Film-Tech Cinema Systems
Film-Tech Forum ARCHIVE


  
my profile | my password | search | faq & rules | forum home
  next oldest topic   next newest topic
» Film-Tech Forum ARCHIVE   » Community   » Film-Yak   » James Dean Reborn in CGI for Vietnam War Action-Drama

   
Author Topic: James Dean Reborn in CGI for Vietnam War Action-Drama
Frank Cox
Film God

Posts: 2234
From: Melville Saskatchewan Canada
Registered: Apr 2011


 - posted 11-06-2019 05:34 PM      Profile for Frank Cox   Author's Homepage   Email Frank Cox   Send New Private Message       Edit/Delete Post 
I wonder how long it will be before there's no longer a need to hire and pay actual actors.

quote:
The cultural icon, who died in 1955, will return to the screen via CGI using actual footage and photos for 'Finding Jack.'

James Dean, who died in a 1955 car crash at the age of 24, is making an unexpected return to the big screen.

The cultural icon, known for Rebel Without a Cause and East of Eden, has been posthumously cast in the Vietnam era action-drama Finding Jack.

Directed by Anton Ernst and Tati Golykh, the project comes from the filmmakers' own recently launched production house Magic City Films, which obtained the rights to use Dean’s image from his family. Canadian VFX banner Imagine Engine will be working alongside South African VFX company MOI Worldwide to re-create what the filmmakers describe as “a realistic version of James Dean.”

Adapted by Maria Sova from Gareth Crocker’s novel, Finding Jack is based on the existence and abandonment of more than 10,000 military dogs at the end of the Vietnam War. Dean will play a character called Rogan, considered a secondary lead role.

"We searched high and low for the perfect character to portray the role of Rogan, which has some extreme complex character arcs, and after months of research, we decided on James Dean," said Ernst, who also produces with Golykh for Magic City Films alongside Donald A. Barton of Artistry Media Group.

"We feel very honored that his family supports us and will take every precaution to ensure that his legacy as one of the most epic film stars to date is kept firmly intact. The family views this as his fourth movie, a movie he never got to make. We do not intend to let his fans down."

While Finding Jack will be live action, The Hollywood Reporter understands that Dean’s performance will be constructed via "full body" CGI using actual footage and photos. Another actor will voice him.

Multiple Oscar nominated songwriter Diane Warren wrote the key song for the film and acclaimed composer Laurent Eyquem is on board to score the film.

Preproduction on Finding Jack starts Nov. 17, with a goal for a worldwide release on Veterans Day 2020. Magic City Films is handling the foreign sales.

The filmmakers are now hoping that the CGI technology used to bring Dean back to life onscreen could soon be deployed on other well-known figures.

"This opens up a whole new opportunity for many of our clients who are no longer with us," said Mark Roesler, CEO of CMG Worldwide, which represents Dean’s family alongside more than 1,700 entertainment, sports, music and historical personalities, including the likes of Burt Reynolds, Christopher Reeve, Ingrid Bergman, Neil Armstrong, Bette Davis and Jack Lemmon.

Added Ernst: "Our partners in South Africa are very excited about this, as this technology would also be employed down the line to re-create historical icons such as Nelson Mandela to tell stories of cultural heritage significance."

Barton added: "Now that we have closed with this iconic figure, we look forward to rapidly closing our remaining actors."

James Dean Reborn in CGI

 |  IP: Logged

Bobby Henderson
"Ask me about Trajan."

Posts: 10973
From: Lawton, OK, USA
Registered: Apr 2001


 - posted 11-06-2019 07:07 PM      Profile for Bobby Henderson   Email Bobby Henderson   Send New Private Message       Edit/Delete Post 
quote: Frank Cox
I wonder how long it will be before there's no longer a need to hire and pay actual actors.
It will be a very long time before that kind of thing happens, if it ever happens.

It's still a far less expensive and far faster, more convenient process to hire a regular actor to perform in a movie than to create, animate/perform and render a realistic looking digital version of the same person. The cost difference is gigantic if you're hiring a "nobody" and paying him around scale or modestly above it. Teams of modelers, animators and engineers are required to do the CGI work. Rendering a digital person in a believable manner takes a hell of a lot of computing horsepower.

Let's also not forget The Uncanny Valley is still a very difficult challenge to overcome when creating artificial versions of people. At best the CGI work these days only succeeds in very fleeting moments.

Right now the biggest motivations for using CGI-based versions of people in movie scenes is either bringing dead celebrities back to life or putting an actor into a visual scene too dangerous or too difficult/costly to pull off using practical methods in real life. I don't see those motivations changing any time soon.

Resurrecting dead celebrities in new movies or TV shows will get expensive. The estates of those celebrities will want to get paid. But we're not going to see widespread use of that kind of thing, at least not until the battle with the uncanny valley can be won.

 |  IP: Logged

Martin McCaffery
Film God

Posts: 2481
From: Montgomery, AL
Registered: Jun 99


 - posted 11-06-2019 08:28 PM      Profile for Martin McCaffery   Author's Homepage   Email Martin McCaffery   Send New Private Message       Edit/Delete Post 
It may come in handy in those movies where the actor dies before shooting is completed, or where he's persona non grata before the movie is released.

 |  IP: Logged

David Kornfeld
Film Handler

Posts: 24
From: Cambridge, MA/USA, USA
Registered: Apr 2017


 - posted 11-06-2019 08:37 PM      Profile for David Kornfeld   Email David Kornfeld   Send New Private Message       Edit/Delete Post 
This is absolutely, utterly, & completely disgusting. Theres no excuse for this: theres an actor out there for every occasion. Find a living one. This is grave-robbing. These people are ghouls. It's wrong on every level. And it should be outlawed, & I mean that.

Let the dead rest.

 |  IP: Logged

Brad Miller
Administrator

Posts: 17775
From: Plano, TX (36.2 miles NW of Rockwall)
Registered: May 99


 - posted 11-06-2019 08:48 PM      Profile for Brad Miller   Author's Homepage   Email Brad Miller       Edit/Delete Post 
Oh this is nothing. Just wait until the technology gets cheaper and the porn industry latches onto it. One day there will be lesbian porn starring Marilyn Monroe and Elizabeth Taylor. [Roll Eyes]

 |  IP: Logged

Bobby Henderson
"Ask me about Trajan."

Posts: 10973
From: Lawton, OK, USA
Registered: Apr 2001


 - posted 11-06-2019 10:43 PM      Profile for Bobby Henderson   Email Bobby Henderson   Send New Private Message       Edit/Delete Post 
There is already crude 3D model porn. So, yeah, CGI celebrity porn fakes are probably a given in the future. And there's no doubt the videos will be creepy as hell. The Uncanny Valley will be delivering some real nightmare fuel.

 |  IP: Logged

Marcel Birgelen
Film God

Posts: 3357
From: Maastricht, Limburg, Netherlands
Registered: Feb 2012


 - posted 11-07-2019 04:43 AM      Profile for Marcel Birgelen   Email Marcel Birgelen   Send New Private Message       Edit/Delete Post 
The latest iterations of so called deepfakes are becoming scarily realistic pretty quickly and those tools aren't just available to Hollywood anymore.

They don't use classic 3D modeling, which has been used by Hollywood in the past for some of those fake actor inserts, but rather a combination of 3D modeling and "AI" image blending techniques.

Just a few examples from already more than a year ago can be found here.

Yes, we currently can still pretty easily spot MOST of those deepfakes, but the technology is progressing more rapidly at this point than I'd expected.

 |  IP: Logged

Bobby Henderson
"Ask me about Trajan."

Posts: 10973
From: Lawton, OK, USA
Registered: Apr 2001


 - posted 11-07-2019 02:02 PM      Profile for Bobby Henderson   Email Bobby Henderson   Send New Private Message       Edit/Delete Post 
Some deep-fake work is little more than use of applications like Photoshop and After Effects to marry together images and footage that closely match in angles of perspective. The approach is more crude, but the upshot is (if the different pieces of imagery match evenly) it can overcome the problems of The Uncanny Valley all too common in completely artificial 3D models.

Since terms like "nightmare" and "scary" were mentioned in this topic I quickly thought of disturbing 1990's music videos from Aphex Twin: Windowlicker and Come To Daddy. FYI, Aphex Twin is the stage for British electronic musician Richard David James. His face was Photoshopped onto the bodies of other people in his music videos.

BTW, if you watch the Windowlicker video the first 4 minutes of it (before any music begins) follows the attempts of a couple very foul mouthed thug-wannabes at picking up a couple ladies on the street. Go to the 4 minute mark to skip all that cussing.

 |  IP: Logged

Marcel Birgelen
Film God

Posts: 3357
From: Maastricht, Limburg, Netherlands
Registered: Feb 2012


 - posted 11-10-2019 07:40 PM      Profile for Marcel Birgelen   Email Marcel Birgelen   Send New Private Message       Edit/Delete Post 
quote: Bobby Henderson
Some deep-fake work is little more than use of applications like Photoshop and After Effects to marry together images and footage that closely match in angles of perspective.
Even those "household tools" like After Effects become more and more powerful in each iteration.

I can now do stuff in After Effects that would've required Nuke just a few years ago. (Although Nuke's node based approach still scales better, but that's another discussion.)

Adobe already offers character animation software (aptly called Character Animator) that does facial tracking. (Some American talk shows use it regularly.) It currently isn't good enough to produce realistic looking human faces, as it is targeted at animated characters, but give this a few years and you can do pretty evil things with it with just a few mouse clicks.

But REAL deep fakes should rely on AI substitution processes. The thing is, they compensate for perspective automatically and based on so many tracking points, it's virtually impossible for a human to do so.

The stuff isn't perfect and can still be easily detected in most cases. You could easily let a public figure do or say things he/she never did. Dependent on the amount of efforts you put into it, you can make it look so real, sufficient people will accept it as such.

 |  IP: Logged

Bobby Henderson
"Ask me about Trajan."

Posts: 10973
From: Lawton, OK, USA
Registered: Apr 2001


 - posted 11-11-2019 10:44 AM      Profile for Bobby Henderson   Email Bobby Henderson   Send New Private Message       Edit/Delete Post 
AI tech in computer graphics still has a lot of room left for improvement in many areas. Human faces are arguably the most difficult thing to re-create artificially. We subconsciously look for many traits in faces with how they look and how all the parts move. Much of it works on the instinct level. Science only has a basic understanding of concepts like left side bias. Computers can't invent the things even we don't fully understand yet without a whole lot of human-based help.

 |  IP: Logged

Marcel Birgelen
Film God

Posts: 3357
From: Maastricht, Limburg, Netherlands
Registered: Feb 2012


 - posted 11-11-2019 03:02 PM      Profile for Marcel Birgelen   Email Marcel Birgelen   Send New Private Message       Edit/Delete Post 
Like I said, we're not there yet, but stuff is progressing at a pretty impressive pace. I don't really know if it's a good or a bad thing we're not entirely there yet. A lot of cool stuff can be done, if we can create believable, life like, virtual characters out of thin air, but it also opens potential new cans of worms.

quote: Bobby Henderson
We subconsciously look for many traits in faces with how they look and how all the parts move. Much of it works on the instinct level. Science only has a basic understanding of concepts like left side bias. Computers can't invent the things even we don't fully understand yet without a whole lot of human-based help.
Well, this is were stuff like deep learning and neural networks can come in. Imagine putting not just like 100 but like 100,000 tracking points on somebody's face and feeding all this input to a deep-layer, self-training neural network.

We don't necessarily need to understand how facial expressions work in all intricate details and how we, humans, are able to perceive them. But, if we feed a neural network with sufficient examples of it, we can eventually train to mimic it.

That's also what those advanced deep fakes are doing, at a still somewhat crude level. But I'm pretty sure that we'll eventually arrive at a level, where you can feed such a system a whole bunch of footage of somebody you want to mimic and it will be able to recreate an almost lifelike animation of that person.

 |  IP: Logged

Bobby Henderson
"Ask me about Trajan."

Posts: 10973
From: Lawton, OK, USA
Registered: Apr 2001


 - posted 11-11-2019 04:23 PM      Profile for Bobby Henderson   Email Bobby Henderson   Send New Private Message       Edit/Delete Post 
There are still very basic visually creative tasks computers still can't figure how to replicate on their own. In computer graphics work there are dozens of visually effects involving lettering which require human intervention. For instance consider a "prismatic" letter with a V-shaped bevel rising to a center line along its stroke. It's a common effect in hand carved or chiseled signs. The computer can manage a full bevel effect on a letter "I". But it cannot for its life figure out how to resolve the effect at corners or joins in the letter strokes. It always bowling out with wacky, weird curves at those points rather than negotiating the corners and joins correctly. It takes a human mind with capacity of abstract, aesthetic thinking to solve the problem.

The human face has so much more taking place in it than the abstract, geometrical shapes of letters. The best a computer can do is extrapolate known variables, even with fashionable terms like A.I. Computers can't invent detail they do not fundamentally know. Computer software has become pretty good at simulating the behavior of different materials (liquids, fabric, glass, etc) but it's all built up on rules and variables encoded in software by people. The CGI application Houdini is about as good as it gets with materials and atmospheric simulations. But even it still has room to improve on the feats it does well.

 |  IP: Logged

Marcel Birgelen
Film God

Posts: 3357
From: Maastricht, Limburg, Netherlands
Registered: Feb 2012


 - posted 11-11-2019 06:52 PM      Profile for Marcel Birgelen   Email Marcel Birgelen   Send New Private Message       Edit/Delete Post 
You're talking about stuff that's based on fixed algorithms and some of them are pretty impressive. The fluidic effects in Houdini, for example, are awesome, if you know how to tame them, but they're based on simulating the real world though pretty fixed algorithms.

But that's not how neural networks work. Sure, they still use algorithms for their internal functioning, but a neural networks aims at recreating how your brain works, with interconnected neurons.

Covering them here would be way off-topic, but it's an interesting topic that you might want to read up on a little if you're new to the concept. Neural networks themselves aren't something new either, but we now have computers that are sufficiently powerful to run those at usable sizes, whereas this used to be relegated to super-computers.

But essentially, you can train a network to act on certain things. You can train it to recognize patterns, but you can also train it to emit a certain pattern at a certain input.

Some advanced image analyses applications use those. Advanced facial recognition can also be done with them. Another application is the creation of "deep fakes", where the substitution process isn't just a simple 3D-corrected image overlay, but uses autoencoders instead.

 |  IP: Logged



All times are Central (GMT -6:00)  
   Close Topic    Move Topic    Delete Topic    next oldest topic   next newest topic
 - Printer-friendly view of this topic
Hop To:



Powered by Infopop Corporation
UBB.classicTM 6.3.1.2

The Film-Tech Forums are designed for various members related to the cinema industry to express their opinions, viewpoints and testimonials on various products, services and events based upon speculation, personal knowledge and factual information through use, therefore all views represented here allow no liability upon the publishers of this web site and the owners of said views assume no liability for any ill will resulting from these postings. The posts made here are for educational as well as entertainment purposes and as such anyone viewing this portion of the website must accept these views as statements of the author of that opinion and agrees to release the authors from any and all liability.

© 1999-2020 Film-Tech Cinema Systems, LLC. All rights reserved.