We have already checked all the videos and submitted the 15 selected ones to the University of Nis! We have also finished the initial projection test on the 3D-print model of the Nis Fortress.
Due to administrative delays related to the venue permit, partly caused by the recent renovation of the nearby bridge, we are awaiting the exact date of the projection event.
We were originally planning to make the montage video once we had the actual footage from the event. Given the current situation, we will iterate on the timeline and announce the winners first! Thanks for flagging out. We will do better in communication.
Is there a surface inside that gate that can receive a projection, or is it open air/too dark for the projector? Also, I assume we can use a negative matte in post if we need to keep stuff within bounds? Cheers!
The input reference can be used as an input in the ComfyUI node workflow to help the AI understand the contours of the object. The overlay must remain stable so the image does not move during the projection.
I just wanted to clarify one thing: will participants receive a mapping template or a reference file with the exact shapes/measures of the fortress walls?
Or should we simply create a standard 1920×1080 video and your team will handle the warping/mapping on your side?
So, when it says loop-able does it mean an unseamless loop where the video loops a few times, a reverse loop or a gif, in which we'll lose resolution? I have multiple graphics in each video so a seamless oop will probably be impossible.
Encountered "Alert 413 - Request Entity Too Large. File upload failed." with the above 3D model file (165MB) in ComfyUI Desktop. Solved it by decimating the mesh down to 6.1MB, and it loaded successfully for now. Hopefully the ComfyUI Desktop could support to upload large 3D model this time for the challenge.
When I project this template reference, it aligns perfectly with the physical object. That’s why it’s important that the overlay video remains exactly the same during projection, using the same reference resolution (1920×1080), so everything stays perfectly aligned when projected back onto the real object. This is true projection mapping, not some flat screen movie theater projection.
Update on the progress:
Thanks everyone for submitting your great work!
We have already checked all the videos and submitted the 15 selected ones to the University of Nis! We have also finished the initial projection test on the 3D-print model of the Nis Fortress.
Due to administrative delays related to the venue permit, partly caused by the recent renovation of the nearby bridge, we are awaiting the exact date of the projection event.
We were originally planning to make the montage video once we had the actual footage from the event. Given the current situation, we will iterate on the timeline and announce the winners first! Thanks for flagging out. We will do better in communication.
The test of the projection went well: https://www.youtube.com/watch?v=6hGkGQUiZl0
Hi Jo, thanks for the update!😊 Could you let us know when you expect to publicly announce the 15 selected projects? Thanks!
Is there a surface inside that gate that can receive a projection, or is it open air/too dark for the projector? Also, I assume we can use a negative matte in post if we need to keep stuff within bounds? Cheers!
Yes, doors will be closed so you can map that part too :)
Thanks! Just to clear one other thing up, we need to also create an (inverted?) depth map from the template to process our videos with?
You can use template .png which have alpha in it. Video have to be with black background.
The input reference can be used as an input in the ComfyUI node workflow to help the AI understand the contours of the object. The overlay must remain stable so the image does not move during the projection.
Thanks for the explanation!
Will you be taking care of the masking on your projection mapping software Nikola? Is Black Background sufficient?
Yes black black background is sufficient. Background behind fortress have to be black, you can use template .png with alpha in it.
Yes, you will recive mapping template .png for accurately following the contours for projection.
where will I receive the .png file ?
Here you go, until admins add it somewhere visible https://drive.google.com/file/d/1PU2RAxO5Weu4i3JG8mnBdFp4fdxbRPTS/view?usp=sharing
Thank you very much for the opportunity !
This looks like an amazing project!
I just wanted to clarify one thing: will participants receive a mapping template or a reference file with the exact shapes/measures of the fortress walls?
Or should we simply create a standard 1920×1080 video and your team will handle the warping/mapping on your side?
Did you receive the file I sent?thx
So, when it says loop-able does it mean an unseamless loop where the video loops a few times, a reverse loop or a gif, in which we'll lose resolution? I have multiple graphics in each video so a seamless oop will probably be impossible.
Encountered "Alert 413 - Request Entity Too Large. File upload failed." with the above 3D model file (165MB) in ComfyUI Desktop. Solved it by decimating the mesh down to 6.1MB, and it loaded successfully for now. Hopefully the ComfyUI Desktop could support to upload large 3D model this time for the challenge.
Sorry about that, this model was optimized for 3D printing.
When I project this template reference, it aligns perfectly with the physical object. That’s why it’s important that the overlay video remains exactly the same during projection, using the same reference resolution (1920×1080), so everything stays perfectly aligned when projected back onto the real object. This is true projection mapping, not some flat screen movie theater projection.