Blender-made movie Flow takes Oscar
88 comments
·March 3, 2025jsheard
eurekin
Cycle's renders are beautiful, but 10 minutes per frame can be a hard sell... I wonder, if anybody tried rendering in cycles to output eevee's primitves. I remember that was one of the tricks that architecture rendering community used - just paint with lights in places that a full blown global rendering/path tracing would do.
leonidasv
Now I'm curious how the film would look if it was rendered in Cycles, there are some lighting aspects that really feel "off". Perhaps now that the film is acclaimed they could release a remaster done in Cycles.
Fluorescence
I kind of hope they don't. I like the humble, democratic, FOSS spirit - it's like Dogme 95 / "Vows of Chastity".
"rules to create films based on the traditional values of story, acting, and theme, while excluding the use of elaborate special effects or technology. It was supposedly created as an attempt to "take back power for the directors as artists" as opposed to the movie studio."
https://en.wikipedia.org/wiki/Dogme_95
I had a negative initial reaction to the animation style but it hooked me in and blew me away. It had virtues far more vital than render quality. In contrast, I bailed on "Inside Out 2" and have no interest retrying. I hope more people are encouraged to create lofi meaningful movies instead of thinking it's the preserve of billion dollar studios and sweat-shop animation factories.
Neywiny
Haven't watched yet but that was my gut reaction. When the engine first got stable released I tried it and was impressed at how quickly it got to a 90% solution, but the now complicated lighting scenarios that it couldn't handle took me back to cycles.
teruakohatu
I am out of touch with the latest and greatest Blender features. If best/highest fidelity renders are required, can Blender scale with a render farm?
Last I heard that was the advantage of the propriety/in-house alternatives.
orlp
An 1.5 hour movie at 24 FPS has ~130k frames to render. As long as you have less machines than that the parallelization is essentially free.
hamaluik
Yup! https://www.sheepit-renderfarm.com/home Is a great community example of this.
ad-astra
I made my own distributed render orchestrator that supports Cycles + custom plugins. It uses Modal’s cloud compute APIs to spawn jobs on up to 20x containers with an L40S GPU (like 80% as fast as a 4090 with tons more VRAM) each. It ain’t cheap but it’s absurdly fast, and much easier in terms of cash flow than outright buying the equivalent GPUs.
https://github.com/stoicsuffering/distributed-blender-render...
zoky
Any embarrassingly parallel task can scale almost infinitely by throwing more resources at it.
touisteur
I think I've seen some amazing Blender hacker put Cycles to the test on a machine with both NVIDIA and Intel GPUs. Love it that their API seems that portable and able to parrallelize on heterogeneous hardware. Amazing software work.
sznio
a stupid-simple approach would be to split up the render betweeen machines by manually starting it on each one and setting different frame ranges to render
pier25
Weren't Blender working on a more efficient cycles renderer?
pilaf
You're probably referring to Cycles X [1], which if I'm not mistaken has already been released.
It will never be on-par with Eevee's performance though as they are fundamentally different approaches to rendering: Cycles is a physically-based path-tracing engine, while Eevee uses rasterization through OpenGL.
pier25
You're right.
I think I confused it with the Eevee Next project released last year.
https://code.blender.org/2024/07/eevee-next-generation-in-bl...
raverbashing
Honestly yeah it won't be "perfect" but neither is a videogame being rendered in real time and it looks pretty good
Since they're not going crazy with effects it seems like a good compromise
null
legitster
I did not find Flow to be a technically impressive movie. The animation was very imperfect. The rendering (especially shadows and textures) were off. The whole movie looked like a video game cut scene.
But oh boy, what an amazing cutscene to watch. I'm worried that the story the media is putting forward is that this was an innovative and cutting edge movie - based only on a superficial appreciation of the (stunning) art design. But the real story is how the director worked within his limitations to make something equally enjoyable and meaningful as the other guys.
Most importantly, this movie passed the Actual Kid (TM) test. My 7 year old and his friends sat raptured through the entire movie without any slapstick, pop music numbers, or even dialogue! Not once, but 4 times now!
brundolf
A takeaway may be that cutting-edge rendering doesn't really matter for cartoon-stylized films, especially for kid viewers
stvltvs
Yep, grew up on cartoons like Transformers, GI Joe, Thundercats, etc. Looking at them now, they are laughably bad in most respects, but they sparked our imaginations and didn't need to be sophisticated to do it.
ChrisMarshallNY
But some of the really old (like 1940s) cartoons were very smooth and well-done. I have DVDs of old Tom & Jerry cartoons, and they are excellent.
My experience, is that the ones made in the 1970s and 1980s had crap quality.
I watched this movie, and think it very much deserved the Oscar, but the character rendering was a bit “scruffy.” The environment rendering was great, and it looks like they optimized for movement, in the characters, which was a good choice. Once I spent some time, watching, the rough rendering didn’t matter.
I had a similar experience, watching Avatar. At first, it seemed like a cartoon, but I quickly became immersed, and the fact it was rendered, didn’t matter.
I read, somewhere, that the movie is being re-rendered. I think they may have the money for that, now.
Hoasi
If the goal is sparking the imagination, these flaws are often a feature, not a bug. You have to do a little bit of work to complete the picture. That's also why the original book is almost always better than the fully rendered movie inspired by the book. No matter the budget.
agumonkey
yes, there's a balance to get and visual "perfection" is nothing real, even star wars had blunders and visibly lesser tricks, but the whole created a deep sense of wonder and you got along
kevinventullo
I think Nintendo figured this out years ago. Their games are not graphically cutting-edge, but they still sell like crazy.
Liquix
nintendo has distilled their own flavor of "disney magic". their flagship games are polished beyond belief, the art direction is a careful choice. they hit the borderlands bullseye over and over - not particularly cutting edge to a gfx professional but unique, cohesive, and beloved.
deskr
The shaky camera is a deal breaker for me. Doesn't matter if it's animated or not, TV series or film. Shaky camera is an instant switch off for me.
crooked-v
I feel like it kind of fits in the same category as Hundreds of Beavers (also a fantastic film), as something using the roughness of low-cost methods as a genuine part of the artistic style.
watt
Recently I reminisced about Blender foundations first(?) effort, Tears of Steel, with the script like "Look, Celia, we have to follow our passions; you have your robotics and I just want to be awesome in space!" - "Why don’t you just admit that you’re freaked out by my robot hand?!"
It's not about the textures and shadows.
seba_dos1
"Tears of Steel" was the fourth Blender Open Movie project. The first one was 2006's "Elephants Dream", then "Big Buck Bunny" and "Sintel".
From the more recent ones I highly recommend "Sprite Fright".
tzs
That's kind of surprising. Academy members are not required to watch all the nominees for Best Animated Feature before voting. In fact they are not require to watch any of them.
Several years ago I remember that after a year where the movie that won best animated was not the one that those in the animation industry overwhelming thought was sure to win some animation industry magazine survived Academy members asking which movie they voted for and why.
What they found was that a large number of the voters thought of animated movies as just for little kids and hadn't actually watched any of the nominees. They picked their vote by whatever they remembered children in their lives watching.
E.g., if they were parents of young children, they'd vote for whatever movie that their kids kept watching over and over. If they no longer had children at home they would ask grandkids or nieces or nephews "what cartoon did you like last year?" and vote for that.
Another factor was that a lot of these people would vote for the one they had heard the most about.
That gives Disney a big advantage. How the heck did Flow overcome that?
Inside Out 2 had a much wider theatrical release in the US, was widely advertised, made $650 million domestic, is the second highest grossing animated movie of all time so far worldwide, and streams on Disney+.
All that should contribute to making it likely that those large numbers of "vote even though they don't watch animated movies" Academy members would have heard of it.
Flow had a small US theatrical release at the end of the year. I didn't see any advertising for it. I'd expect a lot of Academy members hadn't heard of it.
As a guess, maybe Moana 2 is the movie that the kids are repeat streaming. That was not a nominee so maybe those "vote for what my kid watched" voters didn't vote this year and so we actually got a year where quality non-Disney movies had a chance?
CmdrKrool
Interesting. I loved Flow and I'm glad the stars aligned for it on this particular occasion. This article [1] lists a bunch of other Oscar-related firsts:
* Gints Zilbalodis, who is 30 years old, is the youngest director to win the Oscar for best animated feature.
* Flow is the first fully-European produced and funded film to win the feture animation Oscar.
* Flow is the first dialogue-less film to win the feature animation Oscar.
* Flow, made for under $4 million, is by far the lowest-budget film to ever win the category.
It also says the winner of the animated short category, In the Shadow of the Cypress, was unexpected since the Iranian filmmakers couldn't do any of the usual in-person campaigning of Academy voters due to visa problems.
[1] https://www.cartoonbrew.com/awards/underdogs-win-latvias-flo...
chillee
A couple things:
1. The academy has had a significant increase of young voters in the past 10 years or so. Generally speaking, young voters are more likely to take animation as a "serious" medium.
2. These interviews were always somewhat overstated. Of course some voters have stupid rationales, but I don't think this dominates the academy.
3. Disney's Inside Out 2 was nowhere close to winning the award this year - Flow's biggest competition was The Wild Robot, which did gross far more than Inside Out 2, but far below Inside Out 2.
If you look at the past couple years, The Boy and the Heron (Studio Ghibli) won over Across the Spider-Verse (with Pixar's movie Elemental nowhere close) in 2023, Guillermo del Toro's Pinocchio won over Across the Spider-Verse (with Pixar's movie Turning Red nowhere close) in 2022, etc.
I'm curious what year you're thinking about above. Perhaps Toy Story 4 over Klaus in 2019?
stuart78
I would guess this is, to some degree, a generational shift. The Animated category has only existed for ~30 years and was born from the resentment many in the academy felt toward Beauty and the Beast being nominated alongside supposedly serious films for Best Picture. Each generation following that one has grown up with a more diverse slate of animated films available.
The Oscars are the slowest possible reflection of social change, and I’m sure the perspective you share is still held my many members, but this win holds out some hope for sure.
dmazzoni
The Grammy awards for music are the same thing. Members aren't required to listen to the nominated albums, and every member gets to vote in every category.
I had a friend who was a Recording Academy member as a classical musician. He thought it was strange that they asked him to vote for the best hip-hop album since he doesn't listen to hip-hop at all.
So for many of the categories that are a little more niche, it basically turns into a popularity contest, rather than the opinion of true experts.
SoftTalker
Reminds me of when Jethro Tull won best Hard Rock/Metal category, beating out Metallica and AC/DC.
lukasb
Flow was a critical favorite, sometimes that matters at the Oscars.
tootie
I adored Flow. It's hard to say it was truly "better" than Inside Out 2. I think part of the calculation has to be that everyone expected Pixar to deliver something top notch so it only really met expectations. Flow was made by a no-name team from Latvia and was really something unique and interesting. I went into it kinda blind with no expectations and was blown away.
riffraff
I didn't think inside out 2 was a very good movie.
It had good ideas but didn't do very well with them (contrary to the first movie, which was great). I'm not surprised a movie which wasn't "just a sequel" managed to beat Moana and IO2.
autoexec
I was also disappointed by inside out 2. I thought it followed the story beats of the first one a bit too closely. Flow was the better film.
viccis
As someone who started using Blender before 1.8, posting on the old blender.nl forums before its move to BA, it's just been pretty insane to watch it reach this point. Back then it didn't even have ray tracing, and all of the attempts to make long form videos with it were very very rudimentary.
owenpalmer
I remember being a little kid learning Blender, rooting for it to become the industry standard. It's amazing to see how much the project has grown.
kylecazar
I heard him thank Blender in the first few sentences and had to Google it to see if he was talking about THAT Blender!
I remember doing all the tutorials when I was younger and considering game dev.
dang
Recent and related:
First time a Blender-made production has won the Golden Globe - https://news.ycombinator.com/item?id=42620656 - Jan 2025 (49 comments)
maxglute
It shows, Blender has come a long way, but FLOW doesn't look technically incredible. On the otherhand, I just rewatched Shrek recently, and complex graphics isn't everything.
raxxorraxor
What do people mean with technically impressive? There are Blender renders that look quite incredible though and you just cannot differentiate it from real picture anymore
Example: https://images.squarespace-cdn.com/content/v1/58586fa5ebbd1a...
There are probably some flaws here as well, but you need to study the picture in detail. And Flow used the fast renderer of Blender, not the quality one.
Still, it does have a unique style that is much more interesting than many other animated movies. So what is technically impressive, just throwing more compute at it to make it photorealistic?
I think art style will have a larger impact. In a way it is technically impressive as it didn't need a lot of compute power.
netruk44
[delayed]
WhatThisGuySaid
I think people are specifically referring to the movie not looking technically impressive, not that Blender isn't capable of technically impressive renders at all.
raverbashing
No kidding, Shrek probably had to do with 100x less computing (per hour) than a modern production. First Toy Story probably something like 1000x less computing
We did come a long way
bolognafairy
Take a look at the earlier concepts / renderings for Shrek! Before the studio gutted the team(?) and told everyone to pull their heads in. Absolutely off-putting.
febin
Animation of Flow (Blender Conference 2024) https://www.youtube.com/watch?v=fxz6p-QATfs
neoecos
This is the first movie I took my 3 yo kids to watch in a theater. My wife and I are FOSS enthusiasts.
mppm
I am honestly surprised by this. Congrats to the makers and to the Blender community, of course, but to me Flow looked more like a feature-film-length demo reel. And not even the most impressive demo reel, visually speaking. Compared to all the other animation films out there... I don't think it would rank even in the top 100 for me.
lemming
I’m very pleased to see this, Flow was such a lovely film. I didn’t realise it was made with blender!
Not only was it made with Blender, the final renders were done with Blenders semi-realtime Eevee engine rather than its max-fidelity Cycles engine. That reduced the compute required by orders of magnitude - the director said a render farm wasn't necessary because his local workstation could produce final-quality 4K frames in 0.5-10 seconds.
"Proper" production renderers like Cycles do look better of course, but having an alternative which is viable on a shoestring budget is very valuable.