PAL->NTSC - Standardized Conversions to NTSC ie, PAL[Dr. Who]->NTSC

[vhelp]

Because of my limited (lack) of dealings with this source type, I was wondering
a few questions circling my thoughts..


** What is the standard for conversion that those who transfer this source to NTSC
(ie, PAL->NTSC) for NTSC viewers ??

** Is it tipicle to assume that this source type is Edited at either times,
in PAL format, and/or convert to, PAL->NTSC ?

** is there a Film format for PAL.


Some NTSC Examples:

A - NTSC FILM is 24p, where every frame is a progressive frame.
B - while non-FILM is PURE interlace at 29.970 fps (2 fields interlace into 1 frame)
C - while another is Telecine, and at 29.970 fps with 3:2 pulldown framing system,
evidense by 3 progressive frames, 2 interlace frames, or pseudo coded as,
PPP ii PPP ii PPP ii PPP ii PPP ii PPP ii ..


Some PAL Examples:

To my knowledge, PAL services only one speed: 25 fps.

I guess, my question is, does PAL have two (or three) formats:

A - 25 fps PURE interlace ( ie, DV cams )
B - 25 fps where each frame is Progressive ( ie, DV w/ Film feature? )
or,
C - 25 fps w/ a PAL version of Telecine, ie.. 2:2:3:3:2:2 (I just made that one up)
And, where a user could IVTC in the PAL way.


Reasoning behind my questions ...

The reason I'm asking, is because of my latest adventure in PAL->NTSC source
with the new Dr. Who tv show, which is a British source, meaning PAL 25fps.

I'm trying to identify weather or not it is a PAL Film source and a PAL->NTSC
conversion source. Then, I want to undo (after figuring out the mode used) the
PAL->NTSC, using the correct fps in the final MPEG-2 that I would make.

Regarding the Dr. Who series. I think that because this source is PAL, that
in the end, it should be brought back to PAL and 25 fps. I'm guessing that
there are different MODES used in the progress. But, the tricky part to me,
(in my mind) is this..

If the Dr. Who series is a PAL Film source, did they:

A - do a PAL form of Telecine to their source ?? or,
B - leave it as PAL 25 fps progressive frames ??

And, if B, then is this source 1) PAL->NTSC converted, and then somewhere inside this
process, 2) a NTSC Telecine is performed.

I'm getting the impression that its either A or B, (and then what I just mentioned afterward) But I'm just not sure of it all. I'm quite confused w/ Dr. Who
and PAL sources converted over to NTSC


And, for the final conversion to MPEG-2 and DVD medium ...

A - I think it all boils down to this. If it is a PAL source, then I should do a
conversion *back* to PAL 25 fps, and encode to PAL MPEG-2 formats.

B - Or, do a IVTC on the PAL->NTSC[Telecine]->29.970fps, where I bring it back to the [b]PAL->NTSC->23.970fps[/u] Film.

I don't know.

Any advice or explanations would be greatly appreciated.. really.

Thanks,
-vhelp 3894

[jagabo]

The new Doctor Who series was originally 25 fps progressive PAL. The SciFi channel is broadasting it with that crappy partly blended pulldown. You can use AVISynth and the RePal filter to restore it to the original 25 fps but some of the blending will remain.

[vhelp]

The new Doctor Who series was originally 25 fps progressive PAL. The SciFi channel is broadasting it with that crappy partly blended pulldown. You can use AVISynth and the RePal filter to restore it to the original 25 fps but some of the blending will remain.

This is what I wanted to know, in all truth. Thank you

Also, regarding the standards conversion.. this is what I was talking
about. Is this the standard that *everyone* uses when they convert PAL sources
to NTSC. And, if this is so, then the above is what we should expect.

Still, I think its very important to know *HOW* they are doing it, standardwise.
I mean, just because the SciFi channel has that one in the format, does not
mean that other channels will have a different format. I believe that no matter
what the channel or broadcaster, the standards conversion way will be used in
the conversion process, and the expectations would be the same.. blended garbadge
left over, etc.

- As such, I think we should catagorize them all, so that we may be better prepared
... to handle the various soruces that come by our way for processing. Plus, it would
... save us a lot of time - knowing ahead what conversion-type we have

-vhelp 3896

[vhelp]

What a pitty. If this is truth, that we cannot *completely* and accurately
IVTC this conversion-type source

I must admit though. I have tried briefly, TMPGenc and AVIsynth methods
of my own brew, but they both failed. And as I said, I want compelte and accurate
IVTC -- not partial.

jagabo, just out of curiosity. Did you give your hand a try on this conversion-type
source for Dr. Who with TMPGenc (even AVIsynth) ??

I'm just curious to see if you even thought it was worth it or not.

Maybe later, I'll UPLOAD a small sample that has a good detail motion view
of each frame, so that it would be easier to attempt. I had thought about posting
a short "IVTC - best shot competition" for the the Dr. Who source. But I dont' know.

As I was researching my captured source. I was realizing how its very important
to spot every bit of detail (speaking on interlace) of each frame, and particularly
with frames that have good decernable motion in them to determing interlace; blend;
etc. in them.

- You see. I had this idea to create a tool that could aid in this process.
... But I need a clean source to start with. Can't work with blind sources. And this
... source seems to be such. Anyhow.
... .
... With this tool, it would act closely like TMPGenc's pattern generator does. But,
... instead, you look at the frames (fields) for certain consitancty/inconsistancy,
... and log each one. Then, the tool adjusts accordingly, and finally, outputs a
... generated pattern to use - similar to TMPGenc's. But, this could be used for
... AVIsynth's built-in plugin, SelectEvery() for instance.

I don't know. I'll have to give it some thought

In the mean time, thanks gain jagabo.

-vhelp 3897

[FulciLives]

In the case of the new DOCTOR WHO you are probably best buying the UK PAL DVD releases so that you have true progressive PAL video.

The next best thing would be to download the torrents which came from the PAL versions and as such are 25fps progressive.

In other words get a version that hasn't already been converted from PAL to NTSC.

Here is a link to the UK PAL DVD boxset of DOCTOR WHO (complete season one of the new series)
http://www.xploitedcinema.com/dvds/dvds.asp?title=5766

- John "FulciLives" Coleman

[vhelp]

Fulci,

In all honesty, I have no interest in the actual and true 25 Progressive PAL
version. Also, I have no interest in D/L'ing any user-made, true PAL -> NTSC
from the internet. The thrill (as always) is in the hunt.. if you know what I mean.

I'd like to stick with the way it was presented to me by my cables' SciFi channel,
and my capture card(s) capabilities to reproduce this source for processing, all
inside this hobby of mine.

-vhelp 3898

[edDV]

In the case of the new DOCTOR WHO you are probably best buying the UK PAL DVD releases so that you have true progressive PAL video.

Which Dr. Who series are talking about?

I'm almost certain the original was shot in the studio in interlaced PAL with TV cameras much like most BBC/ITV productions of the day. Only the high budget major dramas were shot on film.

Hollywood needed to shoot everything on film because they were targeting all world standards. BBC tended to originate in video since the target audience was mostly PAL. US got standards converted versions.

Correction: This was one of the World market series and was shot on film at the BBC's Ealing and Lime Grove studios which were film.

The editing could have been done in film or video.

[FulciLives]

@edDV

I'm talking about the newest DOCTOR WHO that only aired on UK TV this past 2005

Did you follow the link I provided in my other post?

@vhelp

When I said try the torrents I was referring to the original PAL torrents that exist ... the PAL torrents are from the PAL source before anyone tried to convert it to NTSC.

No reason to live with a buggered NTSC conversion if the original PAL exists.

- John "FulciLives" Coleman

[vhelp]

fulci,

When I said try the torrents I was referring to the original PAL torrents that exist ... the PAL torrents are from the PAL source before anyone tried to convert it to NTSC.

I did realize this. I was speaking for *both* methods of obtaining the whatever
source version there are.

No reason to live with a buggered NTSC conversion if the original PAL exists.

Yes, that is so true. But, I am doing this strickly for the hobby of it all.
You know.. Capture and Encode and Author, yada yada.
And, if I want the dvd, I'll buy it. But I don't want to.. not for when its my
hobby, you know. And at this time, I have no need for it. And, the quality for
the level of this source is satisfactory -- yes, even for when I record it to VHS
tape, *and work off from that source medium* -- Yes, you heard that right

Well, I must give it some more thought and work. My goal for this Dr. Who project
was to figure out the IVTC, if it were possible. I'm still interested in trying
my hand at it again, using a method I deviced for this source when I first started
this, back in March 18th this year.

However, for standards sake, I still believe that there are standard conversion
processes that are performed on such sources. They just have to be identified.
Must like Film is..

A - capture a tv program to HDD
B - identify weather or not it is Telecine, and with known standard field-based patterns
C - knowing which pattern, ie.. for NTSC Film Movie, is a 3:2 pattern
D - using a known tool to process, ie, TMPGenc's Pattern Generator, set to: [10100 ]
E - process and encode to MPEG

This is what I want to do with PAL sources. (OK, Dr. Who sources)
It's not a matter of just using a familiar friends tool or plugin or script or
whatever. That's not always going to work. You have to know the source.
You have to know how to identify the source and with it's attributes are. Much
like the above example..

** we know the source is Film.
** we also know that Film uses 24P, but is Telecined to 3:2 Pulldown to 29.970 fps
** we know that an IVTC for this would require (for TMPGenc) code: [10100 ] to
... completly reverse this back to the original 23.976 fps progressively.

Thus, I'm trying to do the same with Dr. Who, which is PAL based. Oh, but there
is a screw thrown in.. that being a Region Conversion. And because of this, it
is more difficult, if not impossible to reverse, as we did for the NTSC exampe.

But, facts must be true, regarding Dr. Who series for instance. Thus..

A - is Film based PAL always Progressive 25p frames ??
B - is Film based PAL sometimes, PAL Telecined ??
...... and if so, how ?

I'm simply trying to identify and associate the source with the aboves' A or B.
Then, I want to determine what they did as a PAL->NTSC region conversion..

C - they did an A->PAL[Dr. Who]->NTSC, or
D - they did an B->PAL[Dr. Who]->NTSC

Now, if the source was C, it would be a matter of undoing the frame rate, and
identifying the field-blends, etc.. yada yada.

But, if it were a D, then we need to figure out which pattern they used in the
PAL Telecined, in order to properly undo (hopefully) completely and
accurately, back to the PAL version, and finally, MPEG-2 encode and author as
PAL video.

My hunch is that we have to aim for that PAL 25 fps.

But, if the region conversion was a design for a permanent NTSC type showing, then
we have to aim for an NTSC IVTC of..

** 29.970 fps -to- 23.976 fps

..for an MPEG-2 encode and author as NTSC Film based video.

-vhelp 3899

[manono]

Geez, why so many words? Why are you trying to reinvent the wheel? Did you even try what jagabo suggested so many lines of writing ago?

There are pretty much 3 ways to convert 24fps film to PAL 25fps. The usual way is to simply speed up the video and audio, the infamous 4% PAL speedup, where everything remains progressive, even though for DVD it's almost always encoded as interlaced.

The second way is to perform a telecine of sorts. An additional frame is inserted every second. The video remains the same length, and the audio doesn't have to be speeded up. This extra duplicate frame is usually inserted in the form of a field every half second for smoother-playing purposes. When looking at the frames when this method is used, you'll have a half second of progressive frames, followed by half a second of interlaced frames. You get a field reversal every half second whenever a new field is inserted. When doing an NTSC->PAL conversion from an NTSC DVD using DGPulldown, something similar is done.

And the 3rd way to do this conversion is just to blend the hell out of it to go from 24 to 25fps.

Now, when doing the reverse, going from 25 to 24fps for NTSC DVD, the usual procedure is to slow down the video and audio to 24fps and then telecine off of that. Either that or you go back to the original 24fps film elements and create an NTSC master from scratch. However, very often a PAL master is used for the conversion to NTSC. It can be telecined for 25->29.97fps (avoiding the blending), but in my experience, that is almost never done for NTSC DVD. To get to 29.97fps extra frames/fields are often generated by blending the hell out of it.

I don't personally know that a PAL master was used for the Sci-Fi channel Dr. Who shows. I don't capture anything. I'll take jagabo's word for it, though. And it's easy to confirm. Separate the fields, find a place where there's movement, advance a frame/field at a time and look for the blending. When you see blending of the fields, the game is up, and you know you have a bad standards conversion on your hands.

Assuming that's what you find, you can't use TMPGEnc or AviSynth IVTCs (Decomb, TIVTC, etc.) for "IVTCing" it. You have to use an unblender. Otherwise the blends remain, plus you get jerky playback. Some people like Restore24. Personally I think RePAL is the best for removing the blends from PAL sourced video. It'll give you 24.975fps. If for DVD, you can then use DGPulldown set for 24.975->29.97fps, or you can slow both the video and audio to 23.976fps and apply regular pulldown. RePAL does leave behind some blending, particularly after a scene change if the pattern has changed, until it locks onto the new pattern. The sharper and clearer the source, the better the job it will do. Didn't you say you have it on video tape. Then maybe it'll do well, and maybe not. The main thing though, is smooth playback.

AviSynth is the only way to accomplish this. This is my specialty. I have 2 blended PAL2NTSC DVDs on my hard drive at the moment. Both are unblending just fine. Let us know if you need help in setting it up.

By the way, I do know of another way to unblend such material, and the results can be perfect. But it takes an incredible amount of time to accomplish, and is a bit complex to describe. Search for UnblendPattern over at Doom9 in the AviSynth Forum. It does something like you were fumbling for earlier, uses the pattern of blended and unblended fields to reconstruct the video. The problem is that these patterns change all the time.

[jagabo]

Just out of curiosity (I'd already seen the show via PAL satellite rips) I recorded (with my Hauppauge PVR-250) the new Doctor Who series on the first night of it's broadcast a few weeks ago on the SciFi channel. I saw that they use that half pulldown, half blend, frame rate conversion that a lot of PAL material goes through. Another show I know that does this is The Tribe on eWam.

I don't know of any perfect way to inverse telecine this type of material because many of the fields have blending from the other field. The best results I've seen so far has been with AVISynth and the RePal filter.

I have see a few PAL movies with a perfect 3:2:3:2:2 pulldown (the same thing DGPulldown does) on digital cable channels. These were easy to IVTC (I used TMPGEnc) back to 25 fps progessive).

[manono]

Yeah, those PAL movies that have had a real telecine are easy to get back to 24.975fps. You say that this is done for some TV shows? Because you almost never see it like that on DVD. Rather than do it using TMPGEnc, though, you might try an AviSynth IVTC next time. You'll have to properly tweak the one you choose, but:

Telecide().Decimate(6), #using Decomb, or
SmartDecimate(25,60,B) #using SmartDecimate, or
TFM().TDecimate(Mode=0,CycleR=1,Cycle=6) #or some such, using TIVTC