is also a DVD.

[kopmjj]

ok thanks i asked because when i deinterlace using top fields i get jerky playback on some scenes even though Gspot says the dvd vob source is interlaced TFF, maybe it was a pal to ntsc convert so originally it was BFF?

but if i dont have to deinterlace then it will be great time saver because i only use TGMC so its slow as hell!!

ok so i use vob to mpg then edit the mpg which is still interlaced then create a dvd from the interlaced mpg right?
because i tested on a small sample and with Gspot the final VOB said progressive and no info whether its TFF or BFF thats why im confused, i wanna know if i done it right?

[kopmjj]

ok i just checked with sony vegas when i load a vob the info says field order=upper field first,mpeg2,but when i extracted the dvd with tmpge to uncompressed avi and loaded it into vegas the info now says lower field first?
so which is correct? just incase im need to deinterlace it.

i like uncompressed avi better i tried dvd to vob but compression was bad

[lordsmurf]

Look at it in GSpot. That's the correct info, either TFF or BFF there.

Why are you using Vegas?
I have a feeling you're doing something the hard way, using the wrong software.

[kopmjj]

Look at it in GSpot. That's the correct info, either TFF or BFF there.

Why are you using Vegas?
I have a feeling you're doing something the hard way, using the wrong software.

ok maybe you can help this is what im trying to do....

1. extract dvd to uncompressed avi
2. deinterlace
3. do loads of editing with avisynth and save to uncompressed avi again
4. either load avi in tmpge or sony vegas and render to mpeg2 dvd
5. use a good dvd author software to make dvd.

but i would like to miss step 2 deinterlacing so what would be the best way to go about it?
also what ever software i use (if Gspot is showing TFF) do i always have to select upper field even if the dvd is ntsc?
or shall i load the uncompressed avi and choose progressive for rendering to mpeg2 thats the bit im confused progressive or upper field first? after editing avi in avisynth?

basically i do not want to touch deinterlace if possible
thanks.

edit: or how about i rename the vob file to mpg, then do the editing in avisynth and save it to mpg then rename back to vob, and slot it back in my dvd folder without re encoding?

will i lose quality?

[aedipuss]

use vob2mpg to extract the mpeg-2 video from the vobs. use that to edit in vegas with. no de-interlacing required. save as dvd spec mpeg-2.

[Gavino]

You can do basic edits (cuts, etc) or color adjustment without deinterlacing, but any filtering that involves mixing pixels in any way (incl noise reduction or sharpen) will require deinterlacing (or at the very least field separation).

[kopmjj]

You can do basic edits (cuts, etc) or color adjustment without deinterlacing, but any filtering that involves mixing pixels in any way (incl noise reduction or sharpen) will require deinterlacing (or at the very least field separation).

ok thank you

[2Bdecided]

You can do basic edits (cuts, etc) or color adjustment without deinterlacing, but any filtering that involves mixing pixels in any way (incl noise reduction or sharpen) will require deinterlacing (or at the very least field separation).

Not "any processing that involves mixing pixels in any way" - you could easily do a mix-fade etc without deinterlacing, also panning/sliding. horizontal sharpening or bluring, etc.

A decent NLE should handle this internally without you caring.

And yes, the OP is making this job far too hard! Vegas will happily edit MPEG-2 from a DVD without all this messing around. So will VideoReDo - without encoding.

Cheers,
David.

[Gavino]

Not "any processing that involves mixing pixels in any way" - you could easily do a mix-fade etc without deinterlacing, also panning/sliding. horizontal sharpening or bluring, etc.

OK, David - I was deliberately simplifying things to make the decision easier. The essential point is to avoid spatial filtering that mixes the two fields.

A decent NLE should handle this internally without you caring.

And yes, the OP is making this job far too hard! Vegas will happily edit MPEG-2 from a DVD without all this messing around. So will VideoReDo - without encoding.

Are all NLEs decent enough to know when to deinterlace? And do they make a good job of it when they do?

I don't know if VideoRedo does noise reduction or sharpening (as wanted by the OP), but in any case re-encoding would always be required for this.

[2Bdecided]

Sorry, I was looking at post 7 ("loads of editing") which implies Vegas would be great for the job. Looking at post 1 (filtering), obviously AVIsynth and re-encoding is the way to go.

(I also worry that many NLEs are quite stupid, but at least in Vegas if you tell it the input and output is interlaced and how to treat the fields it won't do anything too stupid - just not always optimal)

Very basic question though: are the original DVDs interlaced? Does the OP know how to check?

Cheers,
David.

[lordsmurf]

You can do basic edits (cuts, etc) or color adjustment without deinterlacing, but any filtering that involves mixing pixels in any way (incl noise reduction or sharpen) will require deinterlacing (or at the very least field separation).

This isn't even remotely accurate.

[Gavino]

You can do basic edits (cuts, etc) or color adjustment without deinterlacing, but any filtering that involves mixing pixels in any way (incl noise reduction or sharpen) will require deinterlacing (or at the very least field separation).

This isn't even remotely accurate.

Only slightly inaccurate, I'd say.
I qualified my remark further in post #12 - the point is to avoid spatial filtering that mixes the two fields.
Now, some editors might handle this automatically (though not necessarily well), but the OP was (mostly) talking about using Avisynth.

[vhelp]

I guess it depends on how the op means by editing (which can mean lots of things) and deinterlace.

example..

a) filtering for color or noise enhancements ?
b) restoring sources original frame rate ?

then..

if {a} you: separate the fields and do your enhancements.
if {b} you have to determine your source video if:
ivtc, pal->ntsc, or ntsc->pal, and perform the necessary methods for it.

Now, depending on the editor, (using avisynth in this example) it will depend on the plugin/filter requrements, like field separation and color space are the two common requirements.

Note, the plugin may already do its own field separation, while some may not and you will be responsbile for that aspect.

The frame separation (into fields) is for optimum filter processing and quality. We can use the build in function, SeparateFields() in avisynth if the plugin does not support its own.

Also, some plugins will have a color space requirement, ie YUY2 or YV12, or both, or else RGB, in order to perform the given function.

(Still, there is nothing wrong with filtering whole frames as long as they are progressive. I have done this to a few test videos while working on a few filter algos, delphi code)

There is also determinating the ideal entry point of A with B when performing enhancements to the video. Sometimes it is better done after deinterlace (ie, ivtc) while it may be better to do, before. It all depends on the source and your criterias and how well you know your source(s) and your goals for each video.

-vhelp 5396

[2Bdecided]

The frame separation (into fields) is for optimum filter processing and quality. We can use the build in function, SeparateFields() in avisynth if the plugin does not support its own.

A simple "separatefields" can't be described as "optimum". The resulting "frames" are misaligned and aliased. This might not matter for some processing, but it's not optimum. Some flavour of half-decent bob-deinterlacing which preserves the original fields is usually the way to go.

Cheers,
David.