[HN Gopher] Bolt3D: Generating 3D Scenes in Seconds
___________________________________________________________________
Bolt3D: Generating 3D Scenes in Seconds
Author : jasondavies
Score : 270 points
Date : 2025-03-19 22:30 UTC (1 days ago)
(HTM) web link (szymanowiczs.github.io)
(TXT) w3m dump (szymanowiczs.github.io)
| slowtrek wrote:
| Is anything like this available locally yet?
| emmelaich wrote:
| Here's the repo: https://github.com/szymanowiczs/splatter-image
|
| Apparently you can clone and run the demo locally. But wasn't
| clear at a glance how much is local and what hardware required.
| ashikns wrote:
| Isn't it generating in the browser using webgpu?
| gessha wrote:
| I assume that's for interactive viewing only, not for
| generation.
|
| > Our method takes 6.25 seconds to reconstruct one scene on
| a single H100 NVIDIA GPU or 15 seconds on an A100.
| dvrp wrote:
| I mean, it's the same author but seems like co-authors are
| different.
|
| How do you know it's the actual implementation?
| echelon wrote:
| Your link above (Splatter Image) is not the same code / paper
| / research as Bolt3D.
|
| This is a previous paper/work by the lead author a year
| before they interned at Google Research and produced Bolt3D.
|
| Bolt3D appears to be his intern research project done in
| conjunction with a bunch of other Google and DeepMind
| researchers.
|
| I don't suspect there will ever be publicly available code
| for this.
| kombine wrote:
| This is previous work. Bolt3D uses the same principle, of
| predicting a per-pixel Gaussian splatting representation but
| it also trains a diffusion model, which is only feasible if
| you have substantial compute available.
|
| Given that it's a work done at Google I will not expect them
| to release source code. But it will be reproduced by someone
| else soon enough.
| emmelaich wrote:
| True, I should've said _related_ work.
| diggan wrote:
| Show. Us. The. Wireframes!
|
| Every single time a new "Generate 3D" thing appears, they never
| show the wireframes of the objects/scenes up front, always you
| need to download and inspect things yourself. How is this not
| standard practice already?
|
| Not displaying the wireframes at all, or even offer sample files
| so we could at least see it ourselves, just makes it look like
| you already know that the generated results are unusable...
| ajwin wrote:
| My understanding is that it is not mesh, it's Gaussian
| Splatting. There are tools to convert Splats into mesh though.
| diggan wrote:
| Yeah, but isn't still the expected outcome to end up with
| actual 3D objects, not point clouds? Or did people start
| integrating point clouds into their 3D workflows already?
| Besides for stuff like volumes and alike, I think most of us
| are still stuck working with polygons for 3D.
| text0404 wrote:
| splats augment 3D scenes, they don't replace them. i've
| seen them used for AR/VR, photogrammetry, and high-
| performance 3D. going from splats to a 3D model would be a
| downgrade in terms of performance.
| nine_k wrote:
| Meshes are editable. Are Gaussian splats?
| text0404 wrote:
| kinda but not really in a meaningful way, at least not
| yet. there's some plugins for popular 3D software but
| it's still early days.
| therealpygon wrote:
| Yea, someone can say, "Look, we have just created the
| first color computer and it displays images. Look at this
| first ever real life photo on this digital screen!" There
| will always be the people who ask, "Yeah, but does it run
| Photoshop?"
| dmarcos wrote:
| What kind of edits you mean? You can crop / combine
| splats easily in your browser with supersplat (not
| affiliated)
|
| https://superspl.at/editor
| dmarcos wrote:
| What's the best use of splats that you've seen so far
| that I can try? AR/VR or regular 3D
| tracerbulletx wrote:
| No geometry in the conventional sense. I did a demo of
| rendering a Gaussian splat in React Three Fiber here, you
| can open the linked splat file (its hosted on hugging face)
| if you want to see the data format.
| https://codesandbox.io/p/sandbox/3d-gaussian-splat-in-
| react-... I also have this youtube video about creating
| that demo https://www.youtube.com/watch?v=6tVcCTazmzo
| Legend2440 wrote:
| You can convert splats into meshes using a simple marching
| cubes algorithm.
|
| But the meshes produced are not easy to edit.
| dheera wrote:
| Generating good meshes sounds like a problem for a
| completely different machine learning algorithm to me.
| MITSardine wrote:
| Meshing has been around long before machine learning came
| to prominence, there's plenty of methods to improve
| surface meshes already.
| lmpdev wrote:
| I use pointclouds all the time in Rhino/Lastools/Meshlab
|
| I much prefer pointclouds and nurbs over meshes
|
| Not everything is gamedev
| diggan wrote:
| > Not everything is gamedev
|
| Agree, I'm not sure why you'd think that's the only use
| case for 3D, unless I misunderstand your argument here.
|
| How would you handle visual effects with point-clouds for
| example? There are so many use cases for proper 3D, and
| all I can think of as use-case for point clouds are
| environments with static lightning, which seems like a
| really small part of what people generally consider "3D
| scenes".
| lmpdev wrote:
| > Visual effects
|
| Maybe I missed the mark on "gamedev", but 3D is larger
| than just "aesthetically pleasing 3D VFX" for its own
| sake
|
| Often I'm trying to use something as a reference for a
| design where a 3D model isn't the actual end goal, or I'm
| performing analytics on a 3D object (say in my case for a
| lot of GIS and simulation work)
|
| The whole "mesh is the be all and end all of 3D
| modelling" irks me as while yes it's a really important
| way of _representing_ an object (especially with real
| time constraints), it doesn't do justice to the full
| landscape of techniques and uses for 3D
|
| It would be like 2D sprite artists from the gamedev world
| saying "what's the point of all this vector art you
| illustrators are doing" or "what's the point of all these
| wireframe designs you graphic designers are doing" -
| "these aren't raster images!"
|
| I suppose my snipe was trying to communicate the idea
| that 3D is larger than just a vehicle for entertainment
| production. It intersects many industries that may eschew
| polygons because real time rendering is irrelevant
|
| 3D tooling has uses beyond producing 3D scenes, just as
| Photoshop is used for more than touching up photographs
|
| Edit: for anyone stuck in a rut with meshes come join the
| dark side with nurbs - it makes you think about modelling
| in a radically different way (unfortunate side effect is
| it makes working with meshes feel so so "dirty")
| CyberDildonics wrote:
| _The whole "mesh is the be all and end all of 3D
| modelling"_
|
| No one said this, it seems like you are making up fake
| questions and not dealing with the actual questions that
| the person you replied to asked.
|
| You can view point clouds and you can warp them around,
| but working with them and tracing rays becomes a
| different story.
|
| Once you need something as a jumping off point to start
| working with, point clouds are not going to work out
| anymore. People use polygons for a reason. They have
| flexible UVs, they can be traced easily, they can be
| worked with easily, their data is direct, standard and
| minimal.
| MITSardine wrote:
| Games are the least of it, the vast majority of
| scientific applications to do with physics use meshes
| rather than point clouds.
|
| This is because a point cloud does not represent a
| surface or a volume until the points are connected to
| form, well, a surface or a volume.
|
| And physical problems are most often defined over
| surfaces or volumes. For instance, waves don't propagate
| over sparse sets of points, but within continuous
| domains.
|
| However, for applications where geometric accuracy is
| needed, I think you wouldn't want to use a method based
| on a minimal number of photographs anyways. For instance,
| the Lascaux cavern was mapped in 3D a decade ago based on
| "good old" algorithms (not machine learning) and
| instruments (more sophisticated than a phone camera). So
| these critiques are missing the point, in my opinion.
| These Gaussian Splatting methods are very impressive for
| the constraints they operate under!
| dheera wrote:
| Not necessarily.
|
| If you're using it to render video you don't need to go
| into the mesh world.
| lallysingh wrote:
| Expected by whom? Other researchers in this space? That's
| the audience for this work.
| eMPee584 wrote:
| Isn't https://svraster.github.io/ just superceding gaussians?
| Voxels are also not meshes, but might they not prove even
| more useful for coming rendering engines..?
| text0404 wrote:
| splats don't have wireframes, and they have an embedded webgpu
| viewer in the linked page.
| jsheard wrote:
| They usually don't show the material channels either, which I
| assume is because there aren't any, and instead the lighting is
| statically baked into the asset. That works for a demo where
| you just wiggle the camera in a circle, but it'll immediately
| fall apart if the lighting environment changes or anything in
| the scene moves.
| Legend2440 wrote:
| Think of it more like a 3D picture than an animation model.
|
| There are no materials channels or wireframe. It's a
| volumetric 3D representation, like a picture made up of color
| blobs.
| kookamamie wrote:
| And thus unusable for most things 3D models and scenes
| would be used for today.
| ForTheKidz wrote:
| I see your point, but also consider that interactivity
| comes to mind in part because 3d models are so expensive
| to describe compared to 2d shapes that they're largely
| worth it for interactive stuff. We might see more
| innovation on that front with a low-cost barrier to
| entry.
|
| Plus, scaling dynamic lighting up has _always_ been the
| Big Bad of computer graphics, and precomputation will
| _always_ give us an amazing heuristic to use against it.
| Everything else basically tends towards not mattering: we
| can only absorb a finite number of details, but we live
| in a world with virtually infinite lights.
| teamonkey wrote:
| Honestly, I thought this was the most practical and
| usable example of AI generation I've seen to date. I
| actually found it _refreshing_ after all the guff we
| usually see.
|
| I bet in a couple of years it'll be standard for estate
| agents to show 3D views like this on their web sites,
| architects converting quick paintovers of existing sites
| to 3D models, improvements to Street View, and so on.
| Anywhere where you want a quick 3D view of a space based
| on a few photos taken on a smartphone and where accuracy
| isn't 100% important.
|
| For things like games, it still follows the existing
| photogrammetry workflow (with all of those problems), but
| it might reduce the number of photos needed to create a
| point cloud.
| quitit wrote:
| I think there should be a standard set of images for
| comparison, because I've never seen a mesh generator readme
| that wasn't impressive. I test each one I get my hands on and
| the results are often disappointing.
| noduerme wrote:
| I'm very unclear as to what is supposed to be happening locally
| here, but as soon as the demo finishes loading, it crashes
| firefox on my phone.
| lukan wrote:
| I assume WebGPU related? FF lacks behind in support and I
| assume they use it under the hood. Meaning I would try chrome.
| (I know, I know)
| tmilard wrote:
| Impressed with Bolt3D AI model ! - Speed of the 3D generation, -
| Accurate 3D mesh deduction. It's a wonderfull chock.
|
| I agree, this is the way forward: - "some photos" as imput. -
| Convenient, a camera is in every pocket (Smartphone).
|
| On WE, I have been trying for years to generate 3D from photos.My
| tool now works well, but there is still this big problem of the
| time it takes to "recreate" the 3D mesh from photos. I remind
| that photos are in ... 2D.Not convenient. Here is an example of
| my Tool's generation : https://free-visit.net/fr/demo01
|
| Here, Bolt3d takes away the 4 hours combersome work into a
| automatic process. Wahoo !
|
| So Bravo to the Bolt3d team of researchers.
| oplane wrote:
| Can 3D generation happen for terrain views on a map ?
| bhouston wrote:
| It doesn't seem to work that well. Once you move off the primary
| camera axis, like rotate around, you notice that there are many
| regions with only sparse resolution and there are gaps
| everywhere. It is totally unusable for anything.
|
| Sure, it solves for the primary view, but this is claiming it is
| a 3D scene reconstruction/inference technique and in that claim
| it only sort of works.
|
| For example:
| https://i.postimg.cc/43tj36jv/Screenshot-2025-03-20-at-8-52-...
| gcapu wrote:
| Where is this image from? Were you able to try the model?
| scyzoryk_xyz wrote:
| I'm imagining this approach being combined down the line with
| your typical photogrammetry approach for reinforcing quality.
| antonkar wrote:
| We'll hopefully convert an LLM into a 3D haunted house and
| finally democratize AI interpretability by having millions of
| gamers walking in them
| marianaenhn wrote:
| ok, thx
| burgerone wrote:
| What's the use case for putting AI into everything? Pretty much
| every AI product so far has been and still is subject to
| hallucinations and inaccuracies and on top of that it's hugely
| computing intensive. Sure, it's the best we have right now and it
| allows us to do things that were previously next to impossible
| with manual programming work, but it's far from being something
| that's actually viable. And what would be the use case for
| turning a picture into an approximated 3d mesh that is only
| really complete from one angle? LIDAR does a stunningly accurate
| job at that already, reproducibly (although granted that this
| cannot retroactively be applied to existing photos).
| flykespice wrote:
| Good good, now show me the topology
___________________________________________________________________
(page generated 2025-03-20 23:01 UTC)