"Club Penton" - Ask questions to Hollywood insider "Penton-Man"

[Robert Harris]

Originally Posted by Oliver K

Indeed you managed to do just that

I am not so much shocked about additional care for the opticals but about the 2k scan for two prestiguous 65mm productions.

That sounds very strange to me when recently 65mm productions that made it to Blu-Ray have moved up to an 8k scan, including South Pacific from Fox. Even Patton was scanned in 4k before the 'wizard' with the fancy software had a go at it.

Fwiw I think that a 4k scan that gets reduced to BR resolution could already yield good results especially if 'only' a Blu-Ray is the intended end product but with 4k projection gaining a strong momentum in the markt I would prefer for work to be done in 8k to really extract all of the relevant detail from the OCN and to also allow for 4k projection where I expect 65mm originated movies to shine. BTW: I am still waiting for 4k to make it into consumer projectors with higher on/off CR and I have a little hope that the introduction of 3D to the home theater market might help with that

This sounds quite accurate. Back around 1997, our team went through a number of extremely vigorous tests working within the digital domain, scanning at various resolutions and then recording back out to film. The tests were performed in VVLA.

The bottom line, and I can reveal this now that a reasonable period of time has passed, was that lowering the resolution of the scan was extremely helpful as a workaround when dealing with optical flaws as well as emulsion damage. The final results, recorded back to VVLA were extremely successful, with very pretty images. There is a direct correlation between scanning resolution and the ability of digital software to attenuate both unwanted grain as well as small emulsion problems. Taking this path a bit further, we also proved that recording at a lower resolution also played a huge part toward the creation of a far more pleasant image.

Today these files can be ported to DCP and yield wildly successful images, even on the largest theatrical screens.

RAH

[Doctorossi]

Originally Posted by Robert Harris

The bottom line, and I can reveal this now that a reasonable period of time has passed, was that lowering the resolution of the scan was extremely helpful as a workaround when dealing with optical flaws as well as emulsion damage.

Were these lower-resolution scans contrasted with various levels of down-conversion from higher-resolution scans when evaluating this effect? I would suppose that this quality could be achieved in processing, after producing an archival scan at the greatest resolution available.

Further to this question... On a philosophical level, how do you define an "optical flaw"? Which production artifacts should be obscured from view in the final product? Audiences have different interests and purposes in watching old(er) films. Some wish to be seamlessly immersed "in" the story, while some love the opportunity to witness an historically-accurate reflection of the state of yesterday's art, imperfections and all. These "optical flaws" are the visual equivalent of John Bonham's squeaky bass pedal.

[cjamescook]

Originally Posted by Robert Harris

This sounds quite accurate. Back around 1997, our team went through a number of extremely vigorous tests working within the digital domain, scanning at various resolutions and then recording back out to film. The tests were performed in VVLA.

... The final results, recorded back to VVLA were extremely successful, with very pretty images...

Psst...What's VVLA?

Googling only suggested Victor Valley Libertarian Alliance.

-Jim, sitting in the cheap seats

[Robert Harris]

Originally Posted by Doctorossi

Were these lower-resolution scans contrasted with various levels of down-conversion from higher-resolution scans when evaluating this effect? I would suppose that this quality could be achieved in processing, after producing an archival scan at the greatest resolution available.

Further to this question... On a philosophical level, how do you define an "optical flaw"? Which production artifacts should be obscured from view in the final product? Audiences have different interests and purposes in watching old(er) films. Some wish to be seamlessly immersed "in" the story, while some love the opportunity to witness an historically-accurate reflection of the state of yesterday's art, imperfections and all. These "optical flaws" are the visual equivalent of John Bonham's squeaky bass pedal.

By "optical flaws," I referring to age and handling related problems that have become a part of the image, as opposed to lens flare, etc. And yes, various incarnations of down (and up) conversion were viewed.

[Robert Harris]

Originally Posted by cjamescook

Psst...What's VVLA?

Googling only suggested Victor Valley Libertarian Alliance.

-Jim, sitting in the cheap seats

With apologies...

VistaVision Large Aperture

[4K2K]

http://blog.digitalcontentproducer.c...hines-for-fox/

Surely digital effects techniques like digital wire/rig removal, or grain reduction, etc. is not restoration - restoring it to how it was originally - but is updating old films with current digital effects?

[Robert Harris]

Originally Posted by 4K2K

http://blog.digitalcontentproducer.c...hines-for-fox/

Surely digital effects techniques like digital wire/rig removal, or grain reduction, etc. is not restoration - restoring it to how it was originally - but is updating old films with current digital effects?

Grain reduction is only a part of film restoration when dealing with dupes of different generations, whereby massaging the grain (without losing resolution) helps toward an overall cohesiveness of the image.

[Doctorossi]

Originally Posted by Robert Harris

By "optical flaws," I referring to age and handling related problems that have become a part of the image, as opposed to lens flare, etc.

And it sounds like you're talking about image texture on the kind of fine level that, in my own experience with electronic scanning, I know can be hugely impacted by the subtleties of physical alignment.

In light of this, I would be very curious to look into what kind of image could be produced, in an algorithmic software pass, by interpolating multiple aligned scans of differing resolutions.

[Robert Harris]

Originally Posted by Doctorossi

And it sounds like you're talking about image texture on the kind of fine level that, in my own experience with electronic scanning, I know can be hugely impacted by the subtleties of physical alignment.

In light of this, I would be very curious to look into what kind of image could be produced, in an algorithmic software pass, by interpolating multiple aligned scans of differing resolutions.

From tests performed on The Alamo, while a record with higher resolution is helpful, additional records (scans) of far lower resolution pull the overall quality down measurably. Even with the best in the business dealing with these more complex issues, which IMHO is Reliance (Lowry), the final result is acceptable at best when combining resolutions of disparate quality, and is always below the quality of the highest resolved image in the mix. In our particular case our elements were a VVLA image derived optically from 70mm and a 480i image.

[Jeff Kleist]

Most filmmakers I've talked to WANT them to take out unwanted wires and things, because they were never meant to be seen in the first place.

I mean, I always found it hilarious it took lucas 3 tries to put a blade on Obi-Wan's lightsaber in the shot that was missing it

[Doctorossi]

Originally Posted by Robert Harris

From tests performed on The Alamo, while a record with higher resolution is helpful, additional records (scans) of far lower resolution pull the overall quality down measurably. Even with the best in the business dealing with these more complex issues, which IMHO is Reliance (Lowry), the final result is acceptable at best when combining resolutions of disparate quality, and is always below the quality of the highest resolved image in the mix. In our particular case our elements were a VVLA image derived optically from 70mm and a 480i image.

Wow- 480i. I wasn't imagining anything that low! I'm theorizing that there might be some edifying results from performing some sort of digital boolean operation on, say, a 6k, a 4k and a 2k scan of the same film- not so much for the difference in spatial resolution, per se, as for the simple variance in scanning registration. This would be something conceptually similar to over-sampling, but with a method that might more effectively mimic a product of the analog (optical) domain.

I'm not quite sure how clear I'm managing to be...

[KubrickFan]

Originally Posted by Jeff Kleist

Most filmmakers I've talked to WANT them to take out unwanted wires and things, because they were never meant to be seen in the first place.

I mean, I always found it hilarious it took lucas 3 tries to put a blade on Obi-Wan's lightsaber in the shot that was missing it

And the lightsaber still doesn't look right (that and a lot of other things, but I won't go into that).
I'm kind of torn by this. Part of me (the purist) really wants to see it as it was, warts and all. If that (in the case of Star Wars) means seeing matte boxes on almost every ship, then so be it. The other part of me simply wants a movie the best it can look. If they can clean up the special effects (not replace it with CG, mind you) so that it looks as new, how can you be opposed to that? The Final Cut of Blade Runner, for example, looks extraordinary, and the special effects look fantastic.

[Doctorossi]

Originally Posted by KubrickFan

The Final Cut of Blade Runner, for example, looks extraordinary, and the special effects look fantastic.

Itself being, to be clear, a product of both of the approaches you describe (thoroughly cleaned up, re-composited existing FX and newly created replacement FX).

[captveg]

Originally Posted by micks_address

i'm pretty sure the cover art on the disc i ordered said directors cut?

That's old art work released in 2007-ish. The actual release coming next week has been confirmed as the Theatrical Version.

[Doctorossi]

I watched my Wizard of Oz disc last night. Does anyone have any insight into the decision to windowbox the credits? No explanation in the restoration documentary or any of the online reviews I've looked at.

[Jeff Kleist]

Re-compositing existing elements I find to be my favorite fix of all. Everything is original, it just takes out alll the optical warts and generational loss. I really wish that Disney would do this with Tron. Maybe when Legacy makes them a fortune (that is, assuming that the theaters get a clue and unjack their prices, How to Train Your Dragon got hurt bad by it)

I watched my Wizard of Oz disc last night. Does anyone have any insight into the decision to windowbox the credits? No explanation in the restoration documentary or any of the online reviews I've looked at.

To make sure all the edge text makes it in. I believe this was often done on the actual prints because projection framing isn't constant so everyone's name makes it in

[captveg]

Windowboxed credits are 99% of the time done to assure that TVs with overscan do not crop names, thereby preventing any infringement on any guild/association/etc. agreements.

Of course, when in a 1080p 16x9 image (1.33:1 w/ side bars) overscan is very unlikely to occur. But this master was also used for the new 4x3 DVD.

[SpaceDog]

Originally Posted by Jeff Kleist

Re-compositing existing elements I find to be my favorite fix of all. Everything is original, it just takes out alll the optical warts and generational loss. I really wish that Disney would do this with Tron. Maybe when Legacy makes them a fortune (that is, assuming that the theaters get a clue and unjack their prices, How to Train Your Dragon got hurt bad by it)

Are you sure HtTYD didn't get hurt by Dreamwork's consistant mediocrity in comparison to Pixar? I saw the ho-hum MvA last year in theater, but save the original Shrek, I tend to wait for home video for most Dreamworks product, and most Dreamworks product tends to reinforce my decision in this regard.

[Vincent Pereira]

I'm guessing the 480I material Robert is referring to would be the old LaserDisc transfer of the long version of THE ALAMO. It's amazing that the 70mm print used for that transfer- which aside from being literally one-of-a-kind was supposed to be in virtually pristine condition back in the early 1990s- was treated so carelessly and allowed to fall apart.

Vincent

Originally Posted by Doctorossi

Wow- 480i. I wasn't imagining anything that low! I'm theorizing that there might be some edifying results from performing some sort of digital boolean operation on, say, a 6k, a 4k and a 2k scan of the same film- not so much for the difference in spatial resolution, per se, as for the simple variance in scanning registration. This would be something conceptually similar to over-sampling, but with a method that might more effectively mimic a product of the analog (optical) domain.

I'm not quite sure how clear I'm managing to be...

[SquidPuppet]

Originally Posted by SpaceDog

Are you sure HtTYD didn't get hurt by Dreamwork's consistant mediocrity in comparison to Pixar? I saw the ho-hum MvA last year in theater, but save the original Shrek, I tend to wait for home video for most Dreamworks product, and most Dreamworks product tends to reinforce my decision in this regard.

I thought How to Train Your Dragon was fantastic. Dreamworks best so far. IMO, its in Pixars league. It seemed under promoted to me.