AI generated Super Panavision 70 trailers

[renaissance_man]

I recently found an interesting YouTube channel that has reimaginings of TV shows and films in a very specific style.

I know AI art and moving images are a bit of a contentious subject given their potential to replace VFX artists but I thought the way these have been produced are interesting and at times very freaky.
If you look closely in a lot of these clips you can see the very distinct flaws of AI in generating moving images, hands, mouths and often legs moving in a backwards motion whilst the character is moving forwards are all there.

Just a small sample of the videos on the channel.
I have no affiliation with them, I just found them recently and thought it would be interesting to share.

 

[renaissance_man]

 

[Treadwell]

 

[renaissance_man]

 

[division 6]

There are several channels like that.

 

[AnubisGuard]

YouTube recommends this kind of stuff to me all the time and I never watch it. Too unappealing. There's also something that feels bait-and-switch about it, like it looks high-effort at first, but then you realize it's about the laziest kind of "content creation" you can have.

There are several channels like that.

Also this. The channels seem to proliferate.

 

[CB2001]

I like this one:

 

[division 6]

Abandoned Films
Patrick Gibney
Schrodinger's film club
Warden Cinematic's

​

 

[robstyle]

It's been a rapid jump in what someone on a computer can do alone and quickly. We're at the cusp of reedits and reimagining of existing media on a professional level. I'm not talking this form of "AI" generated content but full films and TV shows. That may be 20 years on but the ground work is being plowed and paved over faster than we can comprehend. Imagine being able to remake Dial of Destiny with a Short Round and a better ending...

 

[Captain Dunsel]

... Imagine being able to remake Dial of Destiny with a Short Round and a better ending...

Even better…use the awesome, colossal powers of our AI overlords to master space and time and travel back to the past to ensure that The Dial Of Destiny never happens at all.

 

[Astyanax]

Even better…use the awesome, colossal powers of our AI overlords to master space and time and travel back to the past to ensure that The Dial Of Destiny never happens at all.

Some of us were spared by you brave ones who went first, and we skipped it entirely.

 

[Captain Dunsel]

Some of us were spared by you brave ones went first, and we skipped it entirely.

And they say that there is no Greater Generation than those who endured World War II…

Ladies and gentleman, I present to you the survivors of the first wave of audiences to have endured The Dial Of Destiny…

 

[yaris]

All of it always looks so goddamn creepy....

 

[Tan Djarka]

Interesting. I think this has potential, both good and bad.

I have an old clip of Doug Henning performing on the Merv Griffin Show. I don't like all the shots the director chose, but it was shot live, so that's all there is. I'd like to see AI generate a theoretical b-roll so you could create a "super edit".

 

[Tan Djarka]

Most of the characters in those videos look just a bit odd, but somehow, very familiar. Then it occurred to me. They look like characters straight of a Norman Rockwell painting!

 

[Captain Dunsel]

All of it always looks so goddamn creepy....

Yes—AI produces images of people, in these trailers, that appear very dreamlike / nightmarish as they move in unnatural ways.

Very creepy, indeed.

 

[publiusr]

It's been a rapid jump in what someone on a computer can do alone and quickly.

Here are some resources:

Modular software for scientific image reconstruction

Scientists use an array of imaging instruments to look inside living organisms, sometimes as they move, and to observe inert objects without altering their state. Such instruments include telescopes, microscopes, CT scanners and more. But these instruments, even when working at maximum capacity...

techxplore.com

Data-driven model generates natural human motions for virtual avatars

Humans can innately perform a wide range of movements, as this allows them to best tackle various tasks in their day-to-day life. Automatically reproducing these motions in virtual avatars and 3D animated human-like characters could be highly advantageous for many applications, ranging from...

techxplore.com

A new framework to generate human motions from language prompts

Machine learning-based models that can autonomously generate various types of content have become increasingly advanced over the past few years. These frameworks have opened new possibilities for filmmaking and for compiling datasets to train robotics algorithms.

techxplore.com

A new model to produce more natural synthesized speech

Recent technological advances are enabling the development of computational tools that could significantly improve the quality of life of individuals with disabilities or sensory impairments. These include so-called electromyography-to-speech (ETS) conversion models, designed to convert...

techxplore.com

Controlled diffusion model can change material properties of objects in images

Researchers from the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) and Google Research may have just performed digital sorcery—in the form of a diffusion model that can change the material properties of objects in images.

techxplore.com

AI trained to draw inspiration from images, not copy them

Powerful new artificial intelligence models sometimes, quite famously, get things wrong—whether hallucinating false information or memorizing others' work and offering it up as their own. To address the latter, researchers led by a team at The University of Texas at Austin have developed a...

techxplore.com

Looking for a specific action in a video? This AI-based method can find it for you

The internet is awash in instructional videos that can teach curious viewers everything from cooking the perfect pancake to performing a life-saving Heimlich maneuver.

techxplore.com

Researchers analyze the characteristics of AI-generated deepfakes

Most of the deepfakes (videos with fake hyper-realistic recreations) generated by artificial intelligence (AI) that spread through social media feature political representatives and artists and are often linked to current news cycles.

techxplore.com

Adobe's VideoGigaGAN uses AI to make blurry videos sharp and clear

A team of video and AI engineers at Adobe Research has developed an AI application called VideoGigaGAN, that can accept a blurry video and enhance it to make it a much shaper product. The team describes their work and results in an article posted to the arXiv preprint server. They have also...

techxplore.com

Research brings together humans, robots and generative AI to create art

Researchers at Carnegie Mellon University's Robotics Institute (RI) have developed a robotic system that interactively co-paints with people. Collaborative FRIDA (CoFRIDA) can work with users of any artistic ability, inviting collaboration to create art in the real world.

techxplore.com

More:

Artificial Intelligence - General News

https://www.msn.com/en-au/money/news/xbox-ai-chatbot-to-takeover-customer-support-and-game-refunds-claims-report/ar-BB1kXqow The ghost of Mr Clippit is apparently being revived for Microsoft’s new AI plans, which are also at the heart of its next gen console. Rather than destroying all...

www.secretprojects.co.uk

CADmium Moves CAD To The Browser

For plenty of computer users, the operating system of choice is largely a middleman on the way to the browser, which hosts the tools that are most important. There are even entire operating systems…

hackaday.com

 

[Too Much Garlic]

I have seen some of these. They are clearly made by people who have never watched movies of the past.

 

[Solo4114]

Alright, here's the deal.

1. Most people don't actually understand how AI works. They can't tell you the difference between machine learning, neural networks, or deep learning. To them, it's all just fancy buzzwords, and they see ChatGPT as a super-powered search engine (it's not -- please don't use it like that).

2. At the core of it, AI is a tool that recognizes and reproduces patterns. That's it. That's all it does. It's a very fancy, very powerful way of gathering information, sorting for patterns, and reproducing what it has determined is a pattern. AI is not creative. AI is derivative, and -- what's worse for artistic purposes -- AI has a tendency to genericize everything.

3. This is the reason why AI art looks both weirdly unreal, and very, very samey. That's because AI is, fundamentally, ill-suited to the actual creation of artistic works. Art requires creativity, and all AI does is spit back common denominators at you. The most creative that an AI gets is when it actually hallucinates and produces something that looks like it crawled out of an H.P. Lovecraft story.

4. The true problem with AI and the arts, and the real reason why AI is so controversial, is not simply that the stuff it produces looks like generic uncanny valley trash. It's that it's all based on copyright infringement. The way AI's "learn" (and I want to say that I am loath to use "alive" terminology to describe what AI actually does, because AI IS NOT ALIVE OR SENTIENT, but we're stuck with this terminology until people are more careful about it) is by inputting a bunch of data into the AI, from which it begins to build ability to recognize patterns. This is often referred to as "training data," which I think conveniently masks what's actually happening. The AI is described as "training" on the data, which makes it sound like the AI is sitting there, reading books, saying "Hmm. Yes, yes. Very interesting. I must remember that part. What an excellent example of metaphor..." But it's not. It's plugging this data into a ****ing computer program so that the program can build derivative works from it.

Anyone who works in a creative field, who makes their living off of their own creativity, should be absolutely infuriated at the notion that an AI might've been trained on the sweat of their brow, and they get nothing for it.

And in the end, it still looks like ass.

 

[renaissance_man]