Welles and New Digital Technologies - Digital Recovery of Missing Scenes

[tonyw]

??? This is a posting designed for Roger Ryan and the more technologically accomplished members of this Group. Last Sunday, I read an article in THE NEW YORK TIMES concerning digital reproduction of the performances of deceased actors or present ones at their former younger selves becoming a possibility due to the development of new digital technology.

Does anyone consider the likelihood of digitally reproducing those missing scenes in Welles's films such as THE MAGNIFICENT AMBERSONS such as the final boarding house meeting between Eugene and Fanny? This would also involve state of the art audio-technology to reproduce the voices of those now deceased actors in the scene.

[Kevin Loy]

Well, Tonyw, I can't speak from the visual angle, but I can assure you that it would largely be a thankless task from the audio angle.

Theoretically, if you knew the dialogue, then you could assemble it. It could easily regress into assembling the dialogue by each individual syllable, but then again, I seem to remember reading somewhere that another director did this for a piece of Welles dialogue after he refused to dub it (Compulsion, maybe?) 50 years ago, so if that story is true, then this is nothing new. The difference is that digital audio editing makes the whole process considerably easier.

But once you read that, you should understand the incredible limitations of such an approach. There are matters such as phrasing, vocal inflection, pacing, etc. that would basically be left up to the sound engineer. This would be coupled with the fact that the only existing boarding house footage (unless the rough cut is floating around somewhere out there, and unless I am mistaken) appears for a few fleeting moments in the Ambersons trailer, and digitally re-creating the scene would be based out of a few seconds of silent footage that cannot possibly be completely representative of the entire scene.

Don't get me wrong, it might be *technically* fascinating to see the scene revived in such a manner, but it certainly wouldn't be Welles' scene. And you wouldn't really be watching Joseph Cotten and Agnes Moorehead on the screen. And I doubt that you could replace any of them in this fashion, let alone all three of them.

[Roger Ryan]

tonyw - I was thinking the exact same thing when I read that New York Times article! The intriguing part is the promise that this kind of CGI visual reconstruction will soon be able to be accomplished quickly and relatively inexpensively.

Kevin - There are actually five or six frame enlargements available from the boarding house scene in "Ambersons" (they can be found in the Bogdanovich and Carringer books) that show specific camera placement for key moments in the scene. In addition we have the few seconds found in trailer plus the description of the shots found in the cutting continuity. I would say a pretty good approximation of the scene as Welles intended could be recreated. Keep in mind that many of the cuts in "Ambersons" came from existing scenes where we already have all (or most) of the visual information needed. Trying to recreate the actors' actual voices is another matter entirely as you pointed out and unnecessary in my opinion. Just bring in some real actors who are good matches to the Mercury performers and have them mimic that style (in the same way Anthony Hopkins dubbed in Laurence Oliver's voice for an entire scene in the restored version of "Spartacus"). What would be important is that the lost sequences could be worked into the film as seamlessly as possible in order to give a better context for the Welles' footage that remains in the released film (which was the purpose behind my reconstruction as well). It's true that the recreated scenes wouldn't actually be Welles' work, but then his name is on the released film and an awful lot of what's there is not his work either!

[Glenn Anders]

At least twenty-five years ago, I brought an Esquire article to the attention of my colleague, "BAMBO-BAMBO" Christianson, a half-brother of Todd Baesen. The article described what then seemed a bizarre process of taking all the available sound and visual recordings of performers who had extensive careers, and giving them new lives, through the use of video tape and computers. As I remember it, the article's hook was the recent death of Elvis Presley.

Since then, we have had the developments which Roger mentions. I would add the quite magical work done by Woody Allen in ZELIG, and a number of TV commercials, several years ago, using the images of actors like Cary Grant.

I went back, through google, to see if I could find the article in my questioning memory -- no luck.

But I did find one, "CASTING CALL AT FOREST LAWN: THE DIGITAL RESURRECTION OF DECEASED ENTERTAINERS - A 21ST CENTURY CHALLENGE FOR INTELLECTUAL PROPERTY LAW" by JOSEPH J. BEARD:

http://www.law.berkeley.edu/journal....xt.html

More recent I assume, it lays out in the form of a legal brief, some of the vast implications involved in using the images of dead performers for other purposes: Copywright, Trademarks, Libel, Slander, Racism . . . Incest, etc. It turns out to be a profoundly metaphysical question.

[Maybe, it's what Beatrice Welles' logic, extended far enough, is all about.]

For instance, as Beard cites, if a film were made using digitalized, computer animated images of Clark Gable, but the actor is identified in the credits as "Bob Spencer," has Gable been reincarnated as someone new? Beard uses the term, "reanimated."

[What if someone, I might add, puts a re-animated Gable/Spencer in a porno movie? Has Clark Gable been libeled. Again, by extension, these are questions which go to the heart of the question of Identity -- a favorite Wellsian theme, I might point out.]

Beard concludes:

"In centuries past, authors and artists enjoyed a kind of immortality denied to performers. So long as their prose, poetry, music, paintings and sculptures endured, the memory of long dead authors and artists persisted in the physical embodiment of their creative genius; works created centuries ago are enjoyed today by the latest inhabitants of this planet. Performers were not so fortunate; performance was gossamer, vanishing upon completion. All that remained were, at best, fragments - a sketch of the performer, a description of the performance, a playbill, nothing more. The twentieth century changed all that; the motion picture and the sound recording lend permanence to the talent of the performer. The voices of Caruso, Madame Schumann-Heinke and McCormack have not been stilled by death. The performances of the Barrymores, Olivier, Tracy, Garbo and Swanson, captured on film, remind succeeding generations of their skill as actors. The twenty-first century may bestow yet an additional measure of immortality to performers via reanimation.

"The prospects of reanimation are exciting and boundless. Acting dynasties may find the living performing with their ancestors; Drew Barrymore may find herself teamed with her grandfather John and her great-aunt and uncle, Ethel and Lionel. Dream casts assembled across generations would enrich the silver screen. Society might be able to rectify, in some small way, social ills of the past. Actors from minority groups who had been given only stereotypical roles during their lifetime or who were cast in films with limited, ethnic distribution could be cast in serious roles with general appeal. Actors denied the opportunity to explore subjects taboo during their lifetimes would have the opportunity through reanimation to address the once-forbidden topics. But, unless the legal issues implicated by reanimation are addressed in a thoughtful, comprehensive and cohesive way, reanimation may find itself strangled at birth.

"Reanimation technology is still in the embryonic stage of development. In fact, it is likely that reanimation will not reach the silver screen until early in the twenty-first century. There is time to explore the legal issues raised by this fascinating innovation in entertainment. This article has discussed a number of issues raised by reanimation; others may occur to the reader. There is time to debate the best ways of responding to the legal issues raised. Too often in the past, the legal profession has lagged behind the scientific community. Playing legal catch-up might impede rather than foster innovation. I do not suggest that laws be drafted before we fully understand the ramifications of the technology. But it is not premature for the legal profession to join the scientific community in discussing the entwined legal and technical issues of reanimation. The readiness is all."

----------------

It would seem that "the readiness" is upon us, here in our 21st Century Wellesnet discussion.

Glenn

[Randy Cook]

I assume the article (or ill-disguised press release) in the TIMES was the one titled "Cyberface: New Technology That Captures the Soul". While the technology is amazing, you aren't going to magically re-create Marilyn or the Duke or Aggie or Jo. That's simply press-agentry.

You'll be able to do an imitation of any dead actor you'd like, of course, but it'll be an imitation. True, imitation worked in SPARTACUS with Hopkins revoicing Olivier's Crassus, but that was pretty unusual situation. Not only is Hopkins a great actor and formidable mimic, but he'd understudied Olivier in the theatre and was in a unique position to intelligently second-guess Olivier's take on the scene. Since the original photography of the scene was used, the only "recreation" of Olivier's performance was vocal, and limited to the ONE ACTOR.

I've read Welles pre-recorded the dialogue for some scenes in AMBERSONS but don't know if he did so for the scenes in question or, if so, whether the audio still exists. So, to recreate Eugene and Fanny one would need a couple gifted voice actors who were good mimics and who could second-guess the performances which Cotten and Morehead might've given (if Fanny's performance in the final scene was as eccentric and chilling as the "it's COLD" moment, I'd wish them luck, and pray for a miracle).

Then you'd need actors to be motion captured giving the performance...maybe the same two as the voice actors, maybe not. Then you would need motion edit "technicians" and animators (and, when we did GOLLUM, we never had any motion capture data which was not edited...often extensively) to further guide and refine the performances.

In other words, any re-created performance in AMBERSONS would of necessity be a collaboration among a number of actors (vocal and visual, with editors and animators being responsible for significant aesthetic decisions further guiding the performance as well). You'd ALSO need someone to sit in Welles' chair and make directorial decisions reflective of his (and I refer to acting only...there'd also be the job of recreating his camera moves, lens choices, and lighting in addition to the creation/rendering of the faces, hands, clothing and hair).

So, a big job, involving many people of highly unusual qualifications. Not impossible, but not the comparatively inexpensive plug-and-play that the technology salesmen would like you to believe.

I haven't the time to do a budget breakdown, but I guarantee that counterfeiting the missing Welles footage from AMBERSONS so that it'd drop seamlessly into the film proper would cost more than, say, the money needed to complete THE OTHER SIDE OF THE WIND.

I'd love to see it tried, but I'd hate to watch an hour of AMBERSONS and suddenly be presented with a scene from POLAR EXPRESS.

[Randy Cook]

Of course, I meant Agnes MOOREHEAD. I was either typing too fast or I've been watching too many James Bond pictures.

[Glenn Anders]

If you are the Cook who designed "Q the Flying Serpent," you can take on Agnes Moorehead, and be in our kitchen any time!

Glenn

[Randy Cook]

Glenn

I'm too busy apologizing for mistakes in my post to apologize for mistakes in my career!

I DID forget to mention, though, the hundreds of facial expression sculptures which'd have to be made for each actor to resurrect same. Even more extensive work to be undertaken by even more accomplished artists.

Randy

[ToddBaesen]

Randy:

Thanks for providing your expertise on the subject of digital doubles for AMBERSONS. I'm sure you have much more idea of what can be done accurately in this area than what was reported in the N. Y. Times. Besides your superb effects work on LORD OF THE RINGS did you also do any work on the restored spider sequence which Peter Jackson & Weta added to the '33 KING KONG?

I thought that was quite an astonishing piece of work, especially in seamlessly matching the style of what Willis O'Brien did originally, at least technically, but there again you run into questions of placing yourself in the shoes of Merian C. Cooper and Willis O'Brien. I personally felt, that despite the great pains that were taken in recreating and matching the style of KONG with the new footage, that they went a bit overboard conceptually, by adding in too many creatures. I doubt that even if O'bie had wanted to include all of those puppets while he was animating on the film back in 1933 that there would have been the time or the money to do that many! So, to me, in it's sensibility, the scene became more like a scene out of Peter Jackson's new version of KING KONG.

So, as you point out, those kind of problems would be multiplied many times if someone were to try and recreate the actors from AMBERSONS digitally. It would be really stepping into another spider-pit, because if anyone wanted to take the time and money to do it, what would be the point, unless it were to match Welles original intentions as closely as possible, and trying to get a line on what Welles thoughts were to the material in 1942 would be part of the problem. I think it would be very difficult for anybody working in movies today to do it, but like the spider-pit in KING KONG, it would still be fascinating to see the results...

[Roger Ryan]

Yes, thank you Randy for your insight on this matter. The New York Times article did seem a little overly optimistic (especially since the photo they used of the "digital actor" looked like it was straight out of an average video game). Even given the article's "press agentry" spin, I still assumed the technology would be decades away from pulling off these facsimiles in a realistic manner. Ultimately, I'm thinking it might be a mistake to attempt to recreate any lost film scene too seamlessly, or to try to fool an unsuspecting viewer. Better I suppose to stick with stills and storyboards and let the viewer's imagination play a role. What makes the recreated "spider pit" scene from "King Kong" such a treat is the setup explaining how the recreation was done which makes the whole thing a superlative DVD extra. Had the recreated footage been incorporated into the actual film, the credibility level would have dropped considerably.

[Randy Cook]

Todd & Roger:

I advised on the Spider Pit recreation, at an early phase (I'm sitting next to Pete on the dvd and shooting my mouth off, but as a stop-motion veteran I defend myself by saying I was caught up in the moment). I did feel that the puppets were off-model (cartoonist's term, I think, it means a given drawing of a character looks nothing like the model sheet used to define the character's appearance), and much bigger than indicated by the drawings produced under O'Brien's supervision.

I agree it was better to leave KONG "as-was", rather than presuming to drop that arachnid forgery into the movie.

But the Spider Pit (another Holy Grail from the RKO trim bin, of course: maybe it's mouldering in the same vault as the AMBERSONS deletions...nawww, just kidding) was excised by the film's creative supervisor as being undesirably extraneous. Footage from AMBERSONS was excised at the behest of the studio's ADMINISTRATIVE supervisors, and is this a different kettle of fish.

If these scenes could be intellegently re-created, they'd NEED to be inserted into the film itself, for only by viewing them in context can the effect of what Welles was going for (no matter how theoretical) be achieved. In my opinion.

God knows the Bremner recording of Herrmann's full score hints at a film with a much different feel. Roger, I've not seen your reconstruction, but I applaud any attempt to throw light on what Welles was up to. I think the ARKADIN reconstruction is a very good job and makes the film (one of my favorites, despite that makeup!) more watchable (to me) than the previously extant versions.

The computer has put sophisticated editing of old films into the hands of many, of course, so we'll see more travesties and more revelations as time plods along. And someday the ability to create photorealistic humans will be in the hands of many, as well. I just don't think that day is here, unless there's unlimited time and money.

Were one to tackle it now, one would do well to build a database of scenes of Cotten, Holt, Moorehead and the rest, culled from other films. Using these, the digital doubles could be roto'd for appropriate key actions, which would at least give one a starting point for the production of forged scenes. Such a database would also be an imperative tool for studying the process of said actors. Intelligent extrapolations would have to be made, naturally, for those actors never played those scenes, or those characters, in other films.

But it'd be a start.

How I wish these amazing tools had been available 35 years ago, when Moorehead & Cotten & Welles were still alive! Makes me want to kick the wastebasket when i think what Welles could have done, rummaging around in the toolbox we lesser talents exploit so successfully today.

Anyway, thanks for including me in the group. I have been reading your posts for a couple of years and feel as though I know you all. It's wonderfull that the amazing work Welles left us is still being discussed, and appreciated.

[Glenn Anders]

Randy: The process of storing the culls is what was envisioned in that Esquire article that I read many years ago. The author (not the lawyer, whose piece I cited) had observed that, after 70 years (at the time), it was theoretically possible to gather enough images and sound bites to create new works, using the "talents" of artists from the past. Some of the personalities could be tracked, in another sense, from childhood to senility. That ability was unique in history.

No doubt the technical problems were, and are, more momentous than the author could imagine. I remember, he said that several companies, which he named, were exploring the field and laying down legal markers.

In that regard, I'm surprised that no one has commented on the other article, the one I did find, in which Lawyer Beard explores the legal implications of "reanimating" players and sequences of the past.

Good to see you here, Randy, and to know that you have cast a bemused eye upon our speculations.

Glenn

[Randy Cook]

Glenn

Thanks for being the voice of reason, here.

I have real qualms about the prospect of creating digital celebrity zombies, as it runs the risk of sullying the actors' legacies.

Personally I do not mind seeing an historical character, or a celebrity, depicted in a biographical picture (and maybe I should) when portrayed by an actor in makeup. The more he or she looks and sounds like the character, the more enjoyment I derive from the accomplishment (or stunt, if you prefer...I think it can be both).

One assumes these pictures are fiction (I also assume written biography is usually fiction as well, but that's a sidetrack which runs for MILES).

But there's something these CGI exhumations which bears more examination from the moral standpoint, as well as from the legal.

In my enthusiasm to theoretically see Ambersons restored in some form approximating what was lost, I neglected to remember that it could turn into a treasure hunt in a graveyard.

It will be most intriguing to see what unfolds, or unravels, as the technology advances... from both the legal and philosophical perspectives.

[Roger Ryan]

Footage from AMBERSONS was excised at the behest of the studio's ADMINISTRATIVE supervisors...

If these scenes could be intellegently re-created, they'd NEED to be inserted into the film itself, for only by viewing them in context can the effect of what Welles was going for (no matter how theoretical) be achieved. In my opinion.

Randy - This is my feeling exactly. In a sense, what should be remarkable about an "Ambersons" reconstruction is not the recreated footage itself, but how the existing Welles-directed footage becomes more meaningful when placed in its proper context. Simply re-editing the scenes found in the released version back to their originally intended sequence (plus dropping in the excised Herrmann music cues) improves the emotional tone of the film as well as its continuity, but one needs to go further to better understand where Welles was coming from. What is desirable is a method which will give the viewer the necessary information to provide a proper context as well as be dramatically effective without the technology becoming a distraction. Could "digital doubles" ever accomplish this? I don't know.

Glenn - For me, there is a difference between recreating a "character" portrayed by an actor and recreating the actor as a personality. Apart from attempting to restore films like "Ambersons" or "Greed", I can only imagine the technology to "resurrect" now-deceased actors being used as a gimmick (Astaire dancing with the vacuum cleaner). Restoration or exploitation; aren't those the only two choices in this matter? If the estates clear the rights for the images of old actors to be used in TV ads, then you would hope they would do so for a restoration attempt. I would think that recreating a "personality" to sell a product would be more questionable legally than recreating a "character" from a lost film segment.

EDIT - I posted the above before reading Randy's last post, but it seems like we're on the same page here.

[Glenn Anders]

It would seem we all agree.

Roger: For your restoration, how extensively do you think, have you drawn on the "complete" Herrmann score we have in the Bremner rendition that Randy mentions?

Glenn