Комментарии:
I originally wanted to make this film with images exclusively in total black and white and 4:3 ratio for aesthetic purposes, I originally lost the entire project years ago from my hard drive as I attempted to restore my pc for reasons I can't remember. I guess it was fortunate that it happened as I stumbled onto other AI services that had only just released that were capable of animating these images. I decided to recreate the project in 16:9 ratio and full color to appeal to a broader audience and again, altered the story for thematic purposes for the same reasons. I also wanted to add a touch of my own personal creativity to it, which may or may not have made it worse lol.
This "film" was made with Runwayml Gen 2 and Midjourney 5. Now with the release of Runwayml Gen 3 and Midjourney 6.1 the difference in AI animation is absolutely mind blowing in comparison to what you see in this video. You can even create lip-sync with dialogue which could actually bring true dialogue to a film made with AI. Everything looks SIGNIFICANTLY more realistic and there is absolutely nowhere near as much warping and artifacts as you see in this video, in fact it almost looks life-like in comparison to what was used to create this video. I suggest you look up Runwayml Gen 3 and Midjourney if you are interested in this AI video revolution or if you yourself are interested in creating a film with AI. I don't know that I would completely recreate this video with Gen 3 and Midjourney 6.1 or even make anything else with it as I currently have no creative goals and am busy with other aspects of life. In a year's time there will most definitely be something like Gen 4 and Midjourney 7 or even 8. Imagine 5 years from now. Hollywood who? Lol.
I thought I'd give an update to this as it seems people are very wary of AI filmmaking, especially after viewing this video. I know most of my other content doesn't even relate to this video and is heavily criticized and I don't blame them. The point of me making this update of sorts is to just bring a more positive outlook at AI and it's potential in the future. It's not as bad as you think it is. This post isn't about me, but more so to open your eyes to the potential of this technology.
At the time of making this comment I've accrued over 4.2k views on this video which is honestly more then I could have ever expected, so thank you for giving it a chance and viewing!
It's interesting (maybe fascinating) as far as an AI-generated film. It's hopelessly off as far as actually having anything to do with _Shadow_. Could've just called this an originaly-AI-produced horror movie. As someone else says on down the page, these early AI things should be cherished for themselves; no point in trying to drag original productions into them.
ОтветитьKill it with fire.
ОтветитьInput from wise human is what is lacking. Not good at all.
Ответитьi love the weird mistakes that AI makes LOL this is great
ОтветитьBeards were out of fashion in the 1920s. And it wasn't raining in the original story.
ОтветитьAmazing ! More !
ОтветитьThe AI speech rhythms are tedious and repetitive. Some images are evocative while too many others are over-the-top and take away from the wonderful sense of mystery in the original tale. The choppy editing doesn’t work either in its attempt to include too many over-literal images. An interesting experiment but it can’t beat such evocative film making as people like Val Lewton used to produce. Now he would certainly have directed an interesting version of this macabre story. The lack of “humanity” in AI work is depressing. It’s all visual style with no truly relatable human content.
ОтветитьI'm an avid reader of anything Lovecraftian or Lovecraft-inspired, it's amazing how ±100 years after this work was written it is still holding up as an iconic work of cosmic horror.
Even with the AI artifact issues being what they are, it still kind of works.
As I'm writing this I am only halfway through but still ... aside from the whole being a little static the shifting animations give it that general sense of unease which Lovecraft tried to capture in his works, mirroring the fragile psyche of the narrator as he is recalling the events from memory.
In this regard even the occasional "jumpscares" where the AI is trying to add a visual to certain descriptions (i.e. the museum's tiara scene) seem particulary apt given the narrators sanity.
This Animation Needs Fixing.
ОтветитьWhare's The Movement.
ОтветитьA very good adaptation of Lovecraft's story, and the only quibble I have is that the AI has not mastered the subtleties of language and sounds odd in some instances where it becomes obvious that it is merely repeating the text without the nuances and natural pauses that the narrative requires.
ОтветитьShow us dont tell us lol has a lot of potential
ОтветитьUgh, 5 minutes in and I have to say, 'AI' has a way to go before I will be willing to sit through any more than 5 minutes worth.
The singing was painful, initially seemed an ok attempt at generating the sound of a motley crew of salty sailors singing shanties...but after a bit it became a very repetitive noise, with no human variation or imperfections.
The waves skipping all over the place in parts were off-putting, and the scene in the bar/mess area was a touch cringy.
Kind of a pity, I was looking forward to watching some HPL.
Ultimately, it's not a bad effort, but it has a long way to go on a number of levels.
If I had to characterize thoughts on it in one word, that would be 'plastic'.
Keep at it though, 'AI' has to go through this process to reach its best potential. ;)
Hated it.
ОтветитьLike the Newbury-Innsmouth-Arkham bus, this art tool cannot yet be trusted. I was easily disturbed by fluid, morphing patches of color meant to form the planes of 3D objects in motion. Mouths changed shape, teeth unstably followed, profiles of faces disjointed as they turned from the 3/4 view. Also, the computer hasn't been taught yet how to make leg motions appear to have weight. Is the young bearded man on the boat the same person as the young bearded man in the village? Are either of them supposed to be the same as the young clean-shaven narrator? Zadok Allen's face kept changing, too, in length of beard, in the sucken-quality below between cheek and chin.
ОтветитьAhahaha godamn the end took me off guard
Ответитьthe ever-shifting figures and landscapes have an eerie, dreamlike quality that will no doubt disappear as AI gets better at this kind of thing. But I cherish these early attempts for precisely those qualities. Sometimes, perfect is just not what is needed. (see Midjourney v.3 vs v.6) I'm only a few minutes in but I find this captivating.
ОтветитьThis is Awesome youre so creative...Narration .is captivating.
ОтветитьFantastic job!
ОтветитьGood experimental work. Check my Black Mirror Episode of 9 minutes here in the channel. I'd like to have feedback.
ОтветитьOnly 6 minutes in. If this was any other genre, how choppy and quirky it is wouldn’t work. But this kinda thing is perfect for Lovecraft.
ОтветитьSo trippy, feels like I'm on shroom.
ОтветитьIt's like a severe case of artifacts, almost difficult to watch, even though it DOES showcase how the AI is thinking of several things at once...a pity.
ОтветитьThis is fucking ass
ОтветитьVery, very nice 👍
ОтветитьSame
Ответитьhere so early i cant even turn the quality up
Ответить