A Few Words About While we wait for A few words about...™ Lawrence of Arabia -- in 4k/UHD Blu-ray (9 Viewers)

[Persianimmortal]

Our Blu-Rays will probably last longer than the studios digital masters. If they're that foolish not to invest in digital preservation, I'll sell them back my collection in 25 years. At least they'll have a 1080p copy.

Isn't this where the new Archival Disc optical format steps in? With up to 1TB of storage per disc, studios can easily store and preserve copies of the digital source master. As long as the discs aren't stored carelessly, they'll last for up to 100 years or more, by which time they would have been transferred to other, better, media.

 

[Michel_Hafner]

Those are some big numbers you are handling - moving or still pictures?

Stills.

 

[Moe Dickstein]

This thread is a general purpose hangout.

 

[Bobby Henderson]

Isn't this where the new Archival Disc optical format steps in? With up to 1TB of storage per disc, studios can easily store and preserve copies of the digital source master.

The materials from just one Hollywood studio movie can consume many Terabytes of storage space, be it on LTO tapes, hard drives, optical discs, etc.

The imagery from video cameras can require a huge amount of storage capacity, especially if it is shot in RAW formats, 4:4:4 color depths and resolutions well above 1080p HD. A lot of movies are shot with the Arri Alexa in 2.8K resolution and ARRIRAW format. Cameras from Red can do 4K, 5K and even 6K. When movies are shot on 35mm or 65mm film the original negative is usually preserved. Do studios preserve the original video camera data?

Movies have a great deal of proprietary data produced in 3D modeling/animation programs, effects compositing programs, color grading software, etc. A "digital intermediate" or "digital master" for a modern movie is really the final container for a lot of separate audio-visual "stems." If a movie studios sees some glaring error in a movie several years old and would like to correct it in the best manner possible it would be best to dig up those original assets and software and re-render. Doing so might be impossible.

There is a CG scene in Jurassic Park that drives me nuts a little bit. It's the scene where Jeff Goldblum is waving the road flare at the T-Rex and takes off running with the T-Rex chasing him. It's a great scene. I remember my mouth falling open the first time I saw the movie, muttering "how the hell did they do that?" Unfortunately the clip has a serious flaw. There is a very noticeable color shift where the live action plate and practical T-Rex gets changed for the CG T-Rex. The edit occurs right at the roar. That color shift can be fixed in various ways, but the best results would be achieved if one could manipulate the original visual effects data. The trouble is that work was done in versions of Alias PowerAnimator and Softimage more than 20 years old -running on Silicon Graphics IRIX-based workstations (a now dead OS). Some custom-engineered software may have been involved too. Another fall-back: scan in the film-out elements and edit that.

Regarding the previous topic of 8K video formats, I certainly don't mind the electronics industry pushing boundaries in that direction. However, I would like to see a shift in priorities to improving lens technology. And if they have to improve camera sensors further, why not do more in the direction of developing global shutter image sensors rather than upping resolution?

We're reaching points of diminishing returns in the megapixel race. And those limits have been surpassed in certain consumer electronics categories -like smart phones claiming 8, 10, 12 or even 41 megapixel resolution from microscopic image sensors and equally microscopic (and diffraction prone) lenses. It's just plain ridiculous.

Lots of people thought DSLR cameras with digital HD video capability would "democratize" movie production, making things much more affordable for a lot of budding movie makers. Unfortunately still camera lenses, even ones costing thousands of dollars, have fatal flaws when used for shooting video. The image grows/shrinks if you rack focus. If you zoom, the lens won't zoom entirely straight. I have a $2500 70-200 f/2.8L "II" lens from Canon that produces outstanding still image quality. It only works for video if you don't mess with the focus or zoom. "Cine" lenses overcome those problems, but at very high costs. The DSLR camera bodies have sensors rife with all their own issues.

Obviously the technology industry has a LOT of improvement work to do regarding "digital" and its replacement of film.

 

[Persianimmortal]

The materials from just one Hollywood studio movie can consume many Terabytes of storage space, be it on LTO tapes, hard drives, optical discs, etc.

The imagery from video cameras can require a huge amount of storage capacity, especially if it is shot in RAW formats, 4:4:4 color depths and resolutions well above 1080p HD. A lot of movies are shot with the Arri Alexa in 2.8K resolution and ARRIRAW format. Cameras from Red can do 4K, 5K and even 6K. When movies are shot on 35mm or 65mm film the original negative is usually preserved. Do studios preserve the original video camera data?

I certainly understand that there are many components that make up the final shot, and some of these original components may take up a huge amount of storage. Still, I'm assuming Archival Disc will be roughly the same physical size as BD or DVD, hence for each movie, even if we're talking 100-200TB of data, that equates to a shoebox's worth of Archival Discs. Sealed and stored properly, they should last for many decades. It should open up the option for studios to preserve all the elements they need in a compact and easy-to-handle format. I wonder what archivists like RAH think of this?

 

[Robert Harris]

I certainly understand that there are many components that make up the final shot, and some of these original components may take up a huge amount of storage. Still, I'm assuming Archival Disc will be roughly the same physical size as BD or DVD, hence for each movie, even if we're talking 100-200TB of data, that equates to a shoebox's worth of Archival Discs. Sealed and stored properly, they should last for many decades. It should open up the option for studios to preserve all the elements they need in a compact and easy-to-handle format. I wonder what archivists like RAH think of this?

My position is simple, and stated previously. I'm a huge fan of digital media, especially as a working solution. Archival data has a shelf life of eternity or seven years -- whichever comes first.If I'm spending half a year saving a film, I love working with all things digital.And then recording back to film as an archival asset.RAH

 

[Peter Apruzzese]

Robert,

Have the archives experimented/tested this idea: store the digital data as a photographic image on 35mm polyester film (similar to how Dolby Digital and SDDS tracks are put on theatrical prints)? If they have tried it, how did it work out?

 

[Bobby Henderson]

That would be an impractical approach, as opposed to simply storing images on film.

A single 2048 X 852 pixel "2K" 'scope image with 8-bits per pixel per channel will consume nearly 5MB uncompressed. LZW compression might reduce the file size by half. Storing the digital data would require every bit (rather than byte) to be reproduced in an analog, printed bit map, 20 million bits of data in the case of a 2.5MB LZW compressed image. The printed bit map would have to be large enough to be read reliably without errors, so just one frame of imagery could easily require multiple frames of 35mm film.

And then there's all sorts of issues regarding the computer-based hardware and software used for reading in the imagery. What works today might not work 20 or 30 years from now. You mentioned SDDS. A lot of new movies still being reproduced on 35mm release prints have SDDS audio tracks on the prints. The funny thing is Sony hasn't made any SDDS film readers or sound processors in over 10 years. There's probably not many theaters left which have operational SDDS gear, especially here in the US. Most of it went to the landfill.

 

[Stephen_J_H]

Robert,

Have the archives experimented/tested this idea: store the digital data as a photographic image on 35mm polyester film (similar to how Dolby Digital and SDDS tracks are put on theatrical prints)? If they have tried it, how did it work out?

Bobby has also pointed out one of the issues with this approach. Even if it were possible to output a digital bitstream to 35mm film, who's to say that readers will still exist in 10-20 years time to read such data? What about after that? One of the reasons for output to film is that film and film projection is robust. It's been around for over 100 years, and it's a lot easier to get a film projector up and running than a 20 year old computer.

 

[seangood79]

My position is simple, and stated previously. I'm a huge fan of digital media, especially as a working solution. Archival data has a shelf life of eternity or seven years -- whichever comes first.If I'm spending half a year saving a film, I love working with all things digital.And then recording back to film as an archival asset.RAH

My big fear is some future studio executive gets tired of paying for the storage of film, and orders all elements digitized and junked. It's happened before with Fox's Great Nitrate Purge.

 

[Dick]

Archival data has a shelf life of eternity or seven years -- whichever comes first.

That's truly one of the funnier (and wittier) remarks I've read in ages. Thanks, Robert!

 

[Peter Apruzzese]

That would be an impractical approach, as opposed to simply storing images on film.

A single 2048 X 852 pixel "2K" 'scope image with 8-bits per pixel per channel will consume nearly 5MB uncompressed. LZW compression might reduce the file size by half. Storing the digital data would require every bit (rather than byte) to be reproduced in an analog, printed bit map, 20 million bits of data in the case of a 2.5MB LZW compressed image. The printed bit map would have to be large enough to be read reliably without errors, so just one frame of imagery could easily require multiple frames of 35mm film.

And then there's all sorts of issues regarding the computer-based hardware and software used for reading in the imagery. What works today might not work 20 or 30 years from now. You mentioned SDDS. A lot of new movies still being reproduced on 35mm release prints have SDDS audio tracks on the prints. The funny thing is Sony hasn't made any SDDS film readers or sound processors in over 10 years. There's probably not many theaters left which have operational SDDS gear, especially here in the US. Most of it went to the landfill.

Thanks, Bobby. I should have remembered that the DD tracks were highly compressed and that actual uncompressed picture data would take a lot more space.

 

[OliverK]

Thanks, Bobby. I should have remembered that the DD tracks were highly compressed and that actual uncompressed picture data would take a lot more space.

Still it is a very good idea to have a visual record of the data - if it was visual it should always be recoverable. The question of course remains: How much space would be needed?

 

[Joe Bernardi]

I've enjoyed reading this thread, but can't wait until we get -
While we wait for A few words about...™ Lawrence of Arabia -- in 4K

 

[Bobby Henderson]

Lawrence of Arabia is available from Sony's 4K movie store to buy or rent, but you have to download it to that $699 round-shaped thingie from Sony. We're probably going to be waiting a while for 4K Blu-ray.

I've heard some complaints about the 4K DCP of Lawrence of Arabia for commercial movie theaters. Digitally restoring a movie and creating a new master is one very important thing. The encoding work for various digital formats, be it Blu-ray, commercial cinema DCP, cable/satellite broadcast, online streaming, etc. is just as important. Comments I've read in the Film-Tech forums suggest movie studios should be doing a better job with the JPEG 2000 encoding work for classic repertory films in DCP format.

Still it is a very good idea to have a visual record of the data - if it was visual it should always be recoverable. The question of course remains: How much space would be needed?

It's easy to figure out data storage requirements in terms of uncompressed image data (lossless image compression generally runs in the 2:1 ratio). You take the image dimensions, such as 2K 'scope (2048 x 852) or 2K flat (1998 x 1080), multiply those numbers for a total pixel count of the image then multiply that number again by how many bits of data are stored in each pixel. A plain RGB image with 8-bits stored in the Red, Green & Blue channels will have 24 bits per pixel. 16-bit channels will double that number. Storage requirements are pretty big just for 2K. Quadruple those numbers for 4K.

A single 4K frame from Lawrence of Arabia might contain more than 80 million bits of data with lossless compression applied and only in basic 8-bit depth. Uncompressed, we're probably talking a video bit rate around 3.8 gigabits per second. That's a lot of data spots to print on some film. It's obviously a lot more practical to just print the image on film instead.

 

[Dee Zee]

May I ask, what is the best biography of Sir David Lean? Other recommended books about Lean are much appreciated.

 

[Robert Harris]

Kevin Brownlow, Stephen Silverman. Very different, both superb.

RAH

 

[Dee Zee]

Kevin Brownlow, Stephen Silverman. Very different, both superb. RAH

Thank you Robert.

 

[Ken Volok]

If I had the resources I'd print out on paper each frame of all my films fror preservation purposes. I'd go as far as having the sound transcribed on disc if I could. While I enjoy as an audience and benefit as an artist by the digital opportunities; I still prefer and trust analog more.
I feel very fortunate to have had the opportunity to work with film both in the camera and on the moviola. Not to boast but I got to the point, if I was working MOS I could hold a strip of film up and know right where to cut. Heck I even love the smell of film.

 

[Josh Steinberg]

I've heard some complaints about the 4K DCP of Lawrence of Arabia for commercial movie theaters. Digitally restoring a movie and creating a new master is one very important thing. The encoding work for various digital formats, be it Blu-ray, commercial cinema DCP, cable/satellite broadcast, online streaming, etc. is just as important. Comments I've read in the Film-Tech forums suggest movie studios should be doing a better job with the JPEG 2000 encoding work for classic repertory films in DCP format.

Somewhat off-topic (other than that the films share a certain restoration guru), AMC theaters is going to be showing "Spartacus" in May for a day. On one hand, excited about the potential to see it on the big screen (even if it's just a multiplex and not a movie palace). On the other hand, if it's from a similar source to what that Blu-ray was made from, yikes.