[HN Gopher] Sakuga-42M Dataset: Scaling Up Cartoon Research
       ___________________________________________________________________
        
       Sakuga-42M Dataset: Scaling Up Cartoon Research
        
       Author : snats
       Score  : 59 points
       Date   : 2024-05-17 13:38 UTC (9 hours ago)
        
 (HTM) web link (arxiv.org)
 (TXT) w3m dump (arxiv.org)
        
       | anonylizard wrote:
       | Wow, this looks like the beginning of the end for in-between
       | animation.
       | 
       | A large chunk of an anime's budget goes to in-betweening.
       | Essentially human interpolation of graphics between two key
       | frames (Its usually like 6-12 inbetweens per key frame). People
       | hate this job, and it is highly unproductive, so generally
       | outsourced by Japan to other countries.
       | 
       | Western animation decided to abandon it altogether, and move
       | first to flash, then to 3d animation. But in retrospect that was
       | a mistake, as it lost so much of the creative flexibility of 2d
       | animation. Anime today is substantially bigger than western
       | animation as a result. Crunchyroll has 13 million subscribers.
       | 
       | AI will solve the problems with 2d animation. Something like SORA
       | fine-tuned on anime-keyframe data like this. Can probably easily
       | solve in-betweening. Then the 2d animation workflow will dominate
       | 3d. Its so much easier to just draw a beginning and end key-
       | frame, then have the AI fill it in. Than to model and rig and
       | render a entire scene.
        
         | CactusOnFire wrote:
         | This is one job I am actually hopeful will be fully automated.
         | I've heard before that the people doing this 'tween' frames
         | make less than minimum wage on contract.
         | 
         | That being said, I'm sure this is going to take a while to
         | fully replace the manual efforts. There will probably be
         | awkward phase in between the outset and the perfect modelling
         | efforts, and I'm sure lower budget shows (or ones looking to
         | cut corners) will be the early adopters.
        
           | anonylizard wrote:
           | Anime already cuts corners by using cheap 3d models in place
           | of 2d hand drawn objects.
           | 
           | The more rigid the object, the better this works. 3d cars
           | look better than 2d cars, even in an otherwise 2d show. Mechs
           | look a bit worse. And human characters look horrible.
           | 
           | Yet anime studios still do it. Including for critical
           | highlight scenes like dance scenes (Check out Love live dance
           | scens), because it is so, so hard to draw humans dancing.
           | 
           | So if anime studios are willing to do something, that looks
           | obviously bad, as a widespread practice. There'll be 0
           | barriers to AI inbetweening adaptation, which would likely
           | look BETTER than human inbetweening within a year of release.
           | 
           | AI anime art has already wiped out the lower-end of patreon
           | artists, and is heavily impacting the mid-tier. Because AI
           | has gotten more technically proficient than the average mid-
           | tier artist. Pretty much only the higher-end can hold their
           | heads above water. Or they have to transition to drawing
           | comics with storylines, instead of just simple images.
        
             | GaggiX wrote:
             | >3d cars look better than 2d cars, even in an otherwise 2d
             | show.
             | 
             | This is quite debatable, if you notice that the car is a 3D
             | object, then something is already wrong.
        
               | Hamuko wrote:
               | Usually you can tell by the fact that nothing is wrong,
               | since the 3D model is very consistent and on model. A
               | hand-drawn car is usually not that.
        
               | GaggiX wrote:
               | The same could be said for a 3D human character, it's
               | very consistent because it's a 3D model but it's horrible
               | to look at.
        
               | mlyle wrote:
               | His point, I believe: artistically interpreting the
               | motion and shape of humans or objects with larger moving
               | parts makes animation look more on-style.
               | 
               | But for "boring" rigid objects, there's less of this
               | advantage; hence, the consistency benefits often are more
               | important.
        
               | numpad0 wrote:
               | unless it's Miyazaki films - in which case most humans
               | are intended to be lifeless rigid objects and every
               | machines are to be lifelike animations (/s)
        
               | speedgoose wrote:
               | Do you notice when Bluey characters are animated from 3D
               | or 2D? The software they use allows to do 2D drawings
               | from 3D animated models.
               | 
               | https://www.celaction.com/en/celaction2d/
        
               | GaggiX wrote:
               | I have never seen Bluey but from the software you linked
               | it is clear that it looks 2D because of how inconsistent
               | the character looks at different angles, for example when
               | you rotate the character the mouth changes position, the
               | hands jump from one sprite to another, it's cel shading
               | with a lot of 2D element on top, it works with simple
               | animation but for anime I'm not optimistic.
        
               | Hamuko wrote:
               | I've seen 3D animation where the people are still quite
               | fluid and not awful to look at. Not as good as 2D
               | animation but still pretty good and more than watchable.
               | 
               | https://www.youtube.com/watch?v=OO9zNw_uHg4
               | 
               | https://www.youtube.com/watch?v=eCc4md8Cuy8
        
               | redwall_hp wrote:
               | It was a major consideration for Initial D, which is
               | pretty much the definitive car anime. Animating movement
               | in tandem with dynamic camera movement is very difficult
               | (also why shows like Love Live use 3DCGI for dance
               | scenes, and why Disney was using 3D elements in films
               | like Beauty and the Beast) and modeling accurate vehicle
               | physics in hand drawn animation is also difficult. It
               | simply wouldn't have been viable without 3D animation.
               | 
               | https://youtube.com/watch?v=YDqKsQu9el4
        
             | TJSomething wrote:
             | A lot of the dancing stuff is about the ability to spin the
             | camera around a moving subject to the music, which is quite
             | difficult otherwise.
             | 
             | There's a lot of impressive work in 3D animation that looks
             | quite good. Outside of Bandai Namco's work on idol anime,
             | Studio Orange has made some of the best looking 3D modeled
             | anime lately and a few other studios have been getting into
             | it. I'm more familiar with video game animation, where
             | Arcsys Works has made great strides too, by using animation
             | on threes, manual tweening, stretch and squish bones, and
             | carefully UV mapped textures for crisp color boundaries.
        
             | Keyframe wrote:
             | Friend, but anime was never about quality of animation. In
             | fact, it was a prime example how to cut corners to get to
             | animations. That was always the case. It doesn't reflect on
             | the quality of character designs, environments,
             | storytelling, camera action, directing, etc. Motion was not
             | one of them; Never was. Anime is the first place I'd expect
             | to see new ground breaking, just like it was with all the
             | tools from 90's onwards (Toonz, anyone?).
        
           | guyomes wrote:
           | > I've heard before that the people doing this 'tween' frames
           | make less than minimum wage on contract.
           | 
           | An employee of an animation company describes in a comic book
           | his experience on working with people drawing the in-between
           | frames. They were paid literally with rice bags [1].
           | 
           | [1]: https://en.wikipedia.org/wiki/Pyongyang:_A_Journey_in_No
           | rth_...
        
         | mcphage wrote:
         | I wondered about that--they stated that they processed videos
         | to pull out keyframes. It makes me wonder how much that will or
         | won't help with in-betweening.
        
       | noobermin wrote:
       | Doesn't this dataset very obviously violate copyright?
        
         | omoikane wrote:
         | The idol dancing example from is clearly from "Oshi no Ko", and
         | the Nyarlahotep example is clearly from "Haiyore! Nyaruko-san".
         | They seemed to have adopted an "ask for forgiveness" instead of
         | "ask for permission" approach with respect to copyrights, and I
         | don't think that's the right way to go.
        
         | molave wrote:
         | It's used as research. How valid is this usage as fair use?
        
         | ronsor wrote:
         | It's a list of links, and they're not rehosting the media, so
         | it's as legal as any search engine or other collection of
         | links.
         | 
         | Of course, for a user to use the dataset, they'd have to
         | download it. Whether or not that's a copyright violation
         | depends on their local laws.
        
       ___________________________________________________________________
       (page generated 2024-05-17 23:01 UTC)