[ad_1]
Then DALL-E gets 2, and things are dialed up to 11.
Text-to-image integration AI applications Like DALL-E 2 Allow anyone to create anything they want with their imagination by simply typing in a few command words known as “queries”. You can imagine a dramatic love story between John Oliver and a cabbage, or turn David Bowie’s lyrics into a real work of art. Worthy of Ziggy Stardust’s own album. But, in the hands of someone as creative as Trillo, these AI tools are the equivalent of using a zoetrope to industrial lighting and following your every need using a magic command line.
“It’s an incredibly exciting and exciting time for creators,” Trillo told me via email. “In one sense, AI democratizes visualization, allowing more verbal people to express themselves visually. It gives people who are already visual a way to improve their work and go down paths they’ve never explored.
“[The experiment] It worked much better than I expected,” says Trillo. “My next impulse was to do a clothes swap, and that’s it I have worked in the previous projects, But I wanted to. Experiment with other ideas Before using with a full body person. That’s where the power of the tool became clear to him. Generating with AI is so fast that “we can get a lot of ideas now,” he says.
Overwhelming, limitless creative power
To guide the AI to create the outfits seen in the video, a number of text prompts ranging from “purple iridescent mylar oversized t-shirt” to “lavender purple jumpsuit” to “lavender purple retro futuristic fashion jumpsuit with mock turtleneck, compact feather shoulder pads, avant garde fashion, Japanese Minimalism from 2040, Barbarella.
“[It] Describing its power as “limitless and overwhelming,” Trillo opened the door to some pretty wild designs that I couldn’t have come up with on my own.
https://www.instagram.com/p/ChInrv1DfwW/
Another DALL-E 2 stop-motion experiment by Trillo, which used text-to-voice AI synthesis to narrate the voice of Sir David Attenborough.
The difference between traditional methods and AI is amazing. According to Trillo, you need to design and build 100 costumes while controlling the camera and then have the model change every frame. Another way is to design and build clothes in 3D, create fabrics and textures, then light them on the video and do the necessary composite work. But DALL-E 2 can do this via text prompt. Not only can it generate an object like a garment, but it can also recreate a photograph and instantly combine an object into that photograph. This is a unique feature for DALL-E 2, which other AI compositing programs do not: “DALL-E analyzes the beauty, lighting, perspective, everything of the original image and seamlessly blends new elements into the original image. . Adding objects to a scene or filling in blanks is surprisingly good, this process ‘Coloring,’” says Trillo.
Making an AI fashion show
Making the perfect video is a straightforward and even tedious process. First you have to take a base video which is a toy shot of the owner Shyama. By garden. Then you need to extract the video frames and feed them into the DALL-E one by one, introducing text to the query for each one. “It’s basically AI-generated stop motion,” he said.
Some fits that didn’t make the cut pic.twitter.com/7pbIRvVZo2
— Paul Trillo (@paultrillo) August 23, 2022
“Once I explored a few different directions to go, I developed a unique style for the wardrobe,” Trillo says, considering the balance of visual design diversity and visual consistency. He ended up using 115 garments, with countless others left on the cutting room floor. Those had to line up in a way that they flowed together organically, but it was also unexpected. Finally, he used another AI program called RunwayML to transfer the image sequence into the source video. And then, to complete the sequence, he created the floating objects that you can see using DALL-E, again using stop motion and placing them in different layers for added depth.
I’m not afraid of AI.
One of DALL-E’s dangers, Trillo believes, is that it sucks. His “imagination” is so amazing that you can easily lose hours and hours in the browsing process. But you can curb that compulsion, he believes, and that’s why it’s a powerful weapon in the creative arsenal.
Trillo knows that DALL-E will have a negative impact on the creative industry, but, “Tt will not take any work from visual effects artists.” If anything, it “creates efficiency to do what you’re already doing. It opens the door to completely new techniques and allows low-budget projects to have photo-realistic VFX.
This makes sense. I can see how creative people are actually safe and encouraged by these new tools rather than being threatened by them. I can imagine people who are good at using After Effects or Photoshop, but limited in their creativity, losing their jobs in the same way that so many other jobs have been lost to technology that allows other people to do amazing things.
Trillo makes another good point: “If everyone could make a show, the show would be boring.”
[ad_2]
Source link
