Jump to content

The ultimate community for Ruby on Rails developers.


Photo

How involved to create jib-jab style video personalization software?


  • Please log in to reply
3 replies to this topic

#1 Estando

Estando

    Passenger

  • Members
  • 2 posts

Posted 14 February 2014 - 08:18 PM

Hola.

 

Not a programmer.  Just talking to some locals here about a project, and trying to get some broader perspective on how involved it would be to create software to automate the process of personalizing videos.

 

More or less what JibJab does.  Not looking to compete or anything.  I'm a videographer looking for a fun and interesting way to share my vids via social media.  I know how to use After Effects or other tools to do this, but looking to have end users just upload their photos, do some basic sizing adjustments, and press go... and have the system do the rest where their face picture follows the preset motion tracking I've programmed in for the video.  Basically, they get their faces on top of my video doing a goofy bobble-head cartoon thing.

 

A couple local guys have mentioned using Rails as their overall framework, so thought I'd ask here to try and get some general sense of the scope of the project.

 

thx.

 

Estando



#2 james

james

    Guard

  • Moderators
  • 221 posts
  • LocationLeeds, U.K.

Posted 15 February 2014 - 11:56 AM

The software to do this already exists, imagemagick and avconv are an awesome combination.

In Rails terms there are gems available to handle the uploading, ropping/rotating etc... of images that use imagemagick behind the sceneses. In fact Imagemagick is at the heart of most recognisable image manipulation software.

 

There are some Railsasts that show how to do this sort of thing but essentially you need to learn how to use avconv and imagemagick

 

Some further reading

http://railscasts.co...ge-manipulation

http://railscasts.co...cropping-images

http://railscasts.co...-images-revised

You will need to pay for a subscription for that last one

 

There are links on the railscasts to the tools that are used and you can find out more about avconv here

http://www.roxlu.com...io-using-avconv

and the official documentation here http://libav.org/avconv.html

 

It's just a case of learning the tools, once you understand how to use them on the command line you just need to put it all together in an app


Programming is just about problem solving!


#3 Estando

Estando

    Passenger

  • Members
  • 2 posts

Posted 15 February 2014 - 03:48 PM

Wow.  Great stuff.  Thx.

 

Looking into.

 

Ultimately, I'd like to have it set up where I give people on a website instructions about how to upload the photo, easy crop tools, identify corners of mouth or whatever, some names for the credits, and the contact info where they'd like it delivered, then the system does the rest meaning it moves the mouth to the vocal audio track, tracks the head to the motion capture to match the movement of the body, fits inside whatever hats or helmets might be present, has multiple sets of faces that appear and disappear to match the scene cuts, puts the user-supplied names in the credits, renders the video, then outputs it automatically to a yt channel, or emails it along with a link for how to purchase an HD version.

 

Worst run-on sentence ever, but you get the idea.

 

Since I'm already familiar with the general functions of After Effects, I'm also digging into whether there might be a way to use scripting between it and Photoshop actions to take care of the steps after the photo is uploaded to cut it out, insert it, have it follow the AE motion tracking, etc.  I'd still need to design a front-end interface, and figure out what to do with it after it's rendered, but there is at least a potential solution there for all the bits in-between.  Ultimately, I'd like this to be running on a hosted server somewhere.  Not sure how that would work exactly as I've never tried to run software anywhere other than my computer, but I'm assuming it's doable.  There are power and connection issues where I live.

 

Again, looking at the moment for a general direction to dig into, so is there any particular reason that comes to mind why it might be preferable to go the imagemagick/ avconv route for the video overlays vs doing some scripting in After Effects?



#4 james

james

    Guard

  • Moderators
  • 221 posts
  • LocationLeeds, U.K.

Posted 15 February 2014 - 09:30 PM

As I am not familiar with After Effects I really can't advise one way or the other, If after effects will do the job for you then go with what you know is the best advice I can give right now.


Programming is just about problem solving!





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users