Home of Professional Cinematography since 1996
Super-16 To 35mm
Looking for some advice on S16mm Blow ups. A project iÕm on is looking to shoot Super16mm with a 2.35: 1 ratio and needs to deliver a 35mm print. The producers are insisting on a 2.35:1 ratio and one of the options they are looking at is to grade to a DigiBeta Master in a Davinci and then get a tape to film transfer at LipSync in Soho. Have any of you used this route before for Super16mm? and if so how did the process handle using such a tiny piece of negative. I am particularly concerned about enlarging a 2.35:1 area from a Super16 frame. Also have any of you dealt with LipSync before?
If you do the calculations, it turns out that the enlargement from super-16 to a 2.35:1 ratio is only 4% greater than the enlargement from standard 16 to a 1.85:1 ratio.
Not that blowing up from standard 16 is so wonderful, but it's not that alarming either.
If you are considering the digital blow-up route, you may want to consider going to a HD transfer rather than a digi-beta. I suspect that the video format may be more of a constraining factor than the original film size.
Group Technology & Services Manager
I'm still curious at the the difference (besides color bit depth issues) between a 35mm "blowup" from a 16mm or S16 original transferred to D5 HD 24P compares to the same 16/S16 original transferred at 2K?
I'm guessing that there's quite a cost savings, and the quality difference, given the resolution of the 16 original, might not be radical. Then again, it might be...
Jeff "HD is cheaper than Data" Kreines
The problems of taking a 2.35 extraction from a S16 frame and blowing it up to 35 pale by comparison to the problems of using DigiBeta as an intermediate. That is perhaps the single worst part of this approach. Either do the extraction and blowup optically or go through a proper digital intermediate process and at least maintain some of the image integrity.
It's all very straight forward except where the producers insist on a 2.35:1 ratio. Here's where ULTRA 16 would be a good possibility. In any event shoot the lowest ASA you can. I think Kodak's 7274 is one of the best emulsions and well suited for the demands of this project. Also, get back from your subject and use longer lenses to counter the less depth of field you have with 16. I like Du-Art. They just about invented 16 blow up to 35.
Edwin Myers, Atlanta dp
If you have to go via video (and you don't - the best result would be to xfer avoiding SD video, and go via HD or a vfx digital film format) - but if you want to keep costs down by going via digital video, then you might be better off using PAL, rather than NTSC.
Also - see if your tk facility can do an anamorphic scan - using all 625 of those video lines, rather than a horrifying 300-odd, which is the best you'd get with a flat image. 300 vertical lines of resolution - that's not much better than VHS.
My recommendation, for what it's worth, would be to do it at 1 or 2k, on a Spirit.
P addy Eason
moving picture company, london
I shot a 16mm feature [free-ture, if you know what I mean!] 2 years ago. It was telecined to Digi-Beta. It was then dumped into an Avid and cut. That image [such as it was then] was exported to a BetaSP [I believe] and then given to a very well known Burbank post facility and a 35mm 'master' neg was struck, resulting in a 35mm print. I held my breath and sat down at the 'premiere' with about 300 others. No one had screened it who knew anything about film...so it was a total shot in the dark as to how it would look.
The picture started......
I almost puked!!!!
Lesson learned: DO NOT EVER EVER EVER EVER CREATE A 35MM PRINT FROM ANYTHING REMOTELY ASSOCIATED WITH AN NTSC IMAGE.
Learn from my nightmare!
This picture would have looked better if I had shot on a VHS camcorder! No joke! And the posthouse did everything they could do, but the damage was done before they got their hands on it.
Jeff Barklage, s.o.c.
www.barklage.com view reel on: www.reelsondemand.com
>But how would you compare S16 - HD24P D5 - 35mm to S16 - 2K data - Film? Do >you thing, with the limitations of S16, that 2K would be vastly superior to HD?
Makes good sense.
Jeff "curious cat awaiting death sentence from cinematic tribunal" Kreines
Define "vastly superior."
My basic answer is a qualified yes. Scanning to a data format and going through a "proper" digital intermediate route preserves a lot more of the color depth (as well as the dynamic range, if a 10 bit log scan is used) and doesn't introduce video noise. Going to current HD formats limits the color depth (to 8 bit) and does introduce video noise, however little it might be. With a 16mm original, grain is your primary enemy, and the data route will likely be kinder to this than an HD transfer. It's not really about pure resolution, it's more about color depth and fidelity, as well as cleanliness of the signal path.
Did you at any time make an s-16 answer print?Also what kind of transfer did the lab do on your S-16 neg and from the Digi Beta what kind of scanner was used to make your 35 mm neg?
There are some variables here that we don't know and quite frankly,even after 23 years of working in the business its been a long time since I struck a print from anything and then only a film negative.I understand there are different resloutions for going from a digi tape master.Right now I backed away from a no budget project that was originated on DVCAM with hopes of going to 35 mm.The director doesn't understand lighting and won't even try to arrange a deal for anything resembling a lighting kit.Now I've worked with the barest of bones lighting gear as I've shot TV news and 2 man crew commercials.In those situations you ammend your project and make compromises based on what you have to work with.Sometimes you even get a better product than what was originally envisioned.This guy doesn't understand that.I'll have to let him find out for himself when he spends what little money he has on a 35 mm neg and print.I hope he gets a lab that's honest.
Independant Shooter and Image Mercenary
Your friend, who wants to do a DV to 35mm print will really have a shock! I did a miniDV short film last year, we had a very large pro crew and grip trucks, HMI's, genny, the whole 9 yards..all donated of course.
The edit was exported to digi-beta [cut on an Avid with Flame effects] and a 35mm neg was created somewhere in Texas [supposedly THE place to do this], a timed answer print was struck and I saw it projected...
I ALMOST FAINTED!! IT WAS UTTER CRAP!
Plus, the final cost of the 35mm neg, the 1st timed answer print and the 2 release prints struck were well more $$$ than if we had shot on 35mm Panavision!!!! There was no savings! [except for the extra speed we got in running around with a 4 pound camera versus a 40 pound camera].
The other movie I mentioned was shot on S-16, the telecine was NOT done on a 2K res machine, instead, the producers decided to go the less expensive route and do a traditional telecine on an older Rank at standard NTSC scan of 700 lines. The digi-beta 'master' tape was then loaded into an Avid and cut, exported to another digi-beta and then 4MC in Burbank did the blow-up. I have no idea what machine 4MC used. I have had great experiences with 4MC in the past, so the poor resolution of this film was not their doing, instead I place the blame on a low resolution telecine transfer of the original camera neg, and then the inherent video compression of an AVID system.
But it was truly horrible!
Jeff Barklage, s.o.c.
The only people who call Texas THE place to do that are the Texans themselves... they use a homemade film recorder (one can speculate what it consists of from looking at their website).
I'm not suggesting that DV-to-35mm is a wonderful thing (hardly) but there are better ways to go about it and probably better cameras than the XL-1 to try it with.
You know, like the A-Minima.
Jeff "uses DV when appropriate" Kreines
>Not a fair comparison since you'd have to pay for answer print and release prints even from a 35mm Panavision negative.
But your point is still good. There isn't an automatic financial saving this way. Although you don't mention what digital-to-film process was used "somewhere in Texas" it's irrelevant. Nothing can make DV look like 35mm film.
There are situations where DV offers no overall cost saving but there is a cashflow advantage, in that you pay the big bucks after you've shot the film instead of before. Of course the same is true of shooting super 16 film, which has a much better chance of being passed off as 35mm.
Also, there are situations where the 4 pound camera instead of 40 pounds can make the difference between shooting a film and not shooting a film. The lightweight feel to the cinematography can sometimes justify the sub-standard* quality of the final print. Once again, super 16 (especially with an Aaton Minima) gives the same benefit.
* I use the term "sub-standard" advisedly, in its traditional, non-pejorative sense of sub-35mm.
The Atlab Group
I agree that the more resolution, the better, and I think the data route would offer the highest quality (and cost!). I would not expect D-Beta resolution to be adequate.
That said, if cost is an issue, is it practical to shoot 35mm 2-perf? It would sure bring the cost of stock down and bring the quality up, plus you're just about at aspect ratio. I'm not familiar enough with labs to know if it's easy for them to optically print from 2 perf to a release print, but I do know that modern telecines can read 2 perf and output a viable digital intermediate.
Just a thought. But in any case you should at least go for a HD transfer over D-Beta!
Shooters Post & Transfer
I would definitely not transfer the Super-16 to NTSC video for an eventual transfer to 35mm. NTSC video is just too low quality, much less resolution than Super-16. You can see it on the commercials that are often run before movies, along with the trailers. The are shot on film and posted on NTSC (SDTV) video and then output to film. They always look blurry and I can often see digital artifacts. Although you'd have the tonal quality of film, your resolution would be the same as if you had shot in the video format you transfer to (Digi-Beta, for example).
I would recommend a transfer to 1080p HD/ 24, making that your electronic master. After having seen both Star Wars trailers on the big screen I believe that 1080p has enough resolution to capture most of the detail in a Super-16 negative. Plus the HD master is a valuable asset for the future exhibition of the completed film.
Something else to consider: 1.77:1 through a 1.33:1 anamophic expansion would be 2.37:1 which is just about perfect. If you could find an anamorphic adapter designed for digital video (the use the 1.33x squeeze) and make it work with your lenses, then you could utilize nearly the entire Super-16 negative (with a slight buffer for gate hairs which is desirable). Also, this 1.33x vertical squeeze is programmed into a lot of electronic equipment these days, and is standard in many non-linear systems such as Final Cut Pro.
Cinematographer/ Camera Technician
Thanks to all the people who responded to this thread both on and off list. Unfortunately the money just isnt available to the producers to go with a HD transfer which is what i asked for initially. After due consideration ive convinced the producers to go with a standard S16 blow up with a 1.85 ratio. They can just about afford that:)
>I replied to Stephen privately, but I just wanted to let you know that we recently completed two feature films (shot on S16 Aaton with spherical lenses) that were blown-up to Super 35mm interpositive, and from there to 35mm anamorphic intermediate negative for contact printing onto positive stock.
>Both the producers and the Dop were very happy with the results, one of the films 'Monte Carlo' was screened for the Netherlands Society of Cinematographers at the suggestion of the Dop Richard van Oosterhout.
>We looked at scanning Hires (2240*1344 pixels) and recording back to film, but neither price nor quality could match the optical blow up.
>Did your production actually test the results of DI vs optical path ? I am certain you had good results but were there any actual face to face tests done?
>Tom Gleeson D.O.P.
>I am certain you had good results but were there any actual face to face tests done ?
>We did about 3-4 minutes of digital opticals in the film, they were scanned and recorded at 2K+ with an Oxberry Cinescan and a Solitaire CineIIIFLX, and they just about didn't fall out too much compared with the rest of the film done optically. To do it from D1 or DigiBeta at 720*576 cropped to 2.35 format would give you almost one tenth of the pixels we had.
>The production was very low budget, several other companies offered to do it fully digital, but none could match the price and quality of our hybrid system.
>This message is related to the previous discussion concerning an eventual digital blow-up from Super 16mm to 35mm.
>I agree in the point that a DATA intermediate (like 2k or better) should be used rather than any SDTV format (D1, D-Beta etc).
>I have recently seen 16mm footage scanned on Spirit at 2k RGB 10 bit Linear resolution and then recorded to 35mm 5245 negative stock. Then this negative was printed on 35mm positive material.The final resolution was very good. However I am sure it is highly possible to improve the result, if a better analysis of the emultion is done during scanning.
>Daniel Henriquez Ilic
>I recently shot a Super 16 short film and am trying to offer the director options on Super 16 blow up. I know the Super 16/digital intermediate feature "Conspiracy" was widely praised on CML for its look. Does anyone know which lab or post house they used? Or have other reccomendations?
>Kodak's Cinesite Hollywood did that work.
>Other companies are now doing D.I. work, like Technique (Technicolor), E-Film, Digital Film Lab (London / Copenhagen), Dubois (France), etc. I
>think labs like CFI, FotoKem, etc. are looking into doing this work too.
>In fact, I think anyone with a Spirit Datacine and an Arrilaser recorder is trying to get into the field now, although you might be better off at a place with some experience at doing it. There's more to it than just making a 2K data transfer and dumping it to 35mm.
>I don't know who in NYC is doing this work... maybe Duart?
>I did a conventional optical printer S-16 blow-up at Colorlab (Rockville, MD) that came out pretty well. Does anyone know who did the S-16 blow-up for "Monsoon Wedding"? I was amazed at how rich the colors were. I suspect Vision Premier print stock... Lately it seems that S-16 is getting better & better in general due to new neg stocks, new dupe stocks, new print stocks, digital intermediates, etc. After seeing how good "Monsoon Wedding" looked, I wondered why more lower-budget indie films aren't shooting in S-16.
>I also think personally that 16mm looks better blown-up and projected in 35mm in general, even compared to a contact-printed 16mm print. It always seems to me that a 16mm answer print is softer and grainier than the final 35mm blow-up -- I think the larger print area of 35mm is a benefit (just as some 35mm looks better on a large screen if blown-up to 70mm first.)
>Note that whether optical printing or digital intermediates are used, these blow-ups look good mainly because they were shot well!
>Cinematographer / L.A.
>I am currently working with Colorlab's office here in NY on a Super16 project. Shot on Super16, all Kodak Vision stocks.
>Colorlab's pitch on a Super16 IP sold me on the idea to blow up to 35 upon the film's completion. The Kodak IP stock is known for its tight grain and even contrast. If your director really wants to, a blow down to regular 16 could easily be accomplished from that same IP. But if the cost of 35 is the biggest concern, the advantages of a print that can be projected at any major theatre, not to mention the increase in respect a "35mm" film receives, far outweigh the additional printing cost.
>The difference in the blow up from the blow down is less than 15%, I think. Call Josh at Colorlab, he can tell you the details.
>>Note that whether optical printing or digital intermediate are used, these blow-ups >look good mainly because they were shot well!
>Cheapest option for a S16 to 35 blow up is to cut the 16mm neg as for a 16mm release (A&B rolls), make a S16 IP then blow up to 35mm dupe neg.
>You'll get better results if you blow up at the first stage (i.e go from S16 neg to 35mm IP). Your lab will probably ask your neg matcher to cut over-length and then fine-cut the IP (long explanations of why, but it works better). It's better (sharper, and often less flare from the blow-up lens) because you have all your duping and printing in 35mm not 16mm. But of course it's dearer.
>At this stage, the dearest, and some say the best, option seems to be the digital scan/record method. This avoids any dupe printing stages, and brings you the glory of a Digital Intermediate. We're talking a 2K scan and a laser record here, not a telecine transfer and kine.
>The relative differences depend a lot on the material you start with, on the exact equipment and techniques the lab uses (optical lenses vary for example). Test, and look at other people's work from the same lab.
>As usual, David Mullen's last word is the best summary, and worth repeating, so I'll repeat it.
>The Atlab Group
>DB Griffith here, experimental filmmaker, shooter, and repairman of Bolex and other 16mm cameras, projectors, etc. I have worked with a guy named Jack Rizzo, owner of Metropolis film lab in NYC (30th and 6th), who does optical blow-ups from S16 to 35 with no IP/IN. He blows up from AB rolled S16 directly to 35mm Vision release print stock. The results have been very good. Furthermore, he will do a S16 answer print before the 35mm answer print to screen on a S16 projector, for checking color correction (further reducing the cost). If absolute lowest cost is an issue, and you won't need too many prints made, this might be the answer. Contact me off list if you want the phone number.
>David is absolutely right, if you do it with someone who know's what they are doing the process can be terrific, even if you can't afford to go with a 2K data transfer. In NYC I have had two wonderful experiences with Tape House Digital Film.
>For a film we screened at Sundance this year they took a digital intermediate (a D2 transfer of a super 16 neg done at Nice Shoes on a Spirit) and blew up to 35mm. With the exception of one or two minor video artifacts, the print was gorgeous - colors very true to what we saw when we did the transfer, less grain that I have ever seen in the optical process (original neg. stock was 7274 & 7279), deep blacks.
>They also did another print for me last year (D5 transfer of a 35mm neg., also done at Nice Shoes on a Spirit) and made both regular and scope prints that had no video artifacts and looked beautiful as well.
>So, although I would do a 2K data transfer if at all possible, I think the key is getting a talented colorist, backing off the contrast just a "skosh" (since you'll get in back in the print), and recording back to film with people who know what they are doing. And when you're doing this with a short film I think the process can even be affordable (depending on what your definition of affordable is I suppose).
>DP - NYC
>Aren't D2 and D5 standard definition formats? Weren't the blow ups kind of mushy looking when projected on a large screen? Or were they shown in venues with smaller screens?
>Los Angeles based Director of Photography
>D2 is a composite (not component) SD digital format. It is basically a digitized analog composite signal. In my opinion, it should never be used as the source for any additional post processing; it is a distribution format only.
>D5 is a multi-resolution component HD format.
>DP - Mariposa, CA
>I'd clarify that a bit and say that D5 is a multi-resolution format. D5-HD is one configuration that requires an HD processor chassis, that can be ordered as an internal module or an external module. The base D5 configuration is standard definition, "601" format component digital video, uncompressed. The HD model employs a degree of data compression for HD recording.
>The S16 to 35 blow-up, thru IP/DN, gives a filmmaker more options. After a S16 IP is struck, the filmmaker can then decide to either blow-up to 35mm, blow-down to 16mm, and/or use the IP in telecine for a video master. This decision is usually determined by the amount of funding available. As such, the S16 IP is a useful (and valuable) element for the filmmaker because it provides him/her with options.
>A direct blow-up produces one 35mm print (hopefully clean), and requires (excessive?) handling of the original A&B rolls. The filmmaker may then have to incur the cost of any additional prints off the A&B's, thereby negating their initial savings. It may be marginally cheaper, but perhaps we can assess the value, or lack thereof, in this procedure...
>I work with a producer who used to be very hard headed about using D2 to master in telecine until one day when I got the colorist to do a simo to D2 and Dbeta. The differences were far more obvious than I would have guessed, and the producer never said the words D2 again.
>DP San Francisco
>Sorry to get back to you late, but in answer to your question, yes and no.
>As Mike Most guessed, we transferred one of the projects to D5-HD, through and Inferno, then out to 35mm. That played on 50 screens of various sizes in New York (it was a film festival promo) and looked great even on the large screens.
>As for the one that transferred to D2, well, yeah, I saw it only on smaller screens like the main Technicolor screening room in NYC, and at the Prospector Theater in Sundance (which although not tiny, wasn't a huge screen). So yes, it was certainly muddier than a 35mm print from a 35mm negative, but I thought a pretty fair competitor to a 16mm blow-up. And though I will protect the names of the innocent, the DP who shot the feature that we preceded thought the film was a print of a 35mm negative (or at least that's what he told me after I had bought him a few drinks...just kidding).
>DP - NYC