Digital to Film vs. Film Negative

[Andrew McCarrick]

Looking to know what the costs comparison would be for a Digtal to Film transfer of 90 minutes, versus a film shot on film and having the negative developed. So basically looking for Transfer costs vs. neg. development.

 

[Dan Hudgins]

DIY vs. LAB

DIY vs. LAB

If you DIY the scan and filmout the costs are low, mostly the print stock and processing, maybe the cost of a sound negative, but sound negatives use both sides so you only need maybe 4000'.

My goal is to bring the cost of making a 88 minute feature to under $2000 total, shoot digital and print direct to Kodak 2383 color print stock, I am close and next year should bring some advances if I make progress.

If you try to cut your negative you WILL see blemishes and dust on the print, digital printing gets rid of most of that issue.

We have done it both ways and are going digital because of the dust mitigation problems with film, even doing the needed manual spotting of the film scans is a cost of sorts.

==

Film to Film printing for an "answer print" from your 35mm negative may be much lower cost than doing a scan and filmout AT A LAB, if you have good control of your exposure you can do a "one light" print to cut the costs more since a so called "work-print" costs about a third of a timed answer print.

If you have a telecine made of the negative and can convert that to 24fps BMP frames (AVI -> BMP in VirtualDUB), then the audio can be edited in my programs for "no cost" and the sound negative for printing made from the WAV using pilot tone, probably WAV -> 35mm Mag -> Optical -> Print. Just be sure to do a head slate on all shots and both a head and tail slate on shots longer than 15 seconds.

Editing would be done in my program using the low res BMP frames to look at as the "work-print" and the negative would be cut using the frame numbers from the head slate in my program's edit list. In other words, you punch the head slate for each shot and include that in the telecine, then you count the punched frame as frame 0 for all the shots, you then cut the negative by seting the syncroniser block to 0 for each shot and count the frames from the punch to splice the negative to conform to your edit (working from tails to reduce the risk of errors since want to cut off the head slate last since that is your ref). That is about the lowest cost way to go shooting film.

==

If you shoot RED then filmout on a Cinevator like CHE #1 and #2 that can be lower cost than doing a filmout to a printing negative.

==

Print stock comes in 2000' rolls so an 88 minute film may cost less than a 90 minute film since the print will be just 4 rolls of print stock.

 

[Andrew McCarrick]

Well what I was looking at doing was either a lab digital to film transfer... or taking a Arri 435 (or something similar), and shooting at a 32-50" HD Plasma, and then doing the negative processing at a lab. So I'm wondering if the Neg. processing would be cheaper then Digital to Film Transfer. I can get the camera and a Ultra Prime lens for $1600 for a day, plus film stock looks like it'll be about $5200.

Judging from the price on DVFilm.com it looks like a HD to Film transfer would be upwards of $40k or so.

 

[David Mullen ASC]

Well, I'm sure pointing a 35mm camera at a monitor will end up cheaper than a laser recorder transfer to a 35mm IN or even a cheaper CRT transfer to a 35mm negative (which is more or less a camera pointed at a high-end CRT monitor, just that I believe it's a monochrome monitor and the color is created in three passes using filter. Or maybe not. And it's probably not being transferred in real time, but frame by frame.)

I mean, you're basically doing the work yourself, so it's bound to be cheaper -- the question is whether the quality will be acceptable. There are gamma and color space issues, and if you are running the camera in real time, there may be issues with monitor sync and motion lag, etc.

 

[Evangelos Achillopoulos]

(which is more or less a camera pointed at a high-end CRT monitor, just that I believe it's a monochrome monitor and the color is created in three passes using filter. Or maybe not. And it's probably not being transferred in real time, but frame by frame.)

Yes this is right David...

Added that is grid less because its monochrome so its electron beam shoots directly to phosphor molecules that glow a very wide spectrum white light that is filtered with high accuracy custom coated filters designed to stimulate the film layers with the closest to real life Red Green and Blue spectrum on every pass...

The monitor is a very special design, it is very long like 2 feet's and very small few inches, it has extremely high resolution that is approaching 8K and the image is exposed at once and not scanned like the laser...

The CRT due to the analog nature of phosphor gives the most beautiful images with natural flare, organic look and natural grain, when images that has being acquired from digital sources (not scanned film) printed on camera negative.

Digital images want > camera negative > Analog CRT recorder

To get film flare, organic look, natural grain, get the camera negative emulsion character and the analog pass from the phosphor layer to take an analog anti aliasing process.

Film scanned (which already are analog) > internegative > ARRI Laser digital recorder

Scanned film is already organic, it has flare from camera, it has grain, it has character we only need to copy the digitized bits to an absolute clinical clean film with no grain or character in order to transfer the original beauty of the camera negative to the duplication process of the internegative.

So its not a matter of price it is a matter of need... film needs ARRIlaser IN and Digital needs CRT and camera negative.

LaserGraphics and Celco CRT film recorders have nothing to be compared to a camera shooting a monitor...

My partner in Belgium has made this on-line offer calculator for my filmout service.

http://www.michaelcinquin.com/cost

 

[Dan Hudgins]

DIY printing negative

DIY printing negative

Well what I was looking at doing was either a lab digital to film transfer... or taking a Arri 435 (or something similar), and shooting at a 32-50" HD Plasma, and then doing the negative processing at a lab. So I'm wondering if the Neg. processing would be cheaper then Digital to Film Transfer. I can get the camera and a Ultra Prime lens for $1600 for a day, plus film stock looks like it'll be about $5200.

Judging from the price on DVFilm.com it looks like a HD to Film transfer would be upwards of $40k or so.

OK,

1) Can you get an animation motor for the ARRI 435?

2) If not can you rent a Mitchell Standard/GC with an animation motor?

3) Using the "Gamma spread" mode in my DANCINEL.EXE (tm) you make three or nine exposures per frame, that expands the tonal range past what the monitor can display and fills in the gaps between the 8 bit video board levels, our tests look good. I plan improvements but for now it should be usable.

4) Doing the transfer at 24fps is not recomended because of the shutter and the monitor refresh rates, you will get roll bar if you shoot faster than about 8fps in the camera. You can make an AVI that will play at slower speeds and run the ARRI on a cristal motor at 8fps or 4fps etc. If your making a negative you then need to measure the length of the head to tail sync marks in a sync block and adjust the WAV file length to match before you make the sound negative for that reel. If you are printing to print stock you can slop process the head and tail leders to so you can measure the exact film length. The main problem with not having the camera and monitor in sync would be that you will get double frames, if you are going from 60p to film that will happen anyway, you make the 60p AVI play 8fps and then you set the crystal motor at (24/60)*8=3.2fps on the movie camera, that does your frame rate conversion and removes the roll bar, good for Canon 5D footage maybe? For going from 24p it is not so good though. If you made a click track WAV to play along with the AVI then you could have the click trigger the animation motor on the camera to keep the camera in sync.

Another way to keep the camera in sync with the monitor if you ARE shooting off the monitor at full speed is to record 60HZ pilot tone on the WAV used to make the 24p AVI, then pipe the 60Hz sound through a power amp to run a 60Hz sync motor on the camera, in that case you would use a Mitchell NC camera with single phase 24fps sync motor, you can then use the shutter indicator on the back of the Mitchel NC to phase the camera to the monitor by making a blink section of the AVI with black and white frames, that should help adjust the shutter phase, once adjusted it will stay in sync if the computer can play in sync.

Stop motion using my DANCINEL.EXE (tm) or some other program would be more likely to not produce waste film, I would think.

5) What do you mean by "digital to film" transfer, any film print will have the same film and processing costs plus markup. It just depends on if you are doing a transfer direct to print stock like 2383, or making a printing negative and then making a print. A printing negative will cost about 4x to 8x the cost of the print stock for one print, about. Print stock and processing are about $0.18 to $0.20 per foot if you get a good rate, you have to talk to the lab and work a deal.

6) You can print direct off the monitor to color print stock but the printing time in MUCH longer than going to internegative to EI 50 Daylight ECN, on EI 50ECN you can print at maybe 1sec at f/5.6 to f/8, but on color print stock the time is 4 minutes because you need a Rosco #22 filter. There is a way around that which I am working on now, you take a 1600x1200 LCD monitor apart and back light using a fresnel lens and 2000w tungsten lamp run at about 85v, that give you enough red light to print in maybe 4 seconds at f/1.6, I will need to do more tests when I get my direct to print stock film recorder finished it is being worked on now. Anyway, you could make one and run of as many projects as you need.

7) You can purchase a used 35mm pin registered camera for less than $500 if you shop around, I saw a wall camera with acme movement go for less than that on ebay. A used LCD monitor goes for less than $300, and you can use a junk computer with my film recorder program for less than $500 plus some 750GB disks for the frame storage.

8) How many prints do you need, if you are just making one print then the extra cost of a printing negative may not be worth it, you can DIY or use a Cinevator for that?

9) AGFA CP-30 print stock is lower in price than Kodak 2383 last time I checked, if its still under $0.12 per foot then 8000' costs, $960 for a 88 minute print, AGFA rolls are metric so they are a little longer, so you might get 90 minutes if you are careful. The real cost of processing color print stock is about $0.035 to $0.05 per foot depending on the water cost and if you do bleach bypass, so if the lab charges $0.15 about 2/3 of that is above the suplies cost.

10) You will not be able to do this on one day. You can do test exposures using a SLR and loading 6' into a 35mm film cassete then taking that to a lab, but you should shoot many tests and have them printed before you spend $10000 on film stock and processing to find out that your camera was a little off frame or something like the filters were not right.

11) My DANCINEL.EXE (tm) lets you use dichroic filters on the LCD backlight or Wratten color separation filters on the camera to get better color if that is what you want.

12) You should use a Wratten #2E filter to cut the UV light, printstock and some other lab stocks are not like camera films, and cannot be exposed to UV without great loss of color quality.

 

[Dan Hudgins]

Samples

Samples

LaserGraphics and Celco CRT film recorders have nothing to be compared to a camera shooting a monitor...

If you can send me some samples of negatives made on such scanners along with the TIF frames they were made from I can scan them close up to see if that is true.

Anyway, most of the resolution in the negative is lost in the contact print made and the projector being out of focus, my tests show that a 1600x1200 LCD is close enough maybe to a Celco in so called "2K" mode which is only 1828 for full with mode since the sound track cuts off part of the image.

Good computer monitors are higher resolution than the images you get on a movie screen in a typical theatre, which are soft and closer to 1280x720 as published tests have shown.

==

Used Celco etc. film recorders come up on ebay from time to time for $1000 to $3000 or so. They would be slower than using a backlit LCD panel, which seems to be what Cinevator uses. When I saw CHE #1 and #2 the Cinevator "2K" print was sharper than the projection lens, and the projector was not in focus enough to see even the limited resolution of the Cinevator.

 

[Evangelos Achillopoulos]

I have print a 4k resolution chart image in my LaserGraphics and looked it with a x200 microscope at Kodak Cinelabs and we found that there are no pixels no scanning traces and the resolution on Kodak 5201 was approaching the 1600 lines in all directions... we were basically watching the grain structure of the film... like big blobs of grain between the lines...

These are not scanners... they are film recorders Dan...

Lasergraphics film recorders...

http://www.lasergraphics.com/us/pages/theproducers1.htm

 

[Dan Hudgins]

These are not scanners... they are film recorders Dan...

I understand what you are talking about, I said IF YOU SEND ME THE FILM THAT came out of the film recorders along with the TIF frames I will SCAN the film samples and enlarge the area from the TIF so I can post the results here so people can SEE the resolution of the film recorder vs. the original TIF frames.

I can then print those same TIF frames in the DIY film recorder I am working on and then we can SEE if a DIY film recorder looks any worse on a print made from the negative samples you send to me, just be sure you send 6 inches or more so I can make a film print without the splice being on the frame.

I can then scan the PRINTS made from your film recorder negatives and the direct print in my DIY film recorder, and post the results of that so people can see how the PRINT made from a commercial film recorder compaires to a DIY direct to print stock film recorder.

 

[Evangelos Achillopoulos]

Dan the resolution is just a small part of the whole equation...

Its probably the 15%... all the rest is color gamut, dynamic range and color management of the film recorder...

With the scanner you will constrained to the limits of the imager-sensor of the scanner...

With a Zeiss x200 microscope you are not constrained since you see with your eyes the film it self...

 

[Dan Hudgins]

Fuse

Fuse

Dan the resolution is just a small part of the whole equation...

Its probably the 15%... all the rest is color gamut, dynamic range and color management of the film recorder...

With the scanner you will constrained to the limits of the imager-sensor of the scanner...

With a Zeiss x200 microscope you are not constrained since you see with your eyes the film it self...

People do not watch movies through x200 microscope from color negative film stocks.

Our scanner used a printing Nikkor the sharpest lens used for motion picture printing, some of the first IC were made with that lens, I can use it at higher powers to blow up the negative to be equal to about 16K full frame, so there is no problem seeing grain and anything else on the negative.

I use color seperation exposures and bracket the sensor to make 27 exposures per color negative frame, those to to 48bpp TIF files and get fused into a single HDR image of the color negative, nothing on the film of interest would be lost.

My system does all the color gamut, dynamic range and color management for both film scanning and film recorder output so that the print made is one-light, no corrections are needed outside my system, if you give me the film samples and matching TIF files then everyone can see the results.

After I get the prints made if someone wants to view them in a projector that should be possable. We have a movie projector that projects still frame with heat filters, we use it all the time to view tests, our normal lens is 4" for projection but I can put a 2" lens in to get a larger view, and could mount a 1" lens to see it bigger if needed, not that that matters since people watch the whole image not a spot from it, right? They could use a film strip projector maybe as well, but before people can look at the prints from the negatives I would first need some sample negatives to print, right?

What matters is what people can see on a screen, what my scanner can show would be BETTER than what people can see on a screen for the most part. I would match the film scans to the source TIF files you used in the film recorder to make the negatives as seems in order, as for the prints those should also be timed to resemble the TIF. If you have a timed example BMP for the frames that would serve as an example for the timing that could also be sent. Small details of timing are not relevant since a RANGE of timings can be made for all source negatives and results, people who have a project can judge the over all effect of the film recorder on the end result, there is never any one "correct" color grading except for the LAD image. Do you have the LAD output to negative?