thibaulttricard.bsky.social
Associate Professor in Computer Graphics in the Maverick team at LJK/Inria Grenoble.
39 posts
208 followers
453 following
Regular Contributor
Active Commenter
comment in response to
post
No, fill out the linked Google form: forms.gle/3JdX3SLF46KG...
comment in response to
post
I see…
Maybe if you combine K complementary blue noise sequences and use the density field as a threshold.
It would limit the number of density level to K, moreover finding the complementary blue noise is not an easy task
comment in response to
post
I’m not sure there is a paper on that, but if you use a white noise distribution that follow the density field and apply a loïd relaxation on the laguere diagram created by the distribution and the density field you should converge towards something close to what you want
comment in response to
post
Supplemental video available here: www.youtube.com/watch?v=6lKA...
comment in response to
post
I am concerned academia is still a "publish or perish" model. How will it impact our ability to get a paper accepted?
Imho, a solution would be to design a stable benching setup and change publication guidelines so that only the results obtained on that setup would be accepted
Easier said than done
comment in response to
post
Special thanks to @iquilezles.bsky.social for allowing us to use his Arlo SDF/shader: www.shadertoy.com/view/4dtGWM
comment in response to
post
Or density fields :
comment in response to
post
Big voxels :
comment in response to
post
You can use it to rig complex animated noise :
comment in response to
post
Our method is rasterization-based and can be integrated with a typical scene graph.
Thus, it's compatible with classical rendering algorithms such as shadow maps :
comment in response to
post
Our method supports self-intersection:
comment in response to
post
In my team only a few machine are high end, but most of our paper are benched on them.
imho two reasons to do that :
- impressive number helps convince the reviewers
- we want to show what the perf of our method on an average pc in 5-6 years ( average AAA game dev time )
comment in response to
post
The video file: inria.hal.science/hal-05046040...
comment in response to
post
on Shadertoy: www.shadertoy.com/view/MXyBDW
comment in response to
post
The video on Youtube: www.youtube.com/watch?v=hcVu...
comment in response to
post
I would say the symbolic violence created by a company that see their consumers as cow to be milked can justify some yelling.
Personally I think that not allowing people to have emotion will lead us to a worse path…
comment in response to
post
Maybe it's not a monopoly (I'm not an economist), but it is a nasty trap for sure
comment in response to
post
imho 2 things:
1) Debuting artists have no real alternative to the Adobe bundle (bundle is cheap)
2) If you're fed up or limited by Adobe, you have to either change all your software at the same time, or keep the bundle and change them one by one until all are replaced by more expensive ones
comment in response to
post
I didn't say that Adobe is preventing interop in the competition.
Two option
1) you have one software that does a lot of stuff
2) you have multiple software that do complementary stuff.
Choose #2 and you're gonna spend a lot of time importing and exporting to switch software (if it's possible).
comment in response to
post
That's the issue, when you pay for Photoshop, you have a toolbox for many practices, replacing Photoshop requires multiple software that might not be interoperable
I use GIMP, it does a good job for small task but it's not user friendly (you can clearly see that it was designed by programmers)
comment in response to
post
I'm not an artist, so I might not be up to date, but last time I checked, there was no equivalent to Photoshop.
And yes, sometimes there are more expensive alternatives, but that's the problem: artists are stuck with Adobe until they can afford something else -> monopoly on mid-range market
comment in response to
post
A company has a quasy monopoly, has a crappy price policy, and has stopped listening to its users.
The users roast the company publicly, what’s wrong here ?
What else can they do ?
The idea that everyone should be nice and polite all the time only profit powerful entities
comment in response to
post
Nerf was slow as hell until they started using gaussian splat, 3D texture, or voronoi cell as rendering primitives.
Turns out having efficient rendering primitive is important 🤯. Who knew ?
Wait until they realise primitives need to support animation …
comment in response to
post
I wouldn’t be so sure about that, ML based rendering have made a lot of progress by ignoring important properties and leaving them to future works, and when they try to solve go back to graphic method
comment in response to
post
This year HPG is in Europe. Just saying.
highperformancegraphics.org/2025/program...
comment in response to
post
I tried asking Dall-E to improve it and it just added some reflections here and there but the lighting model still seems off.
comment in response to
post
Aside from the lack of diversity in the image and the weird anatomy of the two men, Was Dall-e always this bad at lighting, or is this something new?