|
Post by awpeacock on Jan 3, 2022 14:04:47 GMT
I've just got through my Season 17 boxset and I think there was a mixed trick with Shada that could have proved helpful with filling in the missing gaps with the missing episodes. Confession here - this was actually the first time I've seen it completed with the animation (I did have the original VHS with Tom narrating the gaps) and I did actually find it a little jarring switching between the animation and the live action.
The main thought I had watching it though was that, in 2021/22 with the technology we have available to us now, was this not the prime opportunity to have a go at DeepFake to fill in the gaps? We have the audio and, thanks to the scenes that had been filmed, we have reference points for the look and feel of everybody so, with the relatively small amount of footage that needed to be filmed, why not use this as a proof of concept to test the viability and cost of the newest technology. That viral Tom Cruise video (just done by a hobbyist am I right?) plus the Luke Skywalker scene in the Mandalorian shows what could be done, and I'm sure I'm not the only person out there who would rather have missing Who reproduced this way than animation? It would be the perfect opportunity for the Beeb to see what demand there is for it, and the reaction, to see whether to move forward with it for the rest of the missing episodes? It can't be THAT much dearer than the animation?
Or am I living in cloud cuckoo land?
(Actually, knowing the BBC, they'll probably re-release Shada later this year with DeepFake to cash in again, get everyone re-buying the story by itself and then the boxset again!)
|
|
|
Post by Richard Bignell on Jan 3, 2022 18:15:15 GMT
Deepfake isn't anywhere near viable or remotely affordable for something like this.
|
|
|
Post by John Wall on Jan 3, 2022 21:34:17 GMT
The thing about technology is that it continually improves - and reduces in price.
Perhaps in ten, maybe five, years……
|
|
|
Post by jamesvincent on Jan 4, 2022 1:54:04 GMT
Deepfake isn't anywhere near viable or remotely affordable for something like this.
Why? You've got teenagers on youtube doing amazing stuff with it in their bedrooms!
|
|
|
Post by John Wall on Jan 4, 2022 10:23:38 GMT
Deepfake isn't anywhere near viable or remotely affordable for something like this. Why? You've got teenagers on youtube doing amazing stuff with it in their bedrooms!
It’s not a question of practicability but affordability. A DW episode probably needs c. 23 minutes - that’s a lot of animation, or whatever. The obvious DW candidate is Crusades, which would complete Season Two, and has been often discussed here. There are only two episodes to do but there are a lot of characters and costumes - and we’re probably talking about 45-46 minutes total. Then, do you do it as part of the BluRay range or look for a standalone release first to get a double dip? We know that the powers-that-be have been, and are, looking at these things but at the end of the day the income has to exceed the expenditure.
|
|
|
Post by RhysH on Jan 4, 2022 10:53:38 GMT
Personally I would be fine with them releasing as a standalone to help fund the continued animations and then adding to the box set for a complete release. My ideal would be an animation of The Space Pirates next (after TAS) so that we can get Season 6 released as we have seen a lot of Season 5 recently. Same applies to Seasons 1 and 2's remaining missing story in each but appreciate those appear to be harder jobs for the teams to do. So a full Season 6 set seems possible!
|
|
|
Post by John Wall on Jan 4, 2022 11:02:05 GMT
Personally I would be fine with them releasing as a standalone to help fund the continued animations and then adding to the box set for a complete release. My ideal would be an animation of The Space Pirates next (after TAS) so that we can get Season 6 released as we have seen a lot of Season 5 recently. Same applies to Seasons 1 and 2's remaining missing story in each but appreciate those appear to be harder jobs for the teams to do. So a full Season 6 set seems possible! I think Season Five will be completed as it’s all monster stories which should sell. Then it comes out as a BluRay set in a few years time. Space Pirates, I agree, is an obvious candidate although I don’t believe there are Telesnaps. I wonder if it’s been priced? Marco Polo and Crusades are definitely more difficult but it’s certainly possible to contemplate four complete B&W seasons 👍
|
|
|
Post by Richard Bignell on Jan 4, 2022 12:38:00 GMT
Deepfake isn't anywhere near viable or remotely affordable for something like this. Why? You've got teenagers on youtube doing amazing stuff with it in their bedrooms!
The whole basis of deepfake is that you have suitable footage to apply the deepfake technology to in the first place. With Shada, you don't. So, you'd have to get a studio, build sets, create costumes, employ actors and shoot all the material that was never recorded in the first place - all of which has to be paid for. And that's even before you begin to start work on the computer side.
|
|
|
Post by awpeacock on Jan 4, 2022 23:02:58 GMT
Why? You've got teenagers on youtube doing amazing stuff with it in their bedrooms!
The whole basis of deepfake is that you have suitable footage to apply the deepfake technology to in the first place. With Shada, you don't. So, you'd have to get a studio, build sets, create costumes, employ actors and shoot all the material that was never recorded in the first place - all of which has to be paid for. And that's even before you begin to start work on the computer side. OK, so I could be reaching here somewhat but that's why I went for Shada as the proof of concept: There's effectively, what, 2 (or 3) sets that didn't get filmed - Skagra's ship, Shada itself, and there's probably already something put together in BBC world that could be Chris's kitchen. And, heck, with video editing software available these days, you probably don't even need to build those - you could put together a set in Maya and green screen it (or just stick a model in if you want). And the costumes can't be too hard to come by? Tom's/Skagra will probably have been kept somewhere for starters. And just to keep my overly optimistic (naive?) view going ... as all we're looking for is someone to just walk around to be superimposed over, you'd find fans willing to do that for nothing. That Uni recreation of Mission to the Unknown shows that. Yes, I'm probably severely detached from reality here but that's just a characteristic of a sci-fi fan, no?
|
|
|
Post by John Wall on Jan 4, 2022 23:46:31 GMT
I’m not sure how long ago ST TNG did “Relics” and put Scotty on the bridge of the original Enterprise - that was possible as they were able to get a still of the set from TOS. Then DS9 did all sorts of wonderful things with Sisko, etc in “Trials and Tribbleations”.
I’m unsure about Shada but in terms of b&w MEs if there are sufficient telesnaps/production/publicity piccies it may be possible to recreate the sets. What would be really great was if designers had taken a series of piccies of their work without people in the way.
It all comes down to money and technology.
|
|
|
Post by Richard Bignell on Jan 5, 2022 12:59:05 GMT
There's effectively, what, 2 (or 3) sets that didn't get filmed - Skagra's ship, Shada itself, and there's probably already something put together in BBC world that could be Chris's kitchen. Not quite. There's about ten major sets that would need to be rebuilt for the missing material and "Chris's kitchen" is actually an equipped laboratory. If the original recording had gone ahead in 1979, as the studio floor plans show, these sets were substantial enough to require them to be taken down overnight and new ones replaced for the following days. Yes, Tom's costume from this season largely exists in private collections, but nothing else from Shada does (other than the bottom half of a Krarg which appears in the new footage used in the production), so all of that would have to be made from scratch. Hmm. I really don't think that, on a BBC budget, producing a CG set that you're then going to green screen your actors into before replacing all their faces using deepfake is going to look in any way convincing. And wasn't that what a lot of people were criticising about the animation used for The Web of Fear #3? That it was a non-actor trying to perform the roles for motion capture, but they didn't look right?
|
|
|
Post by andrew shutt on Jan 5, 2022 14:05:21 GMT
Marco Polo and Crusades are definitely more difficult but it’s certainly possible to contemplate four complete B&W seasons 👍 Not sure about Marco Polo ,but i can see the Crusade's 2 missing episodes just having Reconstructions ,due not only to the vast anount of costumes and sets needed to be animated but also the poorish sound quality of the surviving audio for the two missing episodes
|
|
|
Post by John Wall on Jan 5, 2022 14:46:05 GMT
Marco Polo and Crusades are definitely more difficult but it’s certainly possible to contemplate four complete B&W seasons 👍 Not sure about Marco Polo ,but i can see the Crusade's 2 missing episodes just having Reconstructions ,due not only to the vast anount of costumes and sets needed to be animated but also the poorish sound quality of the surviving audio for the two missing episodes Mark Ayres has done wonders elsewhere. These factors must be known to those commissioning animations, etc.
|
|
|
Post by jcoleman on Jan 5, 2022 15:46:32 GMT
As Richard has said, Deepfake is miles away from being a useful tool in recreating entire missing episodes technically, practically and financially.
If you were to experiment with it though then the University of Central Lancashire's recreation of Mission to the Unknown would provide the necessary starting point. It has the required raw footage with sets and actors. Someone has already synched their footage to the original audio, which works remarkably well and further shows what a great job the students did.
A key technical issue though would be how similar the student actors are facially to the original actors. Deepfake really needs that similarity to work well and be convincing. There are examples on YouTube where excerpts of the fan-made, not-for-profit Star Trek Continues series have had deepfake applied and it works incredibly well for Captain Kirk because Vic Mignogna bears a resemblance to William Shatner but rather less well for other characters such as Spock. There is also an issue with how much footage of the original actors is available and its quality. There are nearly 80 episodes worth of restored footage of the original Star Trek actors to draw from and they're in character. Suitable samples of Edward de Souza and co are likely to be much harder to come by.
There are also several other obvious non-technical issues.
Firstly, why bother? Mission to the Unknown doesn't feature any of the series regulars so does it really matter that the actors are not the originals? (As an aside, I personally found some of the Dalek props used far more distracting than knowing the actors weren't the originals. It's a shame they couldn't source a couple more 1960s-style ones given how much effort went into faithfully recreating everything else. If you're using deepfake to try to recreate the original episode as accurately as conceivably possible then would you also want to attempt to replace the anachronistic Dalek props in some way?)
Secondly, is it ethical to take the student actors' performances and replace their likenesses with those of other others?
Thirdly, and this is a wider consideration, how well can stand-in actors capture the style of the originals? You may be able to make someone appear to look like William Hartnell or Patrick Troughton but how convincingly could they capture their gestures, movements and mannerisms?
For the foreseeable future at least, deepfake and other similar techniques remain in the purview of fans to experiment with. However, a negative consequence of official missing episode animations has been a drop off in the interest fans used to have for reconstructions and the innovative ways of enhancing these that were beginning to emerge. That this has occurred just at a time when new technology and techniques are becoming available is doubly disappointing for those of us who enjoyed this method of consuming missing Doctor Who. There are some interesting efforts out there but they are few and far between and those producing them rarely continue to do so for long.
|
|
|
Post by awpeacock on Jan 6, 2022 20:43:08 GMT
As Richard has said, Deepfake is miles away from being a useful tool in recreating entire missing episodes technically, practically and financially. If you were to experiment with it though then the University of Central Lancashire's recreation of Mission to the Unknown would provide the necessary starting point. It has the required raw footage with sets and actors. Someone has already synched their footage to the original audio, which works remarkably well and further shows what a great job the students did. A key technical issue though would be how similar the student actors are facially to the original actors. Deepfake really needs that similarity to work well and be convincing. There are examples on YouTube where excerpts of the fan-made, not-for-profit Star Trek Continues series have had deepfake applied and it works incredibly well for Captain Kirk because Vic Mignogna bears a resemblance to William Shatner but rather less well for other characters such as Spock. There is also an issue with how much footage of the original actors is available and its quality. There are nearly 80 episodes worth of restored footage of the original Star Trek actors to draw from and they're in character. Suitable samples of Edward de Souza and co are likely to be much harder to come by. There are also several other obvious non-technical issues. Firstly, why bother? Mission to the Unknown doesn't feature any of the series regulars so does it really matter that the actors are not the originals? (As an aside, I personally found some of the Dalek props used far more distracting than knowing the actors weren't the originals. It's a shame they couldn't source a couple more 1960s-style ones given how much effort went into faithfully recreating everything else. If you're using deepfake to try to recreate the original episode as accurately as conceivably possible then would you also want to attempt to replace the anachronistic Dalek props in some way?) Secondly, is it ethical to take the student actors' performances and replace their likenesses with those of other others? Thirdly, and this is a wider consideration, how well can stand-in actors capture the style of the originals? You may be able to make someone appear to look like William Hartnell or Patrick Troughton but how convincingly could they capture their gestures, movements and mannerisms? For the foreseeable future at least, deepfake and other similar techniques remain in the purview of fans to experiment with. However, a negative consequence of official missing episode animations has been a drop off in the interest fans used to have for reconstructions and the innovative ways of enhancing these that were beginning to emerge. That this has occurred just at a time when new technology and techniques are becoming available is doubly disappointing for those of us who enjoyed this method of consuming missing Doctor Who. There are some interesting efforts out there but they are few and far between and those producing them rarely continue to do so for long. Thanks - this is exactly the sort of reasoned, knowledgeable response I was looking for - if not quite the positive response I was looking for! From yours and Richard's replies, I've clearly been in cloud cuckoo land. Particularly disappointing to hear these sorts of things are being neglected due to the animations which, as I've said, I'm not a big fan of.
|
|