2K scan of Pro8/12 stock (s8 DI research project, Chile)

[sarmoti]

I wonder why TV spots, music videos and features that go through a D.I. process aren't shot on reversal then.

[sarmoti]

Oh, there's no 35mm K40 stock, hey maybe I should open up a company in Burbank called Rev35mm. We'll take apart K40 cartriges and superglue the film together to make 35mm K40 cans.

Sorry, just being sarcastic

[Scotness]

Congrats on the project .I spoke with the film unit in New Zealand and a couple of points.
1. A still photo is not the same as a moving image(in regard to the jpeg image posted).
2. super 8 neg to 35mm may have no advantage over k-40. In fact a k-40 2k scan will probably look better than any negative stock if going back to film (and cheaper).
It would be nice to see a test of k-40 s full potential.I have spoken with a number of post houses and all say k-40 will beat any negative(if going back to film)

Probably in terms of finest grain - but K40 has that distinctive look of lots of blacks and no latititude - you'll need a negative stock to over come that - so it's all subjective preferences really - could you imagine if they release 100D in Super 8 how good that will look in bright sunlight

The size of the files and managing/editing them will be a major issue for any 2k transfer process - but even scanned at 2k and then down sampled to SD it will probably look better than scanning at SD due to (I assume) the benefits of oversampling and the higher accuracy of the computer downsampling algorithms compared to the accuracy of the optics at SD resolution.

All in all though that still frame is amazing - it looks like a normal photo - and not Super 8 at all - but you'll still have the huge DOF of Super 8, and stability issues to overcome before you could think it looks like other film gauges -- I guess then you'd say it's Super 8 at another level - probably it's best - but not a replacement for 16.

Scot

[sarmoti]

100D /12 stock IS available... from Pro8mm, they package the negatives for Kodak anyways so whether you're buying it from them or Kodak directly you're getting the same thing. And Daniel's still is shot on that stock.

[Daniel]

Hello...

100D /12 stock IS available... from Pro8mm, they package the negatives for Kodak anyways so whether you're buying it from them or Kodak directly you're getting the same thing. And Daniel's still is shot on that stock.

May be it is only a typo error... but I prefer to avoid confusion... the frame uploaded was shot through Vision2 100 Tungsten balanced stock, in Pro8mm finishing, code Pro8/12.

could you imagine if they release 100D in Super 8 how good that will look in bright sunlight

May be Scotness was refering to the announced 100 Daylight reversal reversal color stock (Ektachrome identical to EPP still photography stock).

Regards,
Daniel

[sarmoti]

I messed up I meant 100T /12 (5212/7212), I just got the D from Scott's text and didn't give it a thought. Maybe he is referring to another stock but I think he's talking about a negative and the only one that comes to mind is the Vision 2 100T /12.

[alan doyle]

As far as I understand, Kodak outsources their negative packaging to Pro8mm so any issues you have with Pro8mm negs will probably be the same as what you get from Kodak.

sorry this not true,200 neg and 500 neg...are finished and packaged in
france..kodak has used pro8mm for testing small batches,of different stocks..for evaluation only.

my experiance is based on a recent BBC project in which the cheapskate production manager purchased pro8 200asa,so i shot this plus kodaks version...pro8s rolls were more unstable a lot more dirt was visable and i had 2 jams out of seven rolls..
if it ever becomes a fact about pro8 working with for kodak ,it will be a disgrace
because of kodaks complete lack of interest,in super 8,i have decided to use fuji 16mm ...and fuji stills film,which is great...

[Daniel]

congrats,really lovely work..keep going with it...
am i correct in thinking,that the spirit and shadow telecine machines use 16mm optics with a skid plate assembly,which then requires software to do a digital xy zoom into your top quality image..
it would be very interesting,to compare cintels millenium 2 which has a real optical s8 lens at 2 or 4k...with the spirit...
also considering the pro 8 companys poor perforation and dirt issues,with film stocks...
your tests clearly show a quality,i feel not seen on this site before..

WOW 8O

Hello,

Thanks.. I was looking in my source documents before answering this.
I have uploaded some paper documents that I received at Philips/Thomson booth during my NAB 2001 visit. That year they were showing the Shadow and the Spirit with the super8 option.

http://www.filmshooting.com/scripts/gal ... m/thomson1
http://www.filmshooting.com/scripts/gal ... m/thomson2
http://www.filmshooting.com/scripts/gal ... m/thomson3

So the third document (thomson3) explains that " ... the second mechanical element is an adapter plate which mounts in the current Super 16mm gate to provide precise guidance of Super 8mm film. The software with the option adds a separate preset for framing Super 8mm and modifies the zoom parameters to match the smaller image area. The transport characteristics are also modified for the lighter stock, allowing the same cores and film path to be used..."

In the first document (thomson1) there is a picture of the adapter plate, and some specifications including this interesting parameter :
" Optical resolution measured in LP/mm on film equal to Super 16mm "

My current research project only consider to scan the definitive footage through Spirit (1) Datacine at 2K data resolution and 10 bits log. quantization. I do not have budget enough to go further and compare different scanning technologies... it is really an interesting point, that I have considered... but it is a too much expensive research for us at that point. May be if our results are good enough, we would be able one day to extend the research using the same original super8 footage over different scanning tech...

Around this interesting scanning technologies theme, I share the thinking of many film post-production specialist that argue that CCD based scanners (Northlight, Spirit, Shadow...) gives "better results" in terms for smaller gauge (16mm ...) than CRT/PMT technology... "Better results" understood as a quieter image, with grain structure beeing less apparent.
An other important point to consider is the kind and degree of diffusion of the light source within the light path into the scanner... that "simple" parameter makes a big difference in how much you can see scratches and in some maneer grain structure also. (this last parameter is pretty much identical to what happens in a diffused v/s condenser photographic enlarger).

As far as I know.. today super8mm can be scanned to data resolutions (2K +) in a Millennium (1 and 2 with specific film gate), Spirit (1 and 2 with adapter to S16mm film gate), DataMill (with specific film gate), and some Oxberry Cinescan also.
What would be very very nice would be if one day we could see that the new Arri-Scan (based on LED lighting) and/or Northlight are also able to scan super 8mm footage. But anyway what is currently available is very good enough to the majority of film work beeing done in super8 for theatrical applications.

Regards,
Daniel

[Daniel]

Congrats on the project .I spoke with the film unit in New Zealand and a couple of points.
1. A still photo is not the same as a moving image(in regard to the jpeg image posted).
2. super 8 neg to 35mm may have no advantage over k-40. In fact a k-40 2k scan will probably look better than any negative stock if going back to film (and cheaper).
It would be nice to see a test of k-40 s full potential.I have spoken with a number of post houses and all say k-40 will beat any negative(if going back to film)

Hello,
I agree with your point nº1. In fact seeing the 35mm projection, we did see less apparent grain structure than when we were seeing the 2K dpx file on the computer screen. We understand that this is due to the fact that the grains structure is dynamic.
Concerning your point nº2... I rather agree with Scot opinion, as I consider that an advantage of the vision2 color negative line over K-40 is its extended exposure latitude, plus the fact that you can switch between different light sensitivity stock using the different stock of the vision2 line (with almost similar color and gamma characteristics) giving to the filmmaker flexibility. Considering here Pro8/12, Kodak 17, Kodak 18 for instance if we are talking tungsten balanced stock.
If your point is that K-40 is a lower grain stock... than Pro8/12 or Pro8/45... I am not 100% sure that this is really true... but I can't argue that point, and I prefer not to enter to the traditional debate color negative v/s K-40.
What is really true however... is the fact that it is not so easy to scan reversal film ! Some of the top quality scanners like DataMill only can handle a dynamic range up to 3.3 ! So if one pretends to scan reversal... there is some considerations to take in account in order to evaluate what happens with the image information that is precisely within the 3.3 to 4 (aprox) density range of the reversal frames.
However it is likely that this actual scanning limit should be overcome during the next year(s) or decade(s) (?).. if we check out what is the current reality of the still photography top-range scanners.
Regards,
Daniel

[sarmoti]

Daniel, have you gone through the compositing tests? If so what color difference keying (blue/green screen) software will you use? Ultimatte, Primatte, Discreet Master Keyer, etc... I'd be interested to know if you can preserve shadow/transparency detail in the matte without screening through the grain as well on the overall image.

[Scotness]

I messed up I meant 100T /12 (5212/7212), I just got the D from Scott's text and didn't give it a thought. Maybe he is referring to another stock but I think he's talking about a negative and the only one that comes to mind is the Vision 2 100T /12.

Yeah sorry - I'm partly to blame anyway - somewhere along the line I thought Daniel's frame was 200T - but I did mention 100D not T because I've heard Kodak have been developing it and because shot outdoors with heaps of light I assume the grain size will be even less noticeable then 100T shot indoors under artifical light -- and plus I shoot alot outside so I'm always interested in the D stocks first!

Scot

[Daniel]

Daniel, have you gone through the compositing tests? If so what color difference keying (blue/green screen) software will you use? Ultimatte, Primatte, Discreet Master Keyer, etc... I'd be interested to know if you can preserve shadow/transparency detail in the matte without screening through the grain as well on the overall image.

For this early test, I checked the response of the footage that was shot using Rosco Digicomp fabric materials with Pro8/12 and Kodak 17.
The subject was the actress medium shot and Digicomp background.
For tis earlier test I used Dicreet Keyer from Combustion. It worked very well.. with much easier and considerably better result than what you can do with DV25 progressive footage. This was also recorded back to 35mm film and seen on 35mm projection. Anyway I had problems with the fine details of the hair... so for the definitive footage... lighting was improved, using in the backlight Rosco Calcolor complementary filter plus a significative distance between subject and Digicomp background (a bit more than 10 meters). For the definitive footage... "the offline" or "draft" digital compostion work will be done through Combustion, in order to select the shots. Then we plan to use Discreet Inferno keying tools to get high quality digital compositions.
Best Regards,
Daniel

[alan doyle]

My current research project only consider to scan the definitive footage through Spirit (1) Datacine at 2K data resolution and 10 bits log. quantization. I do not have budget enough to go further and compare different scanning technologies... it is really an interesting point, that I have considered... but it is a too much expensive research for us at that point. May be if our results are good enough, we would be able one day to extend the research using the same original super8 footage over different scanning tech...
Daniel

hi ya, i think i can help you here,as cintel have already promised to help me scan super 8 at hi def...2k and 4k...
but i have not got around to taking them up on the offer...
also the people at northlight,already have some of my footage,to scan..
the main problem being transport and the lens assembly,they use would involve a rather big digital zoom..
the guys did say if i had the money they could custom build me a northlight super8 scanner..apart from the cost the speed of the northlight is an issue..
any way daniel, let me know if i can help...

[Daniel]

Hello Alan.

hi ya, i think i can help you here,as cintel have already promised to help me scan super 8 at hi def...2k and 4k...

Well, thank you Alan, but at the moment I would not need your help. .. it is cool to see that people have liked our project and have good intentions in helping..Thank you anyway.

Actually I am only interested in complete the reserch project announced at the beginning of this thread. Just that. And it is a lot of work and dedication. Once the project is finished and released in 35mm (early June 2005) it is possible for me to think in expanding the research using the same super8 media over different scanning technolgies, film recording technologies, different 35mm intermediate stocks, differents 35mm print stock...

I will now leave this internet-thread for some weeks as we now continue the film shooting plan (landscapes shot) in non urban zones in the south.
Bye Bye and thanks again to everybody.
Regards,
Daniel

[mattias]

I guess S8 can be 1/3 cheaper than 16mm if you just get some reversal stock and transfer it with a film chain device. But then you could do the same for 16.

except all reversal 16mm stocks are super expensive.

/matt