PLASK.AI Review | Motion capture accessible for everyone

in #ai8 months ago

tb.png

Hey everyone!

Today's post is a bit different of what I am used to post, because today we are going to review a tool. First of all I was not (of course) paid or even in contact with the team who developp this tool. I have just discovered it on youtube and as I'm using 3d software such as Blender in my life and that I am very interested in 3d moeling / animation I was interested by the project, Plask.ai


What is Plask ?


Plask is a tool developp by a team of the same based in Seoul, South Korea.

Motion capture Source CC-BY SA 3.0
The goal of this tool is make "motion-capture" accessible to everyone, for those who don't know what motion capture is, it's the technology used in industries such as videogame's or movie's to recreate motion from a real object into digital software. Motion capture is known to be pretty expensive and not really easy to do, thus it was used only by big companies. The aim of Plask is to change tat and to make everyone able to use motion capture technology, that's why Plask is free (remember that nothing is really free) and is made to be easy to use. You just need a video of yourself, you can even shot it directly on the website using your webcam and then you movememt will be recreated by a character on their engine thanks to the power of artificial intelligence

How to use Plask ?


The first things you are going to need in order to use Plask are the following:

  • A Video of a movement made by a human that you want to recreate in 3D (I don't know what formats are accepted but I guess the most commons are (MP4, AVI...))
  • A 3D software, Blender for exemple, is an open-source free to use and yet powerful software, a lot of tutorial are available on the web and it may be one of the software with the biggest community.
  • And... That's all you really need in fact, the rest will be done by the software itself.



    First step, create an account, it's easy to do I don't even know why I tell you to do it as it seems obvious

    Then you can get started and you will face this interface:

    interface.png

    The panel on the right side represent the models that you can use my right clicking on them and click on Visualization you have a representation of them.

    The controls with the viewports are:

  • Arrow keys, control the camera
  • alt + right click: allow you to pan
  • alt + left click: orbit the camera (same as arrows)
  • mousewheel : zoom
  • alt + both click : zoom in / zoom out
  • By right clicking anywhere you can also trigger special views.

    The panel on the left is about modifying the character in the viewport, for more details about this you can read the documentation that is a good explanation of how it works

    Importing your video

    It is pretty simple alll you have to do is to drag and drop it on the website are click on the little + next to Library, then you can confirm that you want to extract motion out of the video, it will lead you to that interface:

    interface1.png

    (I am using an export here so that's not the best thing you want to use to extract motion out of it)
    On this interface you can select the fram from which you want to extract the motion and then click on Extract Motion to get the animation of your motion. That you can then easily apply to to one of the models that we saw just before.
    (Some videos are not rendered in this interface and I don't know why, maybe they are too heaving or too long I really don't know)
    To finally export your motion just right click on the model selected on the left panel and extract with th desired file format.

    You can finally import this file into the 3d software of your choice to recreate and animation with it.
    In blender the fbx file format seems to be the best but I had some issues with the textures not showing properly


    How good does it work ?


    So I decided to test some things by myself, the first thing I tried was to record myself head to toe and to do some movements. I then imported the result on Plask.ai but infortunately the result was really bad and it was not matching my movement in the video so it was kind of a fail but I was pretty sure it won't work that good in the condition I was using it.

    I then decided to use some green screen videos from youtube to recreate the motion of the video in blender, for each of the following videos I rendered the 250 first frame of the movement using Plask :



    I was really amazed by the result considering the low effort I had to put on, I then tried doing the same with the video they provide on the website to see if it was legit, and I have to say that the results were really good :

    https://plask.ai/mp4/motion.mp4


    Overall review


    Plask is a great tool considering the fact that it is free and that it want to provide an access to motion-capture-based technology, the AI is really great and the results are good as long as you use really clean videos. So I think its main purpose is well done.
    Considering the user experience, the tool is really easy to use but it is still a bit buggy with some tools not working really well or some button that you have to click a several times in order for it to work as entended.

    I think this tool is more something that people will use in order create draft for their project as it will be a cheap and effective way to make simple motion in a fast way. Very few people will use it for a whole project but it is a good start to create motion and then tweak it a little to have clean results, but the Videos used need to be really good also which can be hard to achieve for everyone, a smartphone may not be enough to create good video for this tool, so it is not really accessible for everyone.


    Thank you for reading that post, it was the first of this kind and I really wanted to show it to you because I think it is a good tool for everyone doing 3d aniamtion.

    Sort:  

    I agree with what you seem to imply here: if Plask is free, that means that Plask is not the product but its users are the product. They are probably building a kind of database from all information (i.e. videos) shared by the users with their application.

    This being said, the results you shared looks not too bad at all ^^

    PS: any chance to see you dancing? :D

    Yes I haven't read th privacy policy or their term of use. But I think they use the videos for their purpose maybe to train AI on a bigger database or anything else.

    Yes the ones I shared don't look bad.

    Considering the dance as I was just dancing macarena and trying to move my body a bit I don't think it would be really interesting to share it 😅

    Well if it is not interesting, it could at least be funny, couldn't it? :D

    Thanks for your contribution to the STEMsocial community. Feel free to join us on discord to get to know the rest of us!

    Please consider delegating to the @stemsocial account (85% of the curation rewards are returned).

    You may also include @stemsocial as a beneficiary of the rewards of this post to get a stronger support. 
     

    Your content has been voted as a part of Encouragement program. Keep up the good work!

    Use Ecency daily to boost your growth on platform!

    Support Ecency
    Vote for new Proposal
    Delegate HP and earn more