Automatically convert to video about motion-canvas HOT 8 CLOSED

It's a hard no for me, unfortunately.

But this is where I draw the line. If there's a need for more complex integration with ffmpeg it should be done via a plugin and not here in the core package.

from motion-canvas.

The FFmpeg exporter is now available as a standalone plugin:
https://github.com/motion-canvas/exporters#installation

from motion-canvas.

Motion Canvas is not meant to be an end-to-end video editing software.
It's supposed to be used in conjunction with Premiere, Da Vinci, Audition, Audacity, etc.

I'm open to a plugin interface that would allow you to create an ffmpeg exporter.
But I definitely don't want to include ffmpeg as a core dependency.

from motion-canvas.

The proposal is simply to make it such that, if the user has ffmpeg installed, they get this QoL boost (is there any common use for wanting the individual frames?). It could be as simple as a try { child_process.exec('ffmpeg ...') } catch {}, which wouldn't require any additional dependencies.

from motion-canvas.

The proposal is simply to make it such that, if the user has ffmpeg installed, they get this QoL boost

Oki but then the post you linked in your proposed solution is a bit misleading because it uses @ffmpeg-installer/ffmpeg.
Anyway, I'm willing to allow it. Two things though:

• It should be opt-in. I have ffmpeg installed and I don't want it to run each time I render something.
  I propose a new flag for the motion-canvas command, for instance --ffmpeg
• The options passed to ffmpeg should be configurable. Maybe also through the flag.

But this is where I draw the line. If there's a need for more complex integration with ffmpeg it should be done via a plugin and not here in the core package.

is there any common use for wanting the individual frames?

I'm not sure if I understand the question correctly, but If you're asking for the reason people may use image sequences then:

• Lossless quality.
• Transparency support.
• Ability to re-render only a portion of the animation.
• No point in encoding a video just to re-encode it again in another software.

from motion-canvas.

@aarthificial How about adding configurable lifecycle hooks to motion-canvas? Instead of building ffmpeg as a specialized flag, have a "post-render" string that is run as a command after motion-canvas is done. I imagine there are probably a lot of things people might want to do with the frames once they finish. Like maybe one person's workflow is actually to make a gif using image-magick. The specific ffmpeg command for rendering could then go into the docs so that people can add it if they want. Granted, that requires that they read the docs, but it decouples your tool from the hundreds of ways people might want to use the frames and still allows people to have their convenient workflows.

from motion-canvas.

@Ross-Esmond That's pretty much what I mean by "a plugin interface that would allow you to create an ffmpeg exporter".
But I think implementing something like this will take some time.
I'd like to switch to using sockets instead of HTTP requests first.
And also explore the possibility of using webpack's hooks architecture in order to not reinvent the wheel.

A simple flag seems like a good temporarily solution

from motion-canvas.

ffmpeg is pretty intrusive dependency for projects due to cross-platform installation requirements, and it not being quickly installable via the CLI (excluding the usage of package managers).

On the other hand, would you be willing to allow the WASM ffmpeg port for exporting directly from the browser? Understood if not, but it'd be nice to just get a "save" button after a render and to be able to get that into your downloads.

If I get the go-ahead, I'd be happy to implement.

from motion-canvas.