Super 8 film, eliminating "pull down" effect

[szilagyic]

Hello:

I'm trying to capture some Super 8 film with sound, by using a projector on a white screen and using a video camera. The projector is running at 18 fps, and the lowest I can set my shutter speed on the camera is 24 fps. The results are fairly good except here and there, there are dark waves that go through the picture. They come and go throughout the film. I know this is because of the mismatched frame rates of the projector and camera.

My question is, what is the best (and hopefully simplest way) to eliminate these dark waves? I've tried using VirtualDub and using the various frame rate settings, including the "3:2 inverse telecine" options, however I can't get the dark waves to disappear. I've also been trying various AVISynth filters. I'm looking for any suggestions on what program(s)/filters should do this.

Thank you very much. I know this is a common subject, but I'm a newbie with capturing film and just need to be pointed in the right direction with this. Thanks!!!

[16mmJunkie]

Hello:

I'm trying to capture some Super 8 film with sound, by using a projector on a white screen and using a video camera. The projector is running at 18 fps, and the lowest I can set my shutter speed on the camera is 24 fps. The results are fairly good except here and there, there are dark waves that go through the picture. They come and go throughout the film. I know this is because of the mismatched frame rates of the projector and camera.

My question is, what is the best (and hopefully simplest way) to eliminate these dark waves? I've tried using VirtualDub and using the various frame rate settings, including the "3:2 inverse telecine" options, however I can't get the dark waves to disappear. I've also been trying various AVISynth filters. I'm looking for any suggestions on what program(s)/filters should do this.

Thank you very much. I know this is a common subject, but I'm a newbie with capturing film and just need to be pointed in the right direction with this. Thanks!!!

To eliminate the screen flicker the projector must be equipped with the right kind of shutter. Secondly, it must be operated at the correct speed. At 15fps the projector must have a 4-blade shutter. At 20fps the projector must have a 3-blade shutter. (Most 8mm projectors use a 3-blade shutter.) At 24fps the projector must have a 5-blade shutter. The reason for these factors is that they must of course be compatible with the TV rate of 30fps. Examples: A 4-blade shutter rotating at 15 rpm equals 60 (4x15), a 3-blade shutter operating at 20 rpm equals 60 (3x20), and a 5-blade shutter operating at 24 rpm equals 120 (5x24).

[szilagyic]

Thank you for the explanation. Currently the projector I have has two modes, selected by a switch, 18 fps or 24 fps. I believe it is a 3-blade model (just had it apart to clean it recently). I guess I was hoping there was a software solution that would help. This seems to be a common question asked, and I've read on all of the filters and such. Is there any way I can fix the captured video, or are you saying the ONLY way is to look at the hardware? Thanks!!

[Khaver]

If you can remove the shutter in the projector, you can use your camcorder to record the image on the screen (use 60i or 60p if you have it). You will get several still frames and some blurred frames where the film is being pulled down to the gate. Using avisynth with some plugins and scripts you should be able to select one frame from each group of good frames.

[filmboss80]

16mmJunkie (I'm one too) explained it quite well. Basically, the timing of the camcorder shutter is not in sync with the projector shutter; thus you get dark frames when the projector shutter is closed and bright frames when it is open. This is a flicker effect; not solely an issue of repeated telecine frames that can be solved through software solutions. In essence, you have frames at widely varying exposure levels. That's something not easily fixed. You have 5 choices, depending on your budget and availability of equipment:

1. Buy a cheap, variable speed Super 8 projector on Ebay and adjust the speed to 20fps.
2. Buy a not-so-cheap Super 8 telecine projector (with 5-blade shutter) on Ebay.
3. Capture to a video camera that records 24p (I am not real sure how well this would work.)
4. Install a 5-blade shutter on your projector (not an easy or practical FYI task).
5. Pay a transfer house to do it professionally.

An afterthought: You do not list your location, so I posted these options based on the assumption that you are in an NTSC country. If not, then everything I wrote will be off base.

[DB83]

Bit confused here as I thought that Super 8 with sound was projected at 24 fps ?????

Silent Super 8 is 18 fps.

[szilagyic]

Thank you for all of the feedback. Yes I am in the U.S. so I'm NTSC.

I already had the video camera set to 24 fps (it's a Panasonic HDC-TM700 that has a special "cinema" mode) which helped quite a bit. So the waves I mentioned are not severe, but noticeable here and there depending on the lighting and timing. I may look for an adjustable projector, the issue though is the film has sound but I am guessing that going from 18 fps to 20 fps, that the sound would still be OK.

Khaver: I have been reading and experimenting with AVISynth (still a novice). I was thinking of doing exactly like you said by using AVISynth to select the "good" frames and somehow reconstruct the video. Do you have the names of any plugins that are known to do this?? I have been experimenting with TIVTC but so far I haven't had much luck.

DB83: I have a commercial reel that's 24 fps with sound, but the home movies I am converting were recorded at 18 fps with sound.

I'll be sure to post any positive results if I find them. Thanks!

[filmboss80]

When altering the speed of a Super 8 sound film, there will be a pitch change in the audio. This is easily remedied with Audacity or other sound editor. I have transferred some 24fps sound films to 20fps, and after fixing the audio pitch, no viewer (other than myself) seemed to notice the slightly slower playback speed. Since many of your films are 18fps, the slight speed up may not be too bad, but you will still have to fix the audio pitch change.

[Khaver]

Khaver: I have been reading and experimenting with AVISynth (still a novice). I was thinking of doing exactly like you said by using AVISynth to select the "good" frames and somehow reconstruct the video. Do you have the names of any plugins that are known to do this?? I have been experimenting with TIVTC but so far I haven't had much luck.

I've got a couple of old projectors that from one I might be able to remove the shutter. I'll then try this method too and post back here with results. This would remedy the flicker problem if we can get it to work.

Read THIS thread.

[filmboss80]

Khaver: I have been reading and experimenting with AVISynth (still a novice). I was thinking of doing exactly like you said by using AVISynth to select the "good" frames and somehow reconstruct the video. Do you have the names of any plugins that are known to do this?? I have been experimenting with TIVTC but so far I haven't had much luck.

I've got a couple of old projectors that from one I might be able to remove the shutter. I'll then try this method too and post back here with results. This would remedy the flicker problem if we can get it to work.

Read THIS thread.

While the thread you linked to is interesting, it seems to me like an extremely convoluted, "Rube Goldberg" telecine method. Perhaps it's doable, but doesn't it require a frame-by-frame analysis, followed by the process of manually discarding garbage frames? Also, it seems like this method could only work with either DV or lossless compression codecs that digitize to individual video frames, as opposed to GOPs (Groups of Pictures), ruling out Mpeg, Divx, HDV, AVCHD, or other such formats.

Getting that method to work would require enormous amounts of time, it seems to me. Nevertheless, it's an interesting concept. Thanks for posting the link.

[Khaver]

It's not really so "Rube Goldbergish". Shooting the film image at 16-18-24fps without a shutter with a camcorder set to 60i and a fairly high shutter speed will give you 3 types of frames.

1 - progressive frame (both fields capture a single frame)
2 - a combination of 1 field capturing a frame and the other field capturing the frame while it's being pulled down)
3 - both fields capturing the frame as it's being pulled down (each field will be different though)

Now each of these may have several of them in a row. Get avisynth to only keep the progressive frames and from these, throw away the duplicates. I think it can be done.

[filmboss80]

It's not really so "Rube Goldbergish". Shooting the film image at 16-18-24fps without a shutter with a camcorder set to 60i and a fairly high shutter speed will give you 3 types of frames.

1 - progressive frame (both fields capture a single frame)
2 - a combination of 1 field capturing a frame and the other field capturing the frame while it's being pulled down)
3 - both fields capturing the frame as it's being pulled down (each field will be different though)

Now each of these may have several of them in a row. Get avisynth to only keep the progressive frames and from these, throw away the duplicates. I think it can be done.

Sorry, but until you actually execute the theory, I will remain extremely skeptical. There is a reason why projectors have shutters. Perhaps shooting a shutterless projection with a camcorder's shutter set at 1/1000 speed will catch good frames; but there will also be loads of misaligned garbage frames in between; and it will take much time to analyze and remove them. (At 1/1000th of a second, it would seem that projector bulb flicker would also enter the equation.) As I read the Doom9 forum link you provided, the OP there seemed to gradually back away from the idea that his theory, put in practice, produced efficient results. Doable, perhaps, but not worth the hoops one would jump through in comparison to other advised methods, such as either a variable-speed or telecine projector. (There are also very good threads in this forum about DIY telecine methods. I recommend perusing through these.)

I'm interested in hearing your real-world results in actual execution -- and I don't mean just a short clip. And the resulting footage will need to be synchronized with the sound.

[16mmJunkie]

I agree with filmboss80. Also with what the OP stated about experience(being a newbie) I would say the learning curve required with this "apparent method" would be very tough. The link provided had no clips and when asked at the end for results no follow-up since 2009.

[2Bdecided]

Sorry, but until you actually execute the theory, I will remain extremely skeptical. There is a reason why projectors have shutters. Perhaps shooting a shutterless projection with a camcorder's shutter set at 1/1000 speed will catch good frames; but there will also be loads of misaligned garbage frames in between; and it will take much time to analyze and remove them. (At 1/1000th of a second, it would seem that projector bulb flicker would also enter the equation.) As I read the Doom9 forum link you provided, the OP there seemed to gradually back away from the idea that his theory, put in practice, produced efficient results.

You didn't read to the point where he'd done over 100 reels with it then?

http://forum.doom9.org/showthread.php?p=1005949#post1005949

Cheers,
David.

[2Bdecided]

I already had the video camera set to 24 fps (it's a Panasonic HDC-TM700 that has a special "cinema" mode) which helped quite a bit. So the waves I mentioned are not severe, but noticeable here and there depending on the lighting and timing. I may look for an adjustable projector, the issue though is the film has sound but I am guessing that going from 18 fps to 20 fps, that the sound would still be OK.

Be aware that the slower the shutter speed, the more blended frames you're going to get. There might be an AVIsynth trick to remove these - I'm not sure. Blending frames isn't a necessarily bad way of converting from 18fps to 24fps anyway.

Cheers,
David.

[filmboss80]

You didn't read to the point where he'd done over 100 reels with it then?

http://forum.doom9.org/showthread.php?p=1005949#post1005949

Cheers,
David.

Of course I did. Perhaps you didn't read the point I made:

Doable, perhaps, but not worth the hoops one would jump through in comparison to other advised methods, such as either a variable-speed or telecine projector.

Also note that there aren't any reflections of the amount of time and effort involved. (However, in post #44 of that link, he said: "I don't have any film with frame numbers on it, but i have spent dozens and dozens of hours going over film, frame by frame, looking for skipped or duplicated frames.") Nor are there any corroborating experiences from others who may have attempted that method.

The clip provided was silent, not synchronized to any sound.

Feel free to try it yourself and post the results.

[2Bdecided]

Feel free to try it yourself and post the results.

I can't figure out what he's doing in the spreadsheets he talks about.

I think the time taken was to get the method right. Having done it, it's now semi-automatic and quick AFAICT.


Other options may be easier, but I'm looking for a frame-accurate transfer. All the "cheap" methods of doing this seem fraught with problems! For now, I'm concentrating on formats that are less likely to be recoverable in a few years. When I've done those, I'll turn my attention to film!

Cheers,
David.

[Khaver]

I am having some success with this shutterless method, but for the best results, I used the "Slow motion" setting on the projector. This runs at about 4fps. This allowed me to capture more progressive video frames of each single frame of film. I then used an "IsCombed" plugin to write out a text file in the first pass that had the frame number and "true" or "false" on each line, true when a frame was combed (film moving) and false when not combed (film still in gate). The text file had a few trues followed by a group of about 4 or 5 falses. I then wrote a program that would process the text file and write out a new file that chose the middle frame of the group of "falses" (the best progressive frame) and using the remapframes plugin in the second pass was able to get a new video with only clean, progressive frames. I'll post results soon.

The projector was just an old kodak where I cut the fins off the shutter. Then used an HD camcorder (Canon HV30) to capture the image projected on white paper.

[filmboss80]

I am having some success with this shutterless method, but for the best results, I used the "Slow motion" setting on the projector. This runs at about 4fps. This allowed me to capture more progressive video frames of each single frame of film. I then used an "IsCombed" plugin to write out a text file in the first pass that had the frame number and "true" or "false" on each line, true when a frame was combed (film moving) and false when not combed (film still in gate). The text file had a few trues followed by a group of about 4 or 5 falses. I then wrote a program that would process the text file and write out a new file that chose the middle frame of the group of "falses" (the best progressive frame) and using the remapframes plugin in the second pass was able to get a new video with only clean, progressive frames. I'll post results soon.

The projector was just an old kodak where I cut the fins off the shutter. Then used an HD camcorder (Canon HV30) to capture the image projected on white paper.

You went through all that trouble, only to shoot the projected image off a white card?

Keep in mind that film projectors do not put out interlaced images, any more than a lightbulb does. That is a characteristic of the camcorder alone, and if you have an HV30, you can always shoot progressive.

Telecine has been around since the early 1950s, when North American television broadcasters tackled the challenge of putting 24fps theatrical movies on their 30fps medium without altering the original film speed. It worked well. We North Americans have spent most of our lives watching 2:3 pulldown movies and television series without ever noticing that some frames were blended and repeated. It is a well-perfected, smooth-motion technology. It is not something that absolutely has to be eliminated.

As of the day and hour of this posting, there are a couple of GAF Super 8 sound projectors on Ebay, with starting bids ranging from $50 to $100. Those units have variable speed controls to get out the flicker during realtime film-to-video transfers. That is just one alternative.

I am not trying to bullishly slam the telecine techniques others have employed. I am only trying to indicate to the uninitiated that there are alternatives that will save enormous amounts of time in accomplishing the task. It is alarming to see the largely unnecessary lengths to which some people will go, just to get their films on video.

[Khaver]

I'm still in the experimental stage. I took a $10 projector I got at Goodwill, cut the fins off the shutter, pointed my camcorder at the projected image and ended up with usable (for an experiment) video.

This method requires to shoot in 60i and not progressive. The script differentiates between progressive frames (when both fields capture a film frame at rest, the IsCombed plugin will output "false") and interlaced frames (when the film is being pulled down, each field will be slightly different and the IsCombed plugin will output "true").

Here's the avisynth script for the first pass:

LoadPlugin("C:\video\dgindex\bin\DGDecode.dll")
Mpeg2Source("G:\\Slow-0100.d2v", UpConv=0)
crop(200,48,-268,-60)
colon = " "
WriteFile("d:\combed.txt", "current_frame", "colon", "IsCombedTIVTC(metric=1,cthresh=20,MI=20)", false)

Here's the script for the second pass:

Mpeg2Source("G:\\Slow-0100.d2v", UpConv=0)
crop(200,48,-268,-60)
remapframes("d:\remapped.txt")
assumefps(16).lanczosresize(640,480)

Here's what the combed.txt looks like:

0 true
1 false
2 true
3 true
4 true
5 true
6 false
7 false
8 false
9 false
10 true
11 false
12 false
13 false
14 false
15 true
16 false
17 false
18 false
19 false
20 true
21 false
22 false
23 false
24 false
25 true
26 false
27 false
28 false
29 false
30 true

And here's what the remapped.txt file looks like:

0 1
1 8
2 13
3 18
4 23
5 28
6 33
7 38
8 43
9 48
10 53
11 58
12 63
13 69
14 74
15 79
16 84
17 89
18 94
19 99
20 104
21 109
22 114
23 119
24 124
25 129
26 134
27 139
28 145
29 150
30 155

More experiments to follow.

[deadrats]

@szilagyic

there's something i can't seem to understand, first you say:

I'm trying to capture some Super 8 film with sound, by using a projector on a white screen and using a video camera. The projector is running at 18 fps, and the lowest I can set my shutter speed on the camera is 24 fps. The results are fairly good except here and there, there are dark waves that go through the picture. They come and go throughout the film. I know this is because of the mismatched frame rates of the projector and camera.

then you say:

Currently the projector I have has two modes, selected by a switch, 18 fps or 24 fps

anything glaring stand out to you?

[2Bdecided]

Telecine has been around since the early 1950s, when North American television broadcasters tackled the challenge of putting 24fps theatrical movies on their 30fps medium without altering the original film speed. It worked well. We North Americans have spent most of our lives watching 2:3 pulldown movies and television series without ever noticing that some frames were blended and repeated. It is a well-perfected, smooth-motion technology.

It's not, and never has been, a smooth motion technology - there's a 12Hz judder from the 3-2 pattern.

More importantly, given that analogue video delivers a sequence of fields, there is absolutely no blending (assuming the film is synchronised to the video - normal these days, but not universal in the early days). You get 3 clean fields from one film frame, followed by 2 clean fields from the next film frame. No blending.

It is not something that absolutely has to be eliminated.

True, film > hard pulldown 3-2 interlaced video is fine, though really suboptimal for MPEG encoding (you need a far higher bitrate to get the same perceived quality, compared with native 24p + flags). Well, it's fine if you want to watch at 60fps.

However, that's not what you get by pointing a camcorder at a projector. This always gives blended fields. This in turn means the picture looks blurred in motion. It looks worse still on a PC, and even worse again if you want to upload it to the web.

The best you can get is to capture each film frame exactly once as one frame. No blends. No repeats. Perfect for uploading to the web. Perfect for viewing on a PC. Ideal for MPEG encoding. Perfect for authoring to DVD (your encoding or authoring software will add pulldown flags. If not, there's always dgpulldown).

I am not trying to bullishly slam the telecine techniques others have employed. I am only trying to indicate to the uninitiated that there are alternatives that will save enormous amounts of time in accomplishing the task. It is alarming to see the largely unnecessary lengths to which some people will go, just to get their films on video.

You can, of course, do it very quickly and cheaply. Many people will be happy with the result. But it's a long long way from the best that can be done.

Cheers,
David.

[filmboss80]

Here in North America, viewers have watch filmed movies and television series with 3:2 pulldown since the early 1950s without noticing anything off kilter. In fact, every single VHS release of a theatrical movie in North America had it. At that speed, there was no perceptible judder or motion blur--although technically, there was plenty of both, if you broke things down frame by frame (but that's not the way most people watch). That is what I meant by smooth motion technology. I am talking about the practical, real-world viewing experience, not the white paper engineering data. It worked fine, and it worked for more than a half century.

(Arguably, someone viewing an American film-to-video transfer in the UK might indeed see a judder effect, but that could be attributed to the combination of pulldown and the NTSC-to-PAL conversion.)

The OP was dealing with the flicker effect, misinterpreting it as a pulldown effect. Some of us have tried to steer him toward a more practical means of transferring old home movies to video.

[szilagyic]

deadrats: Yes I did try to set the projector at 24 fps and the video camera also at 24 fps, however the flickering was actually worse! I am guessing because the projector is not super precise (just a motor that determines fps). Plus like it was pointed out, we would have to deal with the higher pitch sound and adjust that (which I know can be done with software). Running the projector at 18 fps setting, and the video camera at 24 fps has been good, and the flickering is very minimal. In fact, I noticed that when I run the projector backwards in reverse, the flicker is gone. But it does not play the sound this way. Argh, almost there.

Thanks for all of the feedback with the information, and interesting tests. I have also done some testing so far by not removing a shutter but using the projector as-is, and trying to optimize with software. So far the best image quality results have been deinterlacing the video in VirtualDub, then applying the motion blur filter and temporal smoother filter set to around the middle setting, then the sharpen filter maxed out at the far right. I figured that motion blur would blend in the frames better. In fact, the deflicker filters in VirtualDub (by Donald Graft) do exactly this because you can see a trailing effect if you set the filter too high. What I've found is that the image quality is excellent for low motion, but when there is high motion in the film (fast panning for example), there is a trailing effect and the quality is not so hot. It's livable, but I know it can look better.

I think I'm going to pursue looking at a projector with variable speed, and use that to eliminate the flicker from the source video directly by running the video camera at 24 fps. Then use software to adjust the pitch of the sound. I've put in a lot of hours trying to use software, and the results are so-so. I'll also post back any positive results! Thanks again very much.

[2Bdecided]

Here in North America, viewers have watch filmed movies and television series with 3:2 pulldown since the early 1950s without noticing anything off kilter. In fact, every single VHS release of a theatrical movie in North America had it. At that speed, there was no perceptible judder or motion blur--although technically, there was plenty of both, if you broke things down frame by frame (but that's not the way most people watch). That is what I meant by smooth motion technology. I am talking about the practical, real-world viewing experience, not the white paper engineering data. It worked fine, and it worked for more than a half century.

(Arguably, someone viewing an American film-to-video transfer in the UK might indeed see a judder effect, but that could be attributed to the combination of pulldown and the NTSC-to-PAL conversion.)

No, just watching an NTSC film transfer in the UK (all modern European TVs display 60Hz just fine), or for that matter getting on a plane and going to the US and watching your TV, makes the judder horribly apparent.

The reason you don't see it is because you've grown up with it and are used to it. To most "PAL" viewers, the stutter of an "NTSC" film transfer looks like a fault. Stop-Start. 12 times a second.

(The PAL speedup, without pitch correction, sounds like a fault to anyone either side of the Atlantic who has an ear for it - but at least our film>video transfers have no more judder than the original film itself did ).


I do agree that the OP is probably looking for an easy (albeit low quality) solution though.

Cheers,
David.

[filmboss80]

(Arguably, someone viewing an American film-to-video transfer in the UK might indeed see a judder effect, but that could be attributed to the combination of pulldown and the NTSC-to-PAL conversion.)

No, just watching an NTSC film transfer in the UK (all modern European TVs display 60Hz just fine), or for that matter getting on a plane and going to the US and watching your TV, makes the judder horribly apparent.

The reason you don't see it is because you've grown up with it and are used to it. To most "PAL" viewers, the stutter of an "NTSC" film transfer looks like a fault. Stop-Start. 12 times a second.

Not buying the argument. Proper telecine frame blending (from a 24fps source) inherently creates, at worse, motion blur; not a stop-start effect. Exceptions, admittedly, would apply to 16fps (Regular 8mm) and 18fps (some Super8) sources, as there would be more duplicated frames.

If anyone should be able to notice judder, it should be those of us accustomed to 30fps NTSC television, more so than those growing up with 25fps PAL television.

[2Bdecided]

Not buying the argument. Proper telecine frame blending (from a 24fps source) inherently creates, at worse, motion blur; not a stop-start effect.

3 fields followed by 2 fields has no blending and 12Hz judder (stop-start).
(I'm talking about what TV stations use, not pointing a camcorder at a projector.)

If anyone should be able to notice judder, it should be those of us accustomed to 30fps NTSC television, more so than those growing up with 25fps PAL television.

The motion difference between native 50i and native 60i is marginal. People notice the flicker on 50i CRTs, and the lower resolution of SD 60i CRTs - but not so much the slightly different motion portrayal. It's there, but it's not that obvious at all.

Cheers,
David.

[filmboss80]

3 fields followed by 2 fields has no blending and 12Hz judder (stop-start).
(I'm talking about what TV stations use, not pointing a camcorder at a projector.)

Once again, I am not disagreeing with you about the technical aspect. I am talking about what viewers perceive. And to be clear, I do not aim camcorders at projectors. I have utilized telecine systems--from Buhl optical multiplexer setups to Rank Cintel systems--for more years than I care to reveal. When I am speaking about telecine blending, I should clarify that I am not referring to the projector output, but rather the frames (combined fields) as they are captured by a conventional 29.97 NTSC imaging system.

The rules are different nowadays, and I don't hold fastidiously to the old pulldown methods myself. In this age of digital media, it is no longer necessary (or preferred) to display 24fps films at 29.97fps. But in the days of analog TV it was necessary and made to work.

When you saw American TV and airline-displayed US TV episodes, it is very likely you saw judder that resulted from time compression, not telecine pulldown. That is extremely common on shows seen on cable networks like TVLand, as well as old reruns on local stations. This is performed to make room for additional commercials. Time compressed shows do have very noticeable stop-start motion sequences, even to the untrained eye.

Adding on (hours later): http://www.projectorcentral.com/judder_24p.htm -- helps clarify much of the confusion about 24p judder in current digital tv systems and the motion issues of traditional 24fps film telecine transfers (2 separate things altogether).

[2Bdecided]

@ filmboss80 - I've stayed in the USA for weeks. Here in the UK, I have a bank of NTSC TVs sitting across the room. I've converted my own content to/from various formats. I have ten movie DVDs in both "PAL" and "NTSC" format. I have pristine test content from various movie studios that is not available to the public. I know what I'm talking about!

I'm also very sensitive to various motion and coding artefacts. But it's not that unusual to find people in the UK who can see something wrong with 3-2 pulldown.

Motion portrayal on LCD displays is a whole different kettle of fish.

But contrary to what your link claims, the reason TVs now have a "true 24p" mode is because BluRay discs are almost always 24p across the entire world, and this forced 3-2 pulldown on the half of the world that had never seen it before (our DVDs were/are almost always 25p). We didn't like the 24-in-60 which early players and displays defaulted to for 24p material (yes, in Europe, the players spat this out in 60Hz - really!). We didn't like it at all. Hence "true 24p".

Cheers,
David.

[filmboss80]

I probably need to point out that the problems with digital 24p are different from the process of 24fps film being transferred to 29.97 video. I have no doubt that you have seen judder. Everyone sees it at times. However, this is almost always the result of the way in which many 60i television systems process 24p digital video. (This is especially annoying when cable or satellite TV stations reprocess signals and lose the flags from the source.)

At the transfer level of 24fps film (not 24p digital video) to 29.97 video, the negative artifact is just the opposite; there is motion blur and "smearing," especially when the image is panned quickly. (Conversely, image pans of 24p digital on a 60i system will have jerky moves.)

It appears that we have been talking about 2 different things. I assure you, I have lost count of all the miles of celluloid acetate film footage I've had transferred to various forms of video, and it is motion blur, not judder, that plagues that form of 3:2 pulldown. (This was what the link I provided was also trying to explain.)

This whole thread was about dealing with a movie film--not digital video--source. The OP was dealing with projector flicker, and wanted to know if inverse telecine methods would alleviate the problem. Somehow, we've allowed this thread to deviate way off course.