Hi
Barley Studios is an Ai film studio from director Neill Blomkamp. He says he now wants to make a full feature in this format.
https://youtu.be/8Wbtt2JxP7g?si=3bd8ynHtugL1xUT4 I even went to his Twitter account to check whether it was legitimate. https://x.com/NeillBlomkamp
Here is a review by Matt Workman - https://x.com/mattworkman/status/2079353928135712845
There are definitely some impressive shots here and there, the opening scenes of London are stunning, and some of the helicopter shots are genuinely impressive. But something about the editing feels off. It's hard to put my finger on it, but my brain almost feels exhausted trying to work out why the human characters seem to lack energy, vigour, or vitality, do you know what I mean? There is also the lack of lens effects, depth of field, everything is too sharp. I wish he was making District 10, I loved that film.
Anyway, I thought some of you might find it interesting.
Replies
The more you look, the more you notice it getting things wrong.
I'm surprised that someone of his calibre didn't spot the morphing bike that goes under the bus right in the middle of the screen during the opening shot! Thats a simple fix, come on! Also, we drive on the correct side of the road here in the UK - the left. 🙂
Putting important things like the price and consequences of this "pipeline" aside, what exactly is so bad about it?
If we take away tech issues like hallucinations and broken sharpness filter, because those can be fixed eventually, would it make any difference?
If you pretend for a second that it was not made with AI, would it improve the perceived quality of the work?
Is it all look so weird because of technology or director's vision?
Those of his movies I watched looked good, though, must admit, simply not to my taste. Still, they had certain "charm" and it seemed to be centered on the characters/actors who carry the work with emotional performance. This doesn't have any anchor, any pull. Just people talking and stuff happening.
This gives me the same impression as any other AI videos so far: like the prompter was so impressed with sheer tech capabilities, they forgot to put effort into the movie as a whole. And if they really made an effort and think this was good enough, then I have to question their taste as experienced professional.
If this was supposed to be proof-of-concept, I'm not sure what it proves. The tech is impressive, but everything else is not. Is it the fault of the "tool" or the artist? Would it be better if someone else made it?
Naturally it's just an opinion of a mere consumer: my dislike may be simply a matter of taste in movies and heavy bias against questionable tool.
Still, if this is the future, than I'd rather stick with those who are going to be "left behind".
And on first AI mistake you associate the value of it at zero. Since all they did was "prompting" till it clicks. Which it was not this at this level of coherency scene to scene. t2i can't make that coherent image just by prompting alone. In this case most probably complex setup in comfy and not at all text2Image. Practical does look worse but is associated with "high efford" thus "have value" for consumer. When viewer looks at practical FX and sees the usual quirks and thinks of it. Aha! they did a lot!. After many years i learned that there is no universal "good" for visuals. Certain of my clients wanted something that would not be accepted here as good at all
I totally take your point that practical is often better, but the petrol bomb that was supposed to be an atomic blast in Oppenheimer immediately ruined the film for me.