Using Deepfake in Trek... Yes or no?

[Pauln6]

I did my own fan edit of Star Trek TMP. Originally I was just going to stitch the scenes from the DE with the SLV but I decided that it would be nice if I gave Grace Lee Whitney a scene in the second half of the movie.

I went a bit nuts in the end and re-added the rough memory wall footage with dialogue. I have no editing or CGI skill and my laptop struggled to vocode Grace's voice so the end result was quite dodgy but there is some potential there for professionals.

I was able to layer Deepfake over a CGI mesh so I can see how next level animation could work but it would be a legal nightmare. Better to just re-cast.

I'm tempted to Deepfake/edit Grace into a couple of Season 2&3 episodes in place of the extra Jeannie Malone but just for fun. I'm not sure her face is the right shape.

 

[jackoverfull]

Interesting. I’d be curious to see the results.

 

[Richard S. Ta]

I'm not into the idea of deep faking a faux TOS season 4 in the slightest. As has been noted, it would hardly be reflective of the prospective Season 4 we could have had. It would be, no pun intended, fakery and nothing more. At worst as well, as in the finale of The Mandolorian, I'd be taken out of the story. When we got that reveal, any surprise or excitementI might have felt was overtaken by a feeling of 'huh, neat special effect, how'd they do that?'.

But then I'm not even a fan of new CGI in old Trek. Basically, my credo is 'leave the old things as they were and make some new things instead'.

 

[jackoverfull]

But then I'm not even a fan of new CGI in old Trek. Basically, my credo is 'leave the old things as they were and make some new things instead'.

especially the way it was done for TOS-R. I understand the limitations but it always irked me.

On the other hand we have TNG-R and that’s how a remaster should be done in my opinion.

 

[Richard S. Ta]

especially the way it was done for TOS-R. I understand the limitations but it always irked me.

On the other hand we have TNG-R and that’s how a remaster should be done in my opinion.

I understand a lot of love and care went into the TOS remasters and updated effects, but... Well, perhaps I'm odd in finding the way special effects were achieved in the 1960s to be intrinsically more interesting than the idea of layering modern effects over a sixties production. Plus, I think the original shots and effects work are actually quite remarkable in and of themselves.

The TNG remaster, I agree. Sympathetically done and not nearly as jarring. I just wish they'd do DS9, VOY and ENT but... well I guess that's never going to happen.

 

[jackoverfull]

Yeah, too expensive for the moment to redo the CGI shots. With the new upscaling techniques based on machine learning they could do a remaster like the recent Babylon V one, though, so it’s no longer totally off the table.

 

[Pauln6]

Interesting. I’d be curious to see the results.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

Did this using a line from World Enough and Time plus vocoded dialogue.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

I have Tweaked this one a bit more since. I just tweaked it so the space suit matched in the later scene.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

This would benefit from a CGI Rand. So hard to manually extract from a dark transporter room and I would rather she was on screen for longer.

Overall it's watchable if shoddy. Sits ok alongside SLV quality.

 

[dupersuper]

I'm not into the idea of deep faking a faux TOS season 4 in the slightest. As has been noted, it would hardly be reflective of the prospective Season 4 we could have had.

Plus we already have TAS, several novels and a comic book mini series literally called Year Four (by DC Fontana herself), followed by Continues, several more novels, and the current Year Five series, so it's not like the time period hasn't been covered...

 

[Richard S. Ta]

Plus we already have TAS, several novels and a comic book mini series literally called Year Four (by DC Fontana herself), followed by Continues, several more novels, and the current Year Five series, so it's not like the time period hasn't been covered...

I love it all. I'm a huge fan of TOS and TAS but I consider them artefacts and a good thing too. I'd go so far as to say I'm repelled by the idea of a deepfaked TOS season 4.

 

[Pauln6]

I'd rather see an animated Phase II series.

 

[publiusr]

I would have no problem with Deep Fake Planet of the Titans with different actors.

I like the idea of period pieces—as was done with “The Star Wars” with McQuarrie-centric looks. Deep Fake opens up story telling, relaxes shooting. If everything is in post...you can take your time.

Movies and ISS missions are sporting events. Carrier deck ops and theatre too. An all storyboard approach can work if supported. I know it can creep people out...but we’ve done this with Zelig and Forest Gump anyway.

Everything now is a series of moments...

 

[jackoverfull]

I like the idea of period pieces—as was done with “The Star Wars” with McQuarrie-centric looks. Deep Fake opens up story telling, relaxes shooting. If everything is in post...you can take your time.

Wouldn't bet on that. Much more likely they'll lower the budget and raise the profit.

 

[Pauln6]

You probably need to look at the Blade Runner sequel or Captain America to get an idea how difficult and expensive such a project might be. You also need look alike actors because Deepfake can't alter the shape of someone's face so straight away you are doubling up your acting costs!

 

[jackoverfull]

It’s expensive NOW. A few years ago it was something that only big budget movies could afford, now we're already to series, soon it will be as inexpensive as regular CGI (and of course what happened when CGI got inexpensive is that’s we got flooded with movies full of badly done effects).

Of course you still need actors (zf or the time being), but you can negotiate much lower contracts with “nobodies” that could get replaced at any time.

 

[Pauln6]

It’s expensive NOW. A few years ago it was something that only big budget movies could afford, now we're already to series, soon it will be as inexpensive as regular CGI (and of course what happened when CGI got inexpensive is that’s we got flooded with movies full of badly done effects).

Of course you still need actors (zf or the time being), but you can negotiate much lower contracts with “nobodies” that could get replaced at any time.

I think the technology exists to do animation with a layer of Deepfake to make the characters look like the actors but they still look artificial when they move and talk.

Although I can run deep learning on my gaming laptop, it is still not smart enough to run analysis on large batches of images, and couldn't vocode a voice from scratch.

If you want to do live action, you have all the costs of a normal show PLUS additional actors PLUS the cost of the deep learning. Even if costs come down it's still more expensive.

A full on animated show could be easier and cheaper.

 

[jackoverfull]

As I just written: RIGHT NOW is still expensive. In a few years it won’t.

 

[Pauln6]

As I just written: RIGHT NOW is still expensive. In a few years it won’t.

No I get that, I just mean even if it gets less expensive, it's always going to be more expensive than just using actors.

I might experiment with the blonde yeomen in Squire of Gothos, Deadly Years, and Wink of an Eye to see how tricky it is to give Grace Lee Whitney appearances in those episodes. I think the shape of the actresses' heads might make a difference to how good it will look.

 

[jackoverfull]

“always” is a very long time. Given the continuously inflating of prices in the entertainment industry, especially when famous people are famous, and the continuous drop in difficulty of this technology I think that the point where negotiating with an estate to use the likenesses of a deceased star and engaging an impersonator will cost much less than employing a star who is alive and that this will happen in a couple of decades or so. After the obvious fad that will follow we’ll see how changed the entertainment industry will be.

Not to mention that a lot of movies will just get rid of most extras as soon as there are predone libraries to animate virtual background actors convincing enough: hiring extras is a cost and often a huge headache bringer.

 

[Pauln6]

“always” is a very long time. Given the continuously inflating of prices in the entertainment industry, especially when famous people are famous, and the continuous drop in difficulty of this technology I think that the point where negotiating with an estate to use the likenesses of a deceased star and engaging an impersonator will cost much less than employing a star who is alive and that this will happen in a couple of decades or so. After the obvious fad that will follow we’ll see how changed the entertainment industry will be.

Not to mention that a lot of movies will just get rid of most extras as soon as there are predone libraries to animate virtual background actors convincing enough: hiring extras is a cost and often a huge headache bringer.

I think Shatner has sold his likeness to the highest bidder :-) He has certainly sold his voice. Majel Barrett also sold her voice for future use, despite which, I don't think she has voiced the computer since her death?

I'm not sure that the families of the stars who have already passed on would be quite so comfortable with the notion though!

 

[jackoverfull]

Correct: majel’s voice have been recorded to be digitally recreated but hasn’t been used again since she died for whatever reason.

Some families might forbid this, Still were already seeing that other estates have no problem at all in selling the deceased’s likeness for digital recreations