HD 10 for Film Transfers?

[Maheel Perera]

I am a film maker in Sri Lanka. I shoot in DVCAM and transfer the edited material to 35mm film.

Is HD 10 suitable for film transfers.?

Although this has been mentioned in some of the posts briefly I wanted to add a fresh thread to get more related information.

[Eric Bilodeau]

It might. The biggest actual concern is the 30p issue. I am preparing tests for a feature and we will do blowups of these tests in about a month. I will test software 24p renderings as well as direct 30p footage transfers. Who knows but it would be great to be able to transfer some footage.

[Maheel Perera]

Nice to have a reply within 6 minutes.

Thanks.

In DVCAM the biggest drawback was the long shots. Due to the low resolution of the video image long shots were always reproduce as a disturbing out of foucus image. that is where I hope the HD10 will score.

Since all the film work is done under controlled lighting with people who know the craft of lighting some of the problems mentioned regarding the HD10 will not arise..

[Brian Mitchell Warshawsky]

OK, crazy idea ...

Given the high cost of DV to Film transfers, the prosumer HD’s (current) lack of 24P, and knowing that "most" DV feature efforts will never enjoy a celluloid existence, I couldn’t help but notice that at least one DV film (non 24P) listed in the IMDB was transferred the “cheap way”, by non-self-consciously shooting off a monitor. The original footage was not HD, so the results were probably tough on the eyes.

Now, assuming perhaps the DV feature is shot in HD Black and White to avoid color shifts, and assuming a 16:9 plasma screen of the highest order is used, and selecting just the right BW film stock negative, and a precise camera with the highest quality optics shooting 24fps, might this not produce a workable 24FPS film print that would at least stand up to the scrutiny of a film festival? Film festivals have been known to be much more forgiving than regular paying audiences. El Mariachi won the audience prize at Sundance with a poor color print that is a far cry from the color-corrected print eventually released. In the above scenario, the black and white nature of the medium may allow the final print to be very forgiving.

For purposes of studying motion on film, I used to shoot Super 8 film off of a video monitor, and when conditions were perfect, the results were surprising. As crazy an option as this may at first seem, consider that the HD resolution coupled with the new plasma screens may truly be optically better than the previous generation’s professional film transfer equipment.

Much of the expense inherent in current DV to film transfers is in the specialized equipment and the slow render time (at most only 3 frames per second). The transfer time in the above scenario would be the running time of the feature. If this were a workable solution, the final result would be 24FPS, and the cost (still not insignificant, but much less) would be only the cost of the stock and development, plus the answer print.

Are my assumptions unrealistic?

Brian

[Eric Bilodeau]

I know of someone who did a test shooting on a plasma with a 35mm camera and it produced really interresting results. You need a very good plasma though. The test was from a colored source, I don't know about BW but it must be better.

I don't think the idea is crazy at all but you have to consider what are the results you are looking for. It is all a question of perspective. Someone wanting a very sharp and clean image might not try this but someone willing to accept the look it will produce might save a lot of money that way...

Very interresting indeed...

[Maheel Perera]

I have build a film recorder. It is a Kinescope using a high resolution CRT and it works. I have done a few transfers for theatrical distribution in Sri Lanka. We cannot afford 50,000$ transfer since we would need only 24 copies maxium to cover the entire island.

The problem was with the long shots. That is why I thought the HD10 would solve that problem. seems that it adds several too. I have written to several Film Transfer facilities in the USA and almost all have said that the conversion from 30P to 24P will induce motion judder. I have yet to receive a reply from Digital Film Group.

However RevisionFX were confident that the 30P to 24P conversion could be done sucessfully using their Twixtor.

http://www.revisionfx.com/
jasmin@revisionfx.com

Eric , If you could email me your address I could send you several frames of your sample footage transfered to 35mm slides (free of charge ,of course)

Maheel

[Brian Mitchell Warshawsky]

Maheel,

What are the additional problems created by the JVC?

Are you able to transfer 1 foot of an action or motion sequence with your system as a test to determine the extent of any judder?

Is your transfer frame-by-frame or does it work in real time, filming at 24 fps the video output? Wouldn’t the second approach result without judder but with the sacrifice being some varying loss of resolution?

Perhaps the judder referred to by the film transfer companies results from possibly dropping frames, or other incompatibilities with their transfer methodologies which were probably designed with 60i in mind, and have not been properly reworked for 30p.

I know a number of people on this board have been intrigued by what Twixtor offers, and it would be intriguing to compare a short transfer made using your system and a short sequence rendered with Twixtor.

If I understand correctly how Twixtor may function, every frame of the final film might essentially be a digitally modified morph (someone correct me if I am incorrect).

If the above is correct, there would seem to be 3 questions which would need to be answered:

1. How long does Twixtor take to render;

2, Does Twixtor result in a loss of resolution;

3. How natural is the appearance of motion within the final result?

Very much intrigued,

Brian

[Marco Leavitt]

I’m intrigued by the notion of using a plasma screen to make a transfer. Rental costs for the 35mm camera and plasma screen aside, can anyone hazard a guess on how much it would cost for film and developing to create a 35mm print? Under $10,000?

[Eric Bilodeau]

Be caucious, most plasmas do not produce a great image at all and induce a lot of artefacts. A good wide CRT television with HDTV capabilities should produce a better image or even a HDTV CRT.

As for twixtor, it is not fast to render (it is a filter after all), it does not induce la loss in the resolution, it blends frames in a morph like way. It looks nice even with a lot of motion in frame on computer but I do not know how it looks transfered on film. I believe it is an interresting possibility but it takes some time.

[Maheel Perera]

Brian,

For me some of the additional problems are,

1. Narrow latitude

2. Exposure Control (combined with the narrow latitude)

3. Useless focusing ring (Common to most DV cams.)

4. Of course, No 25P (this should be the no 1 problem of my list)


The most talk about point which is the 36 lux rating does not bother me.

At aperture 1.6 a 800 ASA film would need about 4 foot candles. (refer American Cinematographer Manual) 1 fc = 10.76 lux

I stll not decided to buy the HD10. I need it for film transfer work. Firstl I must find a solution to this 30P to 24P issue. So I do not have any materials to transfer. For transfer I have to convert the footage to a sill image sequence. Since I could not work with these m2t files it is not possible.

Re Twixtor I can't add much to Eric's answer. There is a trial version at their site. Try it out for yourself.

[Emilio Le Roux]

The problem with transfering video to 24 fps film is common to most DV camcorders and not only from the HD10.

The resulting 24 fps eliminates every 5th frame from the original 30p footage (6x5=30, 6x4=24) so, if a subject or the camera is in motion, we see a little jump that happens 6 times every second. It is somewhat unconfortable but I can't measure that.

Supposedly most commercials that exhibit at the theathers that came from video cameras have this problem.

If motion is priority over definition, you can get a 24p DV camera. They're being used for film features already.


The SD mode from the HD10 is interesting too. It delivers DV definition but with 60 progressive frames per second.
When converting to 24fps the jump in motion occurs, but it more subtle.

I am not sure if traditional DV 30i footage is better than 30p for transcribe to film.

[Maheel Perera]

The problem (according to many film transfer companies) is the 30P frames. In DV format we get 60i and when it was converted to 24fps they use field blending to minimise the jerkyness. In HDV what we get is 30P.

The need is for more resolution but again we have to find a solution to minimise this jerkyness. Ofcourse, when we have the 25P version this problem will vanish. But how long will we have to wait.?

[Emilio Le Roux]

have you tried the TWIXTOR plug-in yet?

[Barry Green]

<<<-- Originally posted by Emilio Le Roux : The problem with transfering video to 24 fps film is common to most DV camcorders and not only from the HD10. -->>>

Well, no, most DV camcorders shoot 60i, not 30P. 60i can be converted to 24fps without too much difficulty, but 30P cannot, so the problem is primarily with the HD10, HD1, Canon XL1 or other camera in "frame mode", etc.

<<<--
The SD mode from the HD10 is interesting too. It delivers DV definition but with 60 progressive frames per second.
When converting to 24fps the jump in motion occurs, but it more subtle.
-->>>
Actually this mode is quite interesting. It provides for a fairly accurate simulation of 24P, as well as providing exquisite detail for slow-motion shots. But it's only standard-def resolution. As far as motion rendering for film transfer it would be far superior to the HD mode on the HD10, but of course horizontal and vertical resolution are far superior in the HD mode.

<<<--I am not sure if traditional DV 30i footage is better than 30p for transcribe to film. -->>>
It is infinitely superior. Interlaced 60i footage can be converted by most every film transfer house out there, using a reverse 3:2 pulldown scheme. Transfer houses have been converting 60i down to 24fps for years and years, and have gotten quite good at it.

Getting a decent quality transfer from 30P however, is a problem that is yet to be solved. The transfers I've seen were terrible in motion rendering.

The promised next generation of HDV-compliant camcorders should theoretically be much more suitable for the film transfer process, since the HDV standard includes 720/25P and 1080/50i or 1080/60i, and any of those will work much better for transfer to 24fps film.

[David Newman]

<<<-- Originally posted by Emilio Le Roux : have you tried the TWIXTOR plug-in yet? -->>>

Yes, and it does work very well for most scenes -- I was surprized. However it takes forever to do the conversion. But I agree 30p to 24p is the most difficult, and therefore subject to motion artifacts with some types of complex motion.