Jump to content

S16mm at 4k


J Van Auken

Recommended Posts

I am to DP a short in Colorado in January. Lots of exteriors and vistas, etc. For budget reasons we're on super16 with zeiss SS and a mix of the new Fuji vivid stocks.

 

The interesting part is that the post facility is offering a 4k di for marginally more than the 2K we planned on. We've got the reserve money and enough storage, so it's doable.

 

My question is, has anyone here gotten a 4K scan from 16?

 

I want to make sure to squeeze the details out of the landscape, especially for a future blowup, but I'm not sure if there's much resolving power there beyond 2K as I've never gone bigger on 16 before.

Link to comment
Share on other sites

  • Site Sponsor

Do you know what kind of scanner they are using? The Arriscan does a 3K ->2K internally and Northlight is a 4K linear sensor thet down samples to 2K for 16mm. If they are offering a 16mm projector with a 4K camera filming the image go elsewhere.

 

-Rob-

Link to comment
Share on other sites

4k for Super 16 is overkill in my professional opinion. This seems to be something that a lot of clients or 16mm enthusiasts don't understand or like to hear. I did a real-world test myself with the Arriscan we have at work with a professionally shot feature film we were working on. The end sequence had a lot of awful looking blowups/resizing/digital zooms that we were hoping would look marginally better if scanned at 4k. The grain was barely sharper and there was no overall visible difference. This was the case with scenes shot on Kodak 7218 (500T) and Kodak 7212 (100T) that were processed normally and not underexposed. Go with a good overssampled 4k-->2k or 3k-->2k scan. Good luck.

Link to comment
Share on other sites

  • Premium Member

Working in 4k would be hugely time consuming for marginal gains in the final HD output. If you're going to a film out 2k is still plenty of information. This should probably be another thread but I'm not sure 4k makes sense for anyone unless there are special circumstances like needing to pan & scan the negative more than usual.

 

I would be much more concerned about your colorist and the quality of their work than whether it's 4k or 2k.

 

If the decision is 4k for no or minimal additional cost and you put that on a shelf and work from a 2k down sample then by all means get the 4k since it wouldn't hurt unless you work in the format throughout and it massively slows down your process.

Link to comment
Share on other sites

16mm grain gets blown up so much already at 2K, at 4K, no it is probably not worth it unless you're using a slow STOCK with an extraction.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EDIT: STOCK not SHOT

Edited by K Borowski
Link to comment
Share on other sites

Will you get more information using 4k and is it worth the effort to get.

 

16mm grain gets blown up? Its in direct proportion to everything else being blown up although whether being blown up is the right description I'm not sure as its actually being scanned but surely the more clear it is the easier it should be to get rid of?

Link to comment
Share on other sites

S16 @ 4k is 160lp/mm - high-end lenses and certain low-speed b/w-stocks can have usable information in this range - stock shot on Vivid with old Zeiss SS? I doubt it.

It would be interesting to know which scanner is used, massive oversampling (scanning with higher resolution than the actual negative contains information) can lower the information loss caused by scanning.

My guess: it won't help much but when the budget allows it and maximum quality is needed, it's worth a try.

 

"Grain get's blown up" by grain-alaising. Scanning at a very low resolution (<1500ppi with fine-grained stocks) causes barely any noticable grain at all, high resolution (>4000ppi actual scanner resolution) starts to detect actual grain - everything in-between can cause alaising between the size of the grains and the scanner resolution. The result is more "noise" in the finale image than there actually is in the source material (like moire in shooting test charts - huge structures appear in the image where the chart actually shows fine patterns). So material scannend at 4k and downsampled to 2k can look less grainy than 2k directly scanned. But these are numbers for S35 - 2k @ S16 (4000ppi)0 should propably not cause alaising.

 

@Elliot Rudmann

4k S16 with the ARRISCAN? I thought it "only" scans 3k downsampled to 2k output with S16?

Link to comment
Share on other sites

We shot my film using the latest zeiss superspeeds with an SR3 I would loved to have scanned the film to 4k and sure would have got an even better result. Although I never did this it would be good to hear the experiences from someone who has as it might be an option I will try one day.

 

http://cinematechnic.com/super_16mm/resolution_of_super_16mm.html

Link to comment
Share on other sites

  • Premium Member

the post facility is offering a 4k di for marginally more than the 2K we planned on. We've got the reserve money

 

In that case, I'd try it. With good glass and slow film shooting mostly day exteriors, you may well get some sharpness for your extra effort and expense.

 

 

 

 

-- J.S.

Link to comment
Share on other sites

@Elliot Rudmann

4k S16 with the ARRISCAN? I thought it "only" scans 3k downsampled to 2k output with S16?

 

For the 2k option, yes it scans at 3k and downsamples to 2k. Our scanner has the 6k option. Frames scanned at 4k are actually scanned natively at 6k and downsampled to 4k.

 

You don't hear this being done with Super 16 because it doesn't need to be done with Super 16. The difference between oversampling Super 16 at 3k vs 6k in the Arriscan is barely noticeable. 3k is already overkill. The theory that there might be more to gain from S16 from a 4k scan is too far from the reality at this point in film scanning technology. Like Will said, spend that money on something more tangible.

Link to comment
Share on other sites

Thanks for the replies, all. Widely differing opinions here, which is great.

 

In reference to the machine, it is in fact an Arriscan (they nixed my idea of pointing a RED at a moviescope)

 

Budget wise, this is strictly post money we're talking about, and has been allotted as such. We have a great colorist at the house we're using and there's ample time for finishing, and although I'm sure the other dept. heads would love the cash transfered to them, this bit is mine and I'm not sharing.

 

In the end, I'll probably just test it out. I do suppose it says something that in asking if anyone had done it, no one had first hand experience.

 

Thanks again!

Link to comment
Share on other sites

@Elliot Rudmann

 

Are you sure? I always thought the ARRISCAN works with a fixed magnification and only uses half the sensor wideness for S16 and therefore is only capable of3k->2k @ S16? Am I completely wrong and 3k 16mm really means no microscanning but an adjusted optical magnification?

Link to comment
Share on other sites

  • Site Sponsor

The Arriscan has a 3K area cmos sensor, it is the same sensor as used in the D21 but monochrome instead of Bayer-Mask color. The scanner physically moves the sensor for 6K scans. S-16mm 2K scans are made with the full 3K area of the sensor and then down-sampled internally to 2K. The 2K oversampling is good for Nyquist sampling and I don't think working in 4K will get you any advantages.

 

-Rob-

Link to comment
Share on other sites

I do suppose it says something that in asking if anyone had done it, no one had first hand experience.

 

I operate an arriscan almost every day of the week so I'd consider having first hand experience. If you can get a free test by all means do it and see the comparisons for yourself.

 

Georg - The arriscan sensor has a native resolution of 3k by 2k pixels, microscanning is done to create the 6k-4k scans. During the microscanning the sensor shifts and grabs different parts of the image to make up the 6k frame. It does this whether or not you have 16mm or 35mm loaded in the scanner. The amount of optical shifting and microscanning done varies between the 16mm and 35mm frames obviously because of their respective difference in size.

Link to comment
Share on other sites

I had an introduction to the technology a few years ago - but scanning 16mm wasn't the topic.

So the optical magnification is actually changed by moving the lens to always use the full sensor wideness? I tought S16 only uses half the sensor and 3k is achieved by microscanning, therefore not menationing the possibility of 6k/4k with 16mm. It would have been interesting to see the results in the 4k+-document, instead they used an optical element to simulate "10k" !? The difference was quite noticeable, pointing out that current scanning technology still limits the resolution/MTF.

 

That's something I didn't expect, but this forum is also there to learn new stuff, thanks!

Link to comment
Share on other sites

Georg

I have some 8mm films here from the seventies using kodachrome 40 and I can tell you on an 8mm viewer it blows HD out of the water it's almost like looking through a time window its so clear. However I dont expect people to believe that and I dont care. But it's one of the reasons why I think many have been duped about digital and film

 

Film resolution is often judged By scanning to digital I think there is a lot more to this.

Link to comment
Share on other sites

  • Premium Member

I operate an arriscan almost every day of the week so I'd consider having first hand experience. If you can get a free test by all means do it and see the comparisons for yourself.

 

Georg - The arriscan sensor has a native resolution of 3k by 2k pixels, microscanning is done to create the 6k-4k scans. During the microscanning the sensor shifts and grabs different parts of the image to make up the 6k frame. It does this whether or not you have 16mm or 35mm loaded in the scanner. The amount of optical shifting and microscanning done varies between the 16mm and 35mm frames obviously because of their respective difference in size.

 

Hey Elliot... I believe you that you actually scanned S16mm at 4K from an ARRI.... but I've been digging around and can find no info on ARRI's site but the two post houses that give detailed output capabilities of their ARRI 4K's specifically say that it does 35mm at 6K->4K or 3K->2K but that S16mm is only 3K -> 2K.

 

www.goldcrestpost.com/postny/di

 

http://www.digitalfilmcentral.com/english/aboutcentral.html

 

Does yours have some sort of undocumented S16mm 4K option?

 

-Paul

Link to comment
Share on other sites

Undocumented? I really don't think so! Lol. All I can say is that when we have the 16mm gate up, there's a selectable geometrical setting in the arriscan gui that says "16mm-4k" "Sampling 6k" - "Size [px]: 4096x2520" (that 4k pixel resolution fits proportionally within the super 16mm 2k res of 2048x1260). The pitch setting for 16mm-4k does change but I'm not sure how much until I actually see it scan.

Link to comment
Share on other sites

  • 2 weeks later...

... the post facility is offering a 4k di for marginally more than the 2K we planned on..

 

 

I for one would love to see hi-rez screenshots side by side of the same frame, one at 2k and one at 4k.

 

As an ancillary note, I recently got some Ultra 16mm back from Rob who's commented above, and I could only afford the uprezzed SD scans even though he was very flexible about price (I'm not complaining, he wsa super-helpful and took a lot of time explaining the theory and the options available).

 

Even at that low resolution, the difference between 50D in daylight and 500T in low-lit living room was very noticeable and marked. My micro-budget situation also probably meant that the 500T was not up to snuff ( early vision, not 2 or 3, uncertain history.)

 

The 50D looked really good, but my subjective impression was that if it was done at 2k I think there was more detail in it.

 

However, with the marginal 500T, I think it would have just highlighted the seriously grainy images and not helped much.

 

I think it makes sense to keep in mind accumulated resolution, (not sure it's the right name) -I think it means that film doesn't really have a fixed resolution because of the random placement of the grains.

 

So although the single frame resolution might be considered, say, 2000x3000, frame #1 might have a grain at 350.4/1000, frame # 2 at 350.0/1000, and then frame #3 maybe at 350.7/1000.

 

If all these random grain placements are in the image, the brain may merge the conveyed information across the succession of frames and you have a higher aggregated resolution than implied by the rough size of the grains.

 

I read somewhere that 80% of our brains are dedicated to processing visual information. So the tiniest, tiniest differences may be meaningful.

 

I wonder what two of the three last Best Picture Oscar winners, (Wrestler, Hurt Locker--both S16mm) were scanned at?

Link to comment
Share on other sites

  • Premium Member

Hi Alain,

The Hurt Locker was a 2K scan for sure (I know the guy who scanned it). And likely the Wrestler as well. I'm fairly certain there's never been a S16mm feature scanned at 4K. I'm not saying it's a bad idea but I agree with the others that there wouldn't be a whole lot more gained. To me the best path for S16mm (buget permitting) would be to scan oversampled at 4K minumum which should be beyond Nyquist and scale down to 2K on the output. We'll definitely be doing some tests at Cinelicious in the coming months and I'll try to post some stillls.

 

-Paul

Link to comment
Share on other sites

I find it odd that someone would be considering a 4K transfer of S16 when most 35mm scope productions, even 65mm 5-perf. are only scanned at 2K. You can almost count on one hand the number of 4K movies released, ever. Same thing with contact/optical prints.

 

2K is far better than the method that existed before: the optical blow-up.

 

 

 

Maybe for a short film it would make sense, or a movie shot outdoors predominantly on '01, or on under-200T films. You really are just sharpening and blowing up grain with 4K scans. Would it cut down on artifacts for transmission? I notice that HDTV has a terrible time with 16mm grain. Watch something like "Friday Night Lights" (Fuji I think mostly 500T) to see what I mean.

 

"Crash" (I guess just the first season, the Starz TV show not the movie) was S16. I would assume the daylight exteriors were '01 or the Fuji 64D. Those look pretty sharp at 1080P.

 

 

I'd like to see a 4K scan, personally, from S16, I just wouldn't want to have to budget for it ;)

Link to comment
Share on other sites

  • Premium Member

If all these random grain placements are in the image, the brain may merge the conveyed information across the succession of frames and you have a higher aggregated resolution than implied by the rough size of the grains.

 

Yes, that's how it works. Lowry Digital has been able to get higher resolution from film material by combining data from successive frames.

 

Or just look at a freeze frame -- you'll notice that the grain pops up and resolution drops because you're looking at just one grain pattern. That's why smart editors will try to find two frames in a row with no motion between them, and instead of freezing, rock the two.

 

 

 

-- J.S.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...