Adam Pickard

@styletransfer

Art Director trying to make sense of this fast moving technology.
Followers
1,290
Following
463
Account Insight
Score
46.52%
Index
Health Rate
%
Users Ratio
3:1
Weeks posts
I tried to explain the thinking and workflow steps of one of my #comfyui experiments. I hope it makes sense. haha
26 0
1 year ago
I remade the Eameses' "Powers of Ten" using collaged-together search result images as frames. Adam Pickard (@styletransfer ) remade the film with DALL•E. So *of course*, we had to make a two-sided flipbook to show off our two remakes. We'll post more about our respective processes next week but for now, I'm really enjoying the fancy box and zine we created for it. If you want one, I just put them up in my shop! (the link is on my profile page- just click on my name and then there's a link there.)
5,730 76
3 years ago
I’ve been using Apple’s SHARP model to generate 3D Gaussian Splats from videos in ComfyUI. I built some nodes that let you pick camera angles and make new videos from new perspectives. Far from perfect but promising so far.​​​​​​​​​​​​​​​​
5 0
4 months ago
The start and end frames were made using GPT 4o, and the video was generated using the Wan 2.1 Start + End Frame workflow in ComfyUI.
41 7
1 year ago
Wan 2.1 I2V Start + End Frame w/ some ghibli anime style. I made the animation a little longer by using 3 frames, first 1-2, next 2-3 and that’s why you can see the colour change in the middle.
23 1
1 year ago
I saw everyone was using Nerdy Rodent’s video2video WAN ComfyUI workflow to add babies to movie scenes, so I did this. Why? Not sure; I’m sorry. haha
0 2
1 year ago
Generated using Kling elements feature, which uses the four images below to guide the animation.
0 1
1 year ago
Found some slides from Art School and animated them in Kling to make a loop.
0 3
1 year ago
The video above is generated using Kling 1.6 elements feature, which uses the four images below to inspire the animation. Some parts of the video look very realistic and other parts look less so but I really like how the sound creates a bit of confusion for the viewer.
18 2
1 year ago
This started with an AI image I generated over a year ago. I wanted to do something with it.  This track came to mind; I liked its “realness” compared to electronic tracks and the challenge of lipsyncing its fast lyrics. I wanted something that felt as real as I could make it with current technology but with synthetic hints, like the random optical illusion in the first clip. Then, over time, it transitions to something obviously more synthetic. #ai #aivideo
0 4
1 year ago
Quick #sora tests, using image to video in storyboard mode and adding some quick audio. #sorashowcase
0 1
1 year ago
KlingAI released a #ComfyUI node, so I added it to the end of my Flex redux workflow. I originally tried it with the magazine node component, but it quickly broke down and looked strange in the results. So I just went a straight video. You obviously get way more control with models like this using their own portals, but some interesting results. I sped it up and added some sound in After Effects.
40 8
1 year ago