Pulldown of 25 fps to NTSC

Xesdeeni · 10th January 2005, 15:24

Answering a question in http://forum.doom9.org/showthread.ph...9&goto=newpost got me thinking. When I convert PAL 25 fps progressive to NTSC, the output is interlaced. But when we encode 24 fps, the encode can be progressive, and the flags handle the telecine interlacing.

So wouldn't it be possible to create a 25 fps progressive stream of 720x480 and use the flags to do the pseudo telecine to 29.97 fps interlaced?

I think the pattern would be 2:2:3:2:3 instead of 2:3. And the drop-frame (1000/1001) could be handled as well, by dropping one of the set of 3 fields every 1001 fields.

Xesdeeni

tyee · 31st January 2005, 19:13

Xesdeeni
Have you tried this yet. I've just tried your script for PAL to NTSC and my first impression is great. It is interlaced but this new method might work and keep it progressive??

tyee

Xesdeeni · 31st January 2005, 21:50

No. I didn't have much time shortly after I posted the above, and my working machine is broken at present. On top of that, I'm not sure how to create the pulldown order above in an MPEG stream.

Have you tried it?

Do you know how to create this flag order?

Xesdeeni

Guest · 31st January 2005, 21:59

Quote:

So wouldn't it be possible to create a 25 fps progressive stream of 720x480 and use the flags to do the pseudo telecine to 29.97 fps interlaced?

Yes, of course. I've seen several clips that do just that.

Xesdeeni · 31st January 2005, 22:02

So it could be encoded progressively, but use the flags to interlace? That would seem to be the best of both worlds. Although there would be a telecine-like hitch, we're used to seeing this. But the encoding would be progressive, and the audio wouldn't have to be touched.

So do you know what s/w will set the flags accordingly?

Xesdeeni

tyee · 1st February 2005, 21:47

Xesdeeni
No, unfortunately I don't know how to do this. I was looking on the web for methods other that 3:2 pulldown but didn't find any. Maybe someone here will have to write a little app to do it or modify 'pulldown.exe' to do it.

What about the possibility that NTSC DVD players will not accept 25p with flags that aren't the normal 3:2 pulldown sequence?

tyee

Xesdeeni · 1st February 2005, 23:03

That's interesting too. Although, I've never figured out why the flags were necessary in the first place. I'm sure I don't understand completely, but it seems to me, if you encode a 24p DVD, the player knows it is outputting 60i, so it can do the translation. By the same token, a PAL player could know it is outputting 50i, and do the translation (and scaling). But since the flags are there, it seems they might be useful for this.

Xesdeeni

tyee · 11th February 2005, 06:30

Xesdeeni
Looks like we're at a dead end here. I thought Donald might give us a tip but not yet. Any luck on your end?

tyee

Guest · 11th February 2005, 06:40

It would not be very hard for a programmer to write a utility to do this kind of pulldown. Do you think it is important enough to justify the effort? It might take half a day to do.

Paulcat · 11th February 2005, 14:43

It might depend on the player whether it will work or not. A friend of mine has several SVCD's which have a framerate of 25 fps which play fine on his cheap-ass dvd player, but won't play correctly on my sony.

Kind of odd to have a specification and to ignore it...

Guest · 11th February 2005, 14:52

Did you check whether those SVCDs actually use pulldown flags?

Xesdeeni · 11th February 2005, 15:26

I assume it's worth the effort to find out. If we can encode 25p instead of 30i, the quality should be better. But it depends on the player, I guess, whether it will work. So it'd probably take a test stream to see how compatible it would be. I'd try myself, but I haven't seen source for pulldown.exe.

Xesdeeni

Guest · 11th February 2005, 16:05

I'll roll a quick and dirty one this weekend.

tyee · 11th February 2005, 16:08

Xesdeeni
Here's the link to the source code for pulldown.exe.

http://www.inwards.com/inwards/?id=36

Maybe it will help.

tyee

tyee · 11th February 2005, 16:10

neuron2
Wow, thanks Donald. We just posted within a couple of minutes of each other. I just left the link for the source code above. Maybe it will help.

tyee

Guest · 11th February 2005, 20:29

Quote:

Originally posted by tyee
neuron2
Wow, thanks Donald. We just posted within a couple of minutes of each other. I just left the link for the source code above. Maybe it will help.

tyee

I'm 4/5 done with it already, but thanks.

tyee · 12th February 2005, 01:00

Great Donald!!

scharfis_brain · 12th February 2005, 16:32

would the other way around be possible, too?
(NTSC-Film on PAL with progressive & rff MPEG)

using the flags to telecine 23.976fps to 50 fields per second?

so we could go around the PAL-Speedup (for our audiophiles

)
at the cost of 2 stutters per second but without the disadvantage of interlaced encoding.

Guest · 12th February 2005, 16:52

Yes, you can. We need a more general tool than pulldown.exe. And a one field stutter is usually tolerable. That's why 3:2 looks OK.

I'm using this problem as a test bed for my new parsing architecture. I decided not to do it quick and dirty.

scharfis_brain · 12th February 2005, 19:27

And a one field stutter is usually tolerable. That's why 3:2 looks OK.

one field stutter is sadly very common with some broadcasters here in germany.

they do weird things with their movies :
sqeezing movie to show more ads using a speedup to 26fps
(using fieldblending)
hiding censoring with slowdown to 24 fps.
(fieldrepetition without blending)

both conversions are being made using a stupid bobber before, of course. It looks very 'nice' on TV, when the image details are jumping every other half second *grrrrrr*

Because of this I think, some (most?!?) persons would never notice the one field stutter.
(Our audiophiles probably will ignore it

)

but I personally would never make usage of a 24p telecine on PAL, because it is far more off-standard and shows this stutter.

Musical Films often were made using 24p telecine in past.
I've seen it on some older PAL-VHS cassetes (eg: Koyaanisqatsi)

Also I do not like 3:2 telecine. The fluidity of the movies is somehow different from the cinema-look.
PAL with speedup looks better to me.

recapitulating, I want to say, that I just was interested whether it is possible to do such pulldown with MPEG.

Maybe I'll give it a try on some musical movies, when you finished programming.

10th January 2005, 15:24	#1 \| Link
Xesdeeni Registered User Join Date: Aug 2002 Posts: 467	Pulldown of 25 fps to NTSC Answering a question in http://forum.doom9.org/showthread.ph...9&goto=newpost got me thinking. When I convert PAL 25 fps progressive to NTSC, the output is interlaced. But when we encode 24 fps, the encode can be progressive, and the flags handle the telecine interlacing. So wouldn't it be possible to create a 25 fps progressive stream of 720x480 and use the flags to do the pseudo telecine to 29.97 fps interlaced? I think the pattern would be 2:2:3:2:3 instead of 2:3. And the drop-frame (1000/1001) could be handled as well, by dropping one of the set of 3 fields every 1001 fields. Xesdeeni

12th February 2005, 16:32	#18 \| Link
scharfis_brain brainless Join Date: Mar 2003 Location: Germany Posts: 3,653	would the other way around be possible, too? (NTSC-Film on PAL with progressive & rff MPEG) using the flags to telecine 23.976fps to 50 fields per second? so we could go around the PAL-Speedup (for our audiophiles ) at the cost of 2 stutters per second but without the disadvantage of interlaced encoding. __________________ Don't forget the 'c'! Don't PM me for technical support, please.

12th February 2005, 19:27	#20 \| Link
scharfis_brain brainless Join Date: Mar 2003 Location: Germany Posts: 3,653	And a one field stutter is usually tolerable. That's why 3:2 looks OK. one field stutter is sadly very common with some broadcasters here in germany. they do weird things with their movies : sqeezing movie to show more ads using a speedup to 26fps (using fieldblending) hiding censoring with slowdown to 24 fps. (fieldrepetition without blending) both conversions are being made using a stupid bobber before, of course. It looks very 'nice' on TV, when the image details are jumping every other half second grrrrrr Because of this I think, some (most?!?) persons would never notice the one field stutter. (Our audiophiles probably will ignore it ) but I personally would never make usage of a 24p telecine on PAL, because it is far more off-standard and shows this stutter. Musical Films often were made using 24p telecine in past. I've seen it on some older PAL-VHS cassetes (eg: Koyaanisqatsi) Also I do not like 3:2 telecine. The fluidity of the movies is somehow different from the cinema-look. PAL with speedup looks better to me. recapitulating, I want to say, that I just was interested whether it is possible to do such pulldown with MPEG. Maybe I'll give it a try on some musical movies, when you finished programming. __________________ Don't forget the 'c'! Don't PM me for technical support, please.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

31st January 2005, 19:13	#2 \| Link
tyee Registered User Join Date: Oct 2001 Posts: 416	Xesdeeni Have you tried this yet. I've just tried your script for PAL to NTSC and my first impression is great. It is interlaced but this new method might work and keep it progressive?? tyee

31st January 2005, 21:50	#3 \| Link
Xesdeeni Registered User Join Date: Aug 2002 Posts: 467	No. I didn't have much time shortly after I posted the above, and my working machine is broken at present. On top of that, I'm not sure how to create the pulldown order above in an MPEG stream. Have you tried it? Do you know how to create this flag order? Xesdeeni

31st January 2005, 22:02	#5 \| Link
Xesdeeni Registered User Join Date: Aug 2002 Posts: 467	So it could be encoded progressively, but use the flags to interlace? That would seem to be the best of both worlds. Although there would be a telecine-like hitch, we're used to seeing this. But the encoding would be progressive, and the audio wouldn't have to be touched. So do you know what s/w will set the flags accordingly? Xesdeeni

1st February 2005, 21:47	#6 \| Link
tyee Registered User Join Date: Oct 2001 Posts: 416	Xesdeeni No, unfortunately I don't know how to do this. I was looking on the web for methods other that 3:2 pulldown but didn't find any. Maybe someone here will have to write a little app to do it or modify 'pulldown.exe' to do it. What about the possibility that NTSC DVD players will not accept 25p with flags that aren't the normal 3:2 pulldown sequence? tyee

1st February 2005, 23:03	#7 \| Link
Xesdeeni Registered User Join Date: Aug 2002 Posts: 467	That's interesting too. Although, I've never figured out why the flags were necessary in the first place. I'm sure I don't understand completely, but it seems to me, if you encode a 24p DVD, the player knows it is outputting 60i, so it can do the translation. By the same token, a PAL player could know it is outputting 50i, and do the translation (and scaling). But since the flags are there, it seems they might be useful for this. Xesdeeni

11th February 2005, 06:30	#8 \| Link
tyee Registered User Join Date: Oct 2001 Posts: 416	Xesdeeni Looks like we're at a dead end here. I thought Donald might give us a tip but not yet. Any luck on your end? tyee

11th February 2005, 06:40	#9 \| Link
Guest Guest Join Date: Jan 2002 Posts: 21,901	It would not be very hard for a programmer to write a utility to do this kind of pulldown. Do you think it is important enough to justify the effort? It might take half a day to do.

11th February 2005, 14:43	#10 \| Link
Paulcat Registered User Join Date: Jan 2005 Location: Great White North Posts: 326	It might depend on the player whether it will work or not. A friend of mine has several SVCD's which have a framerate of 25 fps which play fine on his cheap-ass dvd player, but won't play correctly on my sony. Kind of odd to have a specification and to ignore it...

11th February 2005, 14:52	#11 \| Link
Guest Guest Join Date: Jan 2002 Posts: 21,901	Did you check whether those SVCDs actually use pulldown flags?

11th February 2005, 15:26	#12 \| Link
Xesdeeni Registered User Join Date: Aug 2002 Posts: 467	I assume it's worth the effort to find out. If we can encode 25p instead of 30i, the quality should be better. But it depends on the player, I guess, whether it will work. So it'd probably take a test stream to see how compatible it would be. I'd try myself, but I haven't seen source for pulldown.exe. Xesdeeni

11th February 2005, 16:05	#13 \| Link
Guest Guest Join Date: Jan 2002 Posts: 21,901	I'll roll a quick and dirty one this weekend.

11th February 2005, 16:08	#14 \| Link
tyee Registered User Join Date: Oct 2001 Posts: 416	Xesdeeni Here's the link to the source code for pulldown.exe. http://www.inwards.com/inwards/?id=36 Maybe it will help. tyee

11th February 2005, 16:10	#15 \| Link
tyee Registered User Join Date: Oct 2001 Posts: 416	neuron2 Wow, thanks Donald. We just posted within a couple of minutes of each other. I just left the link for the source code above. Maybe it will help. tyee

12th February 2005, 01:00	#17 \| Link
tyee Registered User Join Date: Oct 2001 Posts: 416	Great Donald!!

12th February 2005, 16:52	#19 \| Link
Guest Guest Join Date: Jan 2002 Posts: 21,901	Yes, you can. We need a more general tool than pulldown.exe. And a one field stutter is usually tolerable. That's why 3:2 looks OK. I'm using this problem as a test bed for my new parsing architecture. I decided not to do it quick and dirty.