Welcome to manim Discussions!

[3b1b]

👋 Welcome!

We’re using Discussions as a place to connect with other members of our community. We hope that you:

• Ask questions you’re wondering about.
• Share ideas.
• Engage with other community members.
• Welcome others and are open-minded. Remember that this is a community we
  build together 💪.

To get started, comment below with an introduction of yourself and tell us about what you do with this community.

[TonyCrane]

Hi, everyone!

We just enabled the discussion page. Because this is a more free discussion area than issues.

Here, you can ask questions you meet when using manim in the Q&A category.
You can also share the amazing works you have made using manim and the good learning resources you find in show and tell.
In ideas, you can share the new features that you think manim can add, etc.
Of course, you can also freely share anything you want to say in the general category .

Hope that you can make reasonable use of this discussion community and enjoy yourself! :D

[gzhechu]

Hi, I am new to manimlib, and try to write some basic geometry animations. I don't know if it's OK to poste my idea here, so if it's not the right place, please let me know.
I noticed that manimlib uses only one CPU core to render and combine mp4 files. I don't know if there any guy tried use multi CPU core before.
So I did some modification on class Scene and Filewriter, add a "-j" parameter to specify how many job thread, and split all animations to several segement, and deliver to each thread, each thread has different 'start_at_animation_number' and 'end_at_animation_number', and then combine all the small pieces together.
I am using Ubuntu 20.04 with python 3.8, PIP installed manimlib. I met following problems, maybe some of you already know the issues before.

1. It seems each thread shared same 'ffmpeg' stdin pipe in same process, I'm not sure. It caused chaosed mp4 pieces.
2. So I turn to use numpy.save(each frame) to .npy files, and then finally combine to mp4 file, but meet 'cairo-path-fixed.c:839: _cairo_path_fixed_interpret: Assertion `!"reached"' failed'. It seems cario is not thread safe, I'm still trying to figure out.
3. OpenCV cv2.VideoWriter seems another way to dump mp4 file , but I must solve the cario problem first.
4. Now my temprary solution is: wrote a python shell script to parse -j and start several REAL manim process(not thread), and then combine the final mp4 file. it probably practicable.

[TonyCrane]

• Firstly, you’d better open a new discussion here, and choose the “Ideas“ category to post your idea.
• Secondly, the version here in the master branch is a stable version using CPU. If you want to improve it, you can go to ManimCommunity/manim, where the developers may be trying to achieve the same goal as you.
• Thirdly, the version here in the shaders branch is under development, which uses GPU to render, so you don’t need to worry about the efficiency.

[gzhechu]

Got it, thank you. I've clone CE version of manim, and started digging.

[mcurti]

• A similar question here. I use ArrowVectorField to generate a vector field from a class instance which uses a PDE toolbox to create the field. In manim scene I locate pointwise the points requested by the vector ArrowVectorField. Now the problem is that for larger resolution of arrows, this becomes extremely slow, so either I need to parallelize it somehow, or feed directly the arrays with values directly to ArrowVectorField. As far as I can see, at this moment I can use only a function that would accept the input pos
  
  .