Suggestion to RED

[mtseng]

Jim why don t you apply your sensor expertise in a film scanner like arriscan.
I know this will sell a lot and you have all to do that in a affordable price.

Doesn t need speed, make it affordable and will be a extreme hit.

 

[Julio Quintana]

That's like asking Apple to invest in "really affordable" diskmans (portable cd players)...

Focus on the future...

 

[Martin Weiss]

While I do see a bigger market for a quality, sort-of-affordable 8mm/super8mm scanner, I doubt that the market for 35mm scanners is more than a 3 digit number.

Either way, a film scanner would be sort of "with built-in obsolescence"

 

[Jean Déraps]

The whole point of RED is to circumvent film. Once RED really succeeds in surpassing the quality of 35mm film (with Epic?), film scanners will only be used for archival purposes, since no one will shoot on film anymore. Case in point: Only purists shoot stills with film anymore...

 

[mtseng]

The film market sell more this year than last year.
i think it will have some years until it die.
This scanner will be a revolution ,Jim know that.
And this scanner will don't make RED sell less.
It's just make arriscan, spirit sell less.

 

[Dan Hudgins]

Film scanner...

Film scanner...

The film market sell more this year than last year.
i think it will have some years until it die.
This scanner will be a revolution ,Jim know that.
And this scanner will don't make RED sell less.
It's just make arriscan, spirit sell less.

If you want a film scanner you can make one out of an optical printer and an industral digital monochrome camera or DSLR for maybe $1000. Optical printers are going to the dump now (thats what I was told by a lab, he had a fixed price if we did not buy it he was sending it to the dump, seems a shame to me...).

The mechanical workmanship for making a scanner in the US would cost too much. There are many places that scan now, and there will be less film to scan. It is lower cost and faster to print film to film in a used optical printer than to do a DI, but you do not have as many options with color correction and dust removal.

I have done a conversion like this, the DIY software to run such a scanner is "freeish" from my Web site, as is DIY film recorder software, if you have an optical printer you can use the projector for a scanner, and the camera for a film recorder.

Theatres will be converting to 3D Digital projection over the next few years, so if you have a project you want to do on film it needs to be done quickly, as I tell my Brother every day, time for working with film is slipping away.

Since my DI system works with both RED ONE (tm) "footage" and film scans I look at both side by side all the time, and am sad to say that if you light the subject well for the RED ONE (tm) T500 film stock cannot match the smooth images the RED camera makes, if you have film scans that are "better" than the RED one RED LOG TIF 4K frames, please let me download them and look them over close up, nothing we have made comes close from color negative stocks. We did scan some slow, EI 3, black and white frames and did get maybe a tad more detail, but film always has crud on it and unevenness that the digital cinema camera images don't.

It has been hard for me to admit since I have worked with film most of my life, but the point where digital has the upper hand is close or has already past.

Since filmrecorders have a hard time holding detail in the highlights because they are run twice as fast as they should be and the printing negatives do not have the extra density that they would in the highlights when made on an optical printer, the difference between what ends up on the screen in the highlights from RED ONE (tm) footage and film scans gets evened out.

Also graders tend to push the highlights up too far or the shadows down too far for the film print to hold enough detail at the ends of the range after going through the filmrecorder, so everything tends to get a more clipped video look, and in time people will forget what film to film printing looked like, for the most part any current film has no film to film printing all the way through, so the soft focus clipped shadows and highlights DI look is what people think of as "film" now.

Bottom line seems to be, anyone who wants to make their own film scanner can now using an optical printer, and few people would need one, because labs are closing and selling what they have back onto the small market. I talked to some people years ago about maketing a low cost portable film scanner for indy people, no interest. And because of the dust issues most people would rather the lab do such work for them anyway.

If someone wants to try to make a low cost portable scanner I can tell you how it should be made, but with a profit of maybe $500 per unit and maybe sales of 30 units, would you still be interested?

 

[Jarred Land]

this is such an odd request... or is it?

 

[Sarah C.]

Hey Jarred,

I've been looking at the footage we've had scanned for our VFX shots on a certain feature I am working on.. top house in Hollywood doing the scans.. I've been underwhelmed. Don't know if it's the hardware or the operation of it. I've seen better scans over 10 years ago on a 35mm still film scanner by Polaroid!

But then by the time you guys get a scanner out the door.. that show may have moved on. We'll all be shooting Red's for feature production!

~Sarah

 

[mtseng]

this is such an odd request... or is it?

Isn't a request , just trying to help RED to see this market.
The arriscan minute price is very very expensive.
Like $800 p/ minute here in Brazil.
Here you need to wait about a week to have your material scanned.
$1000 the spirit telecine hour

I am sure every post house and the producers that do 35mm commercials and films wanna buy one scanner.
Not just for the money they will save, most for have the speed to scan the materials anytime.
Lot of quality people making commercials through betadigital telecine with spirit.
I am sure they all will love experiment a 10 bits dpx log 2k scanned film
to make chroma key for example.

It will doesn't make people don't buy RED camera. It will just make RED more stronger in the market.

 

[Karl Gustav H.]

Didn't someone do this with a standard RED body as a 8mm scanner?

 

[Dan Hudgins]

Msrp?

Msrp?

Isn't a request , just trying to help RED to see this market.
The arriscan minute price is very very expensive.
Like $800 p/ minute here in Brazil.
Here you need to wait about a week to have your material scanned.
$1000 the spirit telecine hour

I am sure every post house and the producers that do 35mm commercials and films wanna buy one scanner.
Not just for the money they will save, most for have the speed to scan the materials anytime.
Lot of quality people making commercials through betadigital telecine with spirit.
I am sure they all will love experiment a 10 bits dpx log 2k scanned film
to make chroma key for example.

It will doesn't make people don't buy RED camera. It will just make RED more stronger in the market.

What do you see as the MSRP (retail price) for a 35mm 2K film scanner?

If the price is high enough you can have drawings made and take them to shops to have the parts made today, a movie film scanner is not something that requires unknown engineering.

You can ask the buyers you think are out there if they would purchase one and take "pre-orders".

What frame rate do you think the market you see would require?

 

[Matthew Greene]

Didn't someone do this with a standard RED body as a 8mm scanner?

That'd be me. Using a workprinterXP from www.moviestuff.tv connected to the GPI port on RED you can do frame by frame scans of S8 or 16mm. It's a pet project for me and haven't finalized the last few steps which are custom mods to the light source of the system and avoiding standard optics by having a custom lens built.

 

[Matthew Greene]

What do you see as the MSRP (retail price) for a 35mm 2K film scanner?

If the price is high enough you can have drawings made and take them to shops to have the parts made today, a movie film scanner is not something that requires unknown engineering.

You can ask the buyers you think are out there if they would purchase one and take "pre-orders".

What frame rate do you think the market you see would require?

If you take a good pin registered projector (or optical printer) wether it be 35, 16, 8 etc... and do similar modifcations which allow very clean frame by frame transfers such as what Roger Evans does with his products ( www.moviestuff.com ) and team that with a high quality imaging device and then assuming that you have some basic understanding or experience with telecine devices there is no reason that you wouldn't get high quality results with some simple R&D and some cash (I'm in the tens of thousands at the moment). Over the years I have shot lots of S8 along with HD or larger format film formats for it's aesthetic qualities. With my old workprinter XP (pre RED, using Panasonic 3CCD cameras) I was getting better results doing it myself than what I was getting back from Rank Telecines which I can understand many people might have trouble taking my word for. Once I have my system up and running to my satisfaction (maybe over the year end holidays) I'll post something here and although it's not my first order of business, might consider offering it as a service for 8mm, 16mm and perhaps even larger format print and negative film. I do want to finish implementing all my ideas of light sources which give the best CRI, High quality custom optics, multiple passes, particle transfer rollers, wet gate and the option for restoration processing. I want to do this right and when it's ready I'll bring it to this forum first. I'd like to be able to offer a no compromise affordable service as well if I decide to go that route. There's 100+ years of motion picture material out there and I'd love to be able to offer affordable remastering/preservation for the masses as well as for top end clients.

 

[Matthew Greene]

By the way PM me if this is a service you think you might be interested in taking advantage of. I'm doing it for myself mainly but I'll put extra effort into it if there is further interest from the community.

 

[Dan Hudgins]

There color separations...

There color separations...

By the way PM me if this is a service you think you might be interested in taking advantage of. I'm doing it for myself mainly but I'll put extra effort into it if there is further interest from the community.

I would recomend that you use three color separations, that is what we are doing in our scanner. You can use Wratten filters 99, 98, and 70 plus IR and UV 2E+2B. You can also use three LED sets, but you need to check the peak wavelength for the LEDs since some filmstocks have their peak absorption at different wavelengths.

The peak wavelengths for the sensor do not matter as much if you make separation exposures. From a Bayer camera I take just the red for the red separation, just the two greens for the Green expsoure, and just the Blue for the Blue exposure using filtered white light, but if your LEDs are pure you can mix if you do your own de-Bayer since the filters in the sensor are weak, if the de-Bayer you use does not have a monochrome mode to turn off the color matrix, then it can be better to just use the single colors since the matrix will reduce the bits in the other layers.

The compression noise in the RED ONE (tm) frames can be reduced somewhat if you change the brightness of the light in the scanner and make several exposures, then fuse those.

I use 9 scans, three red, three green, and three blue, bracketed, then fuse those into a 48bpp RGB file for doing the color correction. Using 27 exposures would give better results, 9R, 9G, 9B. If the RED ONE (tm) had a RAW mode for single frames that would also help.

My scanner software also works for zero cut editing since that is what it was developed from, it started out as a control program to do zero cut A/B roll printing from 35mm on a step contact printer, and so because of that it has built in light house control for grading/timing. It can be good to adjust the lighthouse shot to shot to keep the histograms centered for very under or over exposed shots.

My DI systems color correction is very good at correcting color in DIY film scans and is fast enough now to output 2K or 4K TIF files for a feature if you use several cheep PC boxes running FreeDOS FAT32 (tm). It should also work with XP Home SP3 since graphics are not required for the frame processing.

About the lens, we have a printing Nikkor which more times than not is too sharp since every scratch, spot, grain, and muck comes through with too much contrast.

You could use a 75mm Macro Switar mounted "backwards", that is you put the "C" mount threads toward the 16mm film. A 75mm Macro Switar should work for Super8, 16mm, and maybe Super16 if you set it to about f/4.5 to f/6.3.

For scanning 8mm I think reverse mounting a D mount 36mm Switar would give good results, you could also put a lens on the RED ONE (tm) to focus through so that the 36mm Switar would be at INF focus, which would look better would need to be testes, since the large size of the RED ONE (tm) sensor would not be so close focus for the 36mm Switar to cause problems if it is stopped down to f/4.0 or so.

The heavy OLPF in the RED ONE (tm) will limit the maximum resolution when scanning 35mm.

If you do not make seperation exposures the color in the scans will not have as good purity and S/N, and have done tests and doing color matrix cannot equal the improvement you get from starting with cleen seperation images. An like I said, single exposures cannot give enough S/N to equal multi-exposure results in pulling all you can get off the film. If you fuse enough frames from a RAW camera you can get 16 bits, with the compressed camera you need to at chaos by adjusting the light house brightness between exposures to have the compression thresholds be penatrated.

When designing a new scanner one should realy use a monochrome sensor to get higher resolution and fewer artifacts. The RED ONE (tm) is fast and does not have a mechanical shutter, so those are good for making for profit scans, but maybe not the best technical results.

We have tried sticky rollers but they do not take all the dust off and can leave an oily deposit on the film that makes a modeled flicker when the film gets its contrast boosted in post. The stickey rollers we had were old and that may have been part of the problem, but since they go bad with time they are an upkeep expence.

A problem with Wet gates is that the thick glass in its window introduces lens aberrations since the scanner lens is not corrected for the extra glass, also you cannot get TCE any longer and solutions other than Perc may not have an index close enough to hide base scratches.

Using defused light gives better results on film blemishes, and the closer you get the defuser to the film the better it works, except that dust on the defuser will also get sharper, so some gap is needed. An intergration dome behind the film might work if you have room for it.

Dust an muck are a big issue, if you can get an ultrasonic film cleaner and an electrostatic air cleaner. We have an electrostatic air cleaner and it does seem to reduce the dust in the air, if you use a laser in your scanner room with the lights off you can see how much dust is floating around. This is the major issue with "home" scanning, the need for "clean room" conditions.

If you would like to compare results I could scan the same film shot you scan in 16mm or 35mm so that we can see how film scanned using a RED ONE (tm) compares, the only way to compair scanners would be to both make a uncompressed 2K TIF file and look at the results split screen. You might learn some ways to get better results?

What software are you using/going to use for dust, scratch, and muck removal of your scans?

Are you going to use IR exposures for dust removal? I am not sure if you can get enough IR on the RED ONE (tm), it seems maybe you can since people need to add IR filters, has anyone shot IR images with the RED ONE (tm) to see what the S/N would be?

I have been thinking of the best way to write the automatic blemish removal software for my system. My Scanner program can do IR exposures I think, and my FUSE can make four color 48 bit files so far, but the other steps are not in release yet.

 

[Matthew Greene]

Thanks Dan, I'd love to talk to you in more detail. I'm in the middle of a theatrical production load-in and I don't have time for a full response to your post at the moment. I really appreciate the sharing of your knowledge... here I go, I'll type fast and forgive my grammar/spelling since I probably won't proofread this. I'm going to jump through many of my experiments so forgive me if it's hard to follow, keep in mind that many are completely different and unrelated to each other.

The best results I've gotten so far are with 3 color separation and a monochromre imager (with no bayer filter or OLPF) wiong custom OLPFs (or plain diffusion) depending on the film format being transferred and whether it was diffused in camera when shot or not (in other words manual/adaptive OLPF). I see the RED ONE currently being a good option for S8 an 16mm but not for larger formats as you suggested.

I've also gotten some good results on negative film by using a (approx)10,000K diffused light source (using a high end systems color pro meant for theatrical illumination which has controllable color mixing) with some cyan additional filtering (to further ofset the orange base of negative film) and capturing bayered (full color) exposures and carefully spatially offsetting 3 exposures and then re stiching them into an interpolated "master frame". This and the following are all purely experimental. My MO is getting the best image even before the light hits the film, followed by minimal very high quality optics (currently frankenstein versions of lenses assembled with elements from leica, zeiss and edmund optics).

For reversal I could consider bracketed exposures perhaps but for negative film I find multiple exposures are unecessary since the full latitude of the film is "compressed" in a range where a high bit depth imager can capture it's full latitude without the need for more complication (I'm not talking RED here but I wouldn't mind them giving me an option of ordering an imager with no OLPF or IR filter). The issue then becomes having enough bits to expand this to video color space and grade it with the least amount of banding and noise.

If the film is ultrasonically cleaned pre transfer and I'm using PT rollers I find the need for electronic/IR dust removal greatly unecessary as well. My father worked with Kodak for 33 years and I have had the honor of collaborating with the late Jon Pytlak who created alot of this technology and mentored/guided me for a few years on my projects.

I've had varying results with wet gates, first off getting the chemical is complicated to start with and I find it unecessary for anything but heavily damaged film. I'm also testing running the footage through some post processing including off the shelf software (Red Giant Film Fix) and some custom code written by an associate of mine which is my preferred solution.

I've played in the past with other methods including a modified Nikon Coolscan 9000 with some custom optics and other experiments. I do all this for fun, it's kind of masoquistic but in many cases I've gotten better results than what I've been getting back from Rank, Millenium and Datacines of course with alot more complication, unreliability and headaches.

I'm finally thinking of building a machine that I can use to never have to outsource scanning/telecine again. It's been a fun journey and it was never intended to be a commercial endeavour, I was trained as a colorist before going to film school, my affiliation with Kodak has helped me and I have a passion for the topic so it's all in my blood. I think I want to finish my pet project.

By the way I've done the same with film recording... at age 14 I was already playing with a mitchell animation camera and transfering 3D animations and composites I had done (using my own color difference travelling matte algorithms based on the info learned from the American Cinematographer's Manual) to 16mm using Amiga computers Multisync monitors and color separations frame by frame by doing primary color passes with dichroic filters.

 

[PeterM]

Only my two pence.
Does a Filmscanner use a Bayer sensor. I thought it use a mochrome sensor and filters.
Second.
The marekt on Filmscanner is not as big es for 2 3 or 4K Camereas.

 

[Dan Hudgins]

Film and clean scans...

Film and clean scans...

Thanks Dan, I'd love to talk to you in more detail. I'm in the middle of a theatrical production load-in and I don't have time for a full response to your post at the moment. I really appreciate the sharing of your knowledge... here I go, I'll type fast...

If someone does contract to make a low cost portable film scanner as far as I can see from my experiences with movie film scanning the biggest issue the end user will have is the dust and muck on the film.

If you pay to have your film cleaned, that cost can be about the same as the cost of harddrives to store footage shot using a RED ONE (tm), so the extra cleaning costs with film make using a Digital Cinema camera even more of bargain.

If you do not clean both before the scan and to retouch the scans later, the results look distracting, audiences no longer tolerate the amount of gunk film shows.

You may not have noticed this so much yet since 35mm has a larger film area and so each frame has many more dust spots than an 8mm frame would have.

Also the printing Nikkor is so sharp that you can see things in the negative like little cracks that are smaller than the grains, tiny pinholes, scum, jet backing spots, all kinds of stuff that are larger than one or two pixels at near 4K resolution. The grains themselves are between as much as 3 and 5 pixels wide on faster negative stocks, and the circle of confusion of the lens after film dye bleed is maybe as much as 8 pixels wide.

You may be able to improve your results by low pass filtering the separation images so that the bandwidth of the blue image is lower than the green and red image.

The main reason for shooting multi-exposures and fusing them is that the Canon XTi we are using as the test camera on our scanner does not have a S/N ratio anywhere close to even 6 bits in the blue most of the time.

The same issue comes up with the RED ONE (tm), it may output 12bits in its files, but those are not all "good" bits, or you would not see ANY noise from one frame to another. To find the good bits you difference them, and the bit delta that changes from frame to frame in each pixel are non-data, so if only the top 6 bits stay put, then you have 6 bits not 12 bits of data. You may need to step the light brightness to overcome the compression delta threshold to get a true noise value.

When you change the brightness of the light source and take more exposures you get closer to the real value of the density on the film.

Using the wet gate inside a sealed room that uses an electrostatic air cleaner could be lethal or explosive. If you use Perc you could pass-out or get damage to your health being in an enclosed space since the film dries to the room air after coming out of the projector wet gate, you cannot open a door or window since you will let dust into the scanning room. If you use Isopropyl alcohol 100% you could have an explosion since the electrostatic air cleaner and relays/switches on the scanner make sparks, also it can cause health problems with the lungs and heart, and this link says coma(?)

http://en.wikipedia.org/wiki/Isopropyl_alcohol

So if you want to use a wet gate you have some issues to deal with on getting fresh air in without dust, and who are you going to vent the vapors onto with the exhaust? If you use refrigeration vapor recovery that adds to your costs of operation.

There will probably be a need to digitize film that has been shot in the past before they are sent to recycling or landfill, but with no standard for how films should be archived in digital format, footage will probably end up in some compressed HD format resulting in irretrievable loss of image data.

Using uncompressed 4K TIF 48bpp to archive all movie film is not going to happen because of the time, space, and costs, even though it should be the MINUMUM standard for preservation.

The good thing about having your own scanner is being able to scan at full resolution and then make workprint resolution frames for editing, that way you only need to put the film through the scanner once that can reduce the number of scratches and cinch marks it picks up. But I am not sure most people could deal with the film clean room aspects enough to be happy with the results they would get. You see the crud more when you watch an uncompressed AVI than when you look at a MPEG2 or other compressed file type since some compressed files are only maybe 4 to 6 fps for all parts of the frame so some of the dust and muck does not get displayed. If you are going for a filmout you may not realize how bad your scans look until you watch your filmout at 24fps, when you see every frame in full. It sure has been an eye opener for us, and made us respect more what the good labs can do and why they charge so much.

So, although I thought there would be a market for a small end user scanner a few years ago, I cannot see much need now that Digital Cinema cameras will sell for less than the cost of a few rolls of film. Can anyone here make a good argument for their personal purchase of a say $10000 scanner at this time that scans maybe 4 to 8 frames per minute (dust and all)?

 

[Dan Hudgins]

Bayer vs. Mono...

Bayer vs. Mono...

Only my two pence.
Does a Filmscanner use a Bayer sensor. I thought it use a mochrome sensor and filters.
Second.
The marekt on Filmscanner is not as big es for 2 3 or 4K Camereas.

I think some people have tried to use Bayer in one exposure per frame to scan with, the J&K scanner seemed to work that way, maybe some others. Not all labs let you into their scanner room to see what they are using, and there is pressure to run everything as fast as you can, so no commercial scanner is going to make 27 exposures per frame, even if you can see a difference.

I think some 2 1/2 sensor backs have been used in some labs for special projects, and those must be Bayer?

I was talking with a guy in a LA lab who wanted to use a Nikon DSLR and start a scanning lab, so people have the idea that one exposure is enough. My tests show that you can get an images, but the color is not as good as separations even with extra software processing, like they say, "garbage in, garbage out".

The one good thing about the Canon XTi for making seperations is that it has good dynamic range on the brackets, and I can hold both ends of most dense negatives in three exposures per color. But its noise level is way to high and with 27 exposures per frame it will break after a few thousand feet scanned if that much. So DSLRs are not a good idea for anything other than testing your projector, lens, and light house, at some point getting a good monochrome camera will probably be on the purchase roster...

 

[Matthew Greene]

Only my two pence.
Does a Filmscanner use a Bayer sensor. I thought it use a mochrome sensor and filters.
Second.
The marekt on Filmscanner is not as big es for 2 3 or 4K Camereas.

Depends on the make/model