Content Creator | Specializing in VFX, and AI-Driven Video Solutions. Business Contact: [email protected]

Joined July 2010
Another example of using masking to add elements into a scene. For this one, I used ComfyUI with Wan 2.2 Animate and masking to fill in the area where Wolverine’s mask would be. For me, using AI in subtle, surgical ways is the key. It helps keep things more convincing instead of feeling overdone. I know people will nitpick it, and that’s fine. I’m just excited to explore what feels like the future of VFX.
82
326
30
3,600
418,064
A chance to win 10k or an RTX 5090👀
This quoted post is unavailable.
1
1
14
1,801
I tested Astra for the first time to edit a video using footage I shot at IBC Amsterdam. Is it actually any good at editing? I put my experience and thoughts below 👇
2
1
16
1,779
I imported everything, picked a song, and gave it a rough idea of what I wanted at the beginning, middle, and end. What worked really well: It automatically labeled my clips based on what was in them, which made a huge folder of footage much easier to search. It also built out the timeline nicely and picked some shots where I immediately thought, “Yeah, that’s a great choice.” Where it struggled: There were also clips where I had no idea why it chose them. I still had to remove shots and find others I knew I wanted in the edit. The labeling helped a lot here because finding replacements was much faster. The surprising part: I asked it to create some masking transitions, the kind you see a lot in Reels, and it actually did a pretty good job. Once I gave it more specific directions, it handled those well too. I still went in myself to add some extra polish like pulses, lens flares, and other transition details. My takeaway: Astra can definitely speed up editing, especially organizing footage, building a first cut, and even handling some effects. But I wouldn’t hand it footage and expect a finished edit. It doesn’t really know your taste yet. You can direct it, but sometimes it’s still faster to get your hands in there and make the creative decisions yourself. For me, it feels more like an editing assistant than an editor.
1
3
741
Whenever I go to events and meet people using AI for production work, I notice that a lot of them aren’t very active on social media. They don’t share much of their work or even their tests online. Being on here can make it feel like we’re seeing everything that’s happening with AI, but there’s so much interesting work being done in production that simply goes unnoticed because it’s never shared online.
16
3
1
103
4,572
The idea started while I was in Berlin last week. I shot the main video in my hotel room on my phone, and the rest came together using a mix of different tools. I’ll list them all below.
15
29
7
484
34,985
Tools: SAM 3D Body Blender Astra Seedance 2.5 iPhone
1
30
1,766
Now in different locations
4
11
185
9,712
Been experimenting with different use cases for SAM 3D Body. I think this one is a good one 🥳
SAM 3D Body is open-source, so it can be run locally or on Comfy Cloud. Extract motion from a recording → bring it into Blender → control the camera → use the result with Seedance 2.5 More control over body motion and camera movement. To try this workflows, links below 👇
7
118
13,595
I feel like there are some really fun potential use cases with SAM3D. Still playing around with ideas, but here I used LTX to create a clean plate and a SAM3D workflow to place a 3D model into the shot, similar to how Autodesk Flow Studio handles character replacement, except this was done entirely locally in ComfyUI. The next step would be replacing that 3D model with whatever model I want, which would probably require jumping into Blender and continuing the process there. There are also some interesting possibilities with the motion capture data and being able to control the camera inside Blender while keeping the original movement. That could potentially be really useful for things like MiniMax V2V workflows. Still very early in my testing, but I’m excited to keep experimenting and dive deeper into what’s possible with this.
8
22
1
242
12,928
Another test with Blender Camera movements and Minimax. Results not so great cause of the Minimax jank but still interesting.
1
1
10
927
A side by side version of the H3 inpaint video I showed before.
4
16
145
12,866
Made with MiniMax H3 inpainting workflow in ComfyUI, running locally on my PC.
13
22
5
312
73,374
Inpainting using MiniMax H3 😎 github.com/drozbay/MaskVidEx…
13
104
6
1,045
97,949
And now it’s local 🎉🎉
Comfy MCP is now local and open-source! The #1 ask after Comfy Cloud MCP shipped in June. Your agent sees your actual install. So every node, custom nodes included, and every model on disk. It fetches files, starts your instance, gets a workflow to the point where it runs. It is now the easiest way to help with your local Minimax H3 workflows! Cloud MCP still does everything it did. Tell your agent where a job goes, or let it decide.
2
2
25
2,648
Say the line! 😑
19
7
2
98
16,977
Can you figure out which one is the original and which one was done with MiniMax H3 using masked inpainting? Testing it on the infamous Justice League scene where Henry Cavill’s mustache was digitally removed. 👀
13
7
2
111
13,607