18
github
7y

Automatic(or Manual) Devrant Cartoon Creator

Project Type
Open source idea
Summary

Automatic(or Manual) Devrant Cartoon Creator

Description
This is gonna be either extremely complex or not feasible project. Anyone seen devrant cartoon clips based on a rant. Example: https://youtu.be/KyFoJv3Otfg I would like to initiate discussion on how we can contribute towards it. Either we can provide an Open Source Cartoon Maker Studio Software(if already existing, then a devrant flavored plugin) which enables everyone to easily build a cartoon or clip for any rant with predefined characters and voices. They should still have flexibility to add new items and that will be available to others too. OR Some intelligent system which given a rant, extracts its Context and provides a bootstrap cartoon of it.
Tech Stack
TBD
Comments
  • 3
  • 2
    Example Devrant Cartoon:

    https://youtu.be/KyFoJv3Otfg
  • 1
  • 2
    In order for this to take place, dfox should provide a sorting option only for these cartoons bcuz the feed will soon be filled with cartoons.
  • 3
    @CozyPlanes we can scope down to rants with no image attached.
  • 3
  • 2
    @github kinda complicated
  • 4
    @github this sounds like fun ;D
  • 1
    @nickj I accept.
    What's your opinion on creating cartoons for a rant will be appreciated by this community?

    Have you seen devBanners. It was simple and nice. And people liked it.
  • 1
    @github difficult. Potentially worthwhile.
  • 2
    @theNox yup. Though either it's complex or I am too dumb to think of a simple solution. Or atleast start with a simple POC.
  • 1
    @github And you want the feed to be filled with devToons again, lol

    Why don't you come to our team and create some frontend? (basic web stuff without fancy frameworks)
    https://github.com/devBanner
  • 2
    @github sounds like a good idea, I however have no clue how to make this. I would help but school is a bitch.
  • 2
    @github maybe ask trogus for the avatar layers and move them around using some kind of keyframe system?
  • 1
    @CozyPlanes devToons.. haha. That's a good name. 😂👍

    I looked into devBanner, it uses .NET and no offense, I just don't like C#(for no reason, just that currently I am working on improving my Java and Python and don't want to start a new language)

    Btw, you will be credited for naming the tool if it can happen some day.
    I am little more inclined towards using ML to derive the entities that can be useful and later dive into producing a cartoon.
    I haven't personally started any research(===just normal googling) yet, in the phase of brainstorming with everyone here.
  • 1
    @github also, I would probably use electron to make this due to the fact that that's what I'm most comfortable in and electron usually isn't appreciated here.
  • 1
    @Dacexi it's fine. I can understand the school part.

    And great suggestion.

    @dfox @trogus tagging if you can share some thoughts.

    Summarizing some premature requirements:
    1. Scope to only include non image rants
    2. Using avatars to build dynamic characters in the devToon™ (named by @CozyPlanes)
  • 1
    @Dacexi Well. I know how comfortable you are with electron.

    Let's discuss regarding the stack later. Let's see how much requirements can be defined and check it's feasibility.
  • 1
    @github uhh. Way to complex for my time right now I believe
    We're speaking of full animated movies right?
  • 1
    @Kimmax Currently, yes. Best case is the same as the YouTube animated one.
  • 1
    @github to be honest I don't think that's going to happen
    It's just sooo may things to consider, you will probably have to render some thousand frames where you render each object in, layer by layer, and encode afterwards
  • 2
    While it sounds really good, it is indeed quite complicated as everyone mentioned. But good luck with it! I will try to contribute if I have time.
  • 3
    @CozyPlanes can I join the devbanner team? I'm a frontend person.
  • 0
    @ewpratten
    NP! Go https://github.com/devBanner
    And find the gitter chat join button somewhere.
  • 3
  • 1
    This is waaaaaaaaaaaaaaay too complicated, it is not that you throw some code and end up with a good result. But nothing is impossible :)
  • 1
    @CozyPlanes I've done some thinking and this might not be that hard. Instead of encoding everything, why not just use electron which has built in screen recording?
  • 2
    @Dacexi @github and a keyframe system should be quite average difficulty to do. I'll see if I can set aside some time after Christmas to help :)
  • 1
    @Dacexi I don't know about keyframe systems. Will look into it.
  • 2
    @github look inte animejs
  • 1
    Just add js at the end of all the 'features'/problems. (Not helpful commentary sorry)
  • 2
    I think basically you would need a:
    -script/input
    -ML parser/tokenizer to extract the scenes
    -a library of image assets and templates
    -and then some ML app that takes each sceneDef and transforms it to a visual using the assets

    I'm just saying.... But def would be interested
  • 2
    @billgates thanks for your input. Summed up pretty well.
  • 2
    @github np, and for the time being, technically u can replace all the ML parts with Human.

    For scripts, i guess you can just pick the top rated stories of the month/week

    But yea i'd like to do some open source stuff but mostly backend... work is just too boring and it aint gonna get any better...

    Maybe try making a few manually, just the storyboard/sketch and then see if which parts can be most easily automated/generalized first.

    I might actually try one sometimes.... cant really say when... today is basically my "day off" for the next couple of weeks.
Add Comment