Needs advice
on
FFMPEGFFMPEG
and
OpenCVOpenCV

I want to make a video template where we can change images and make a new video with template animation text and song but the images are our and the image which is chosen from a user can convert into different shapes just like cartoon or childhood or old age and at the last show both images in the template video. How I can do it in Android Studio?

READ LESS
5 upvotes·23.3K views
Replies (2)
Lead Software Tools Engineer at Leonardo UK·

Since I believe Android Studio now supports Python plugins &/or invoking Python scripts I personally would use MoviePy (https://zulko.github.io/moviepy/) as a wrapper around FFMPEG &

ImageMagick (https://www.imagemagick.org/script/index.php) - this combination will allow you to have a script that takes in an image or video, resizes it to a specified size, adds the text, (either scrolling it in or fading in/out or even with some special effects), then saves the output video. There are example scripts on the MoviePy site on how to do these. While you could do it yourself with FFMPEG & ImageMagick plus command line parameters I personally find using MoviePy much more clear & maintainable.

If you want the end user to be able to supply the image/video then you will need to set up a server somewhere that can take in the POSTed image and return the processed results.

READ MORE
4 upvotes·9.1K views

Hi! I will start by asking questions before giving specific advice, because there is the need to understand better what you want to do, and there will be a number of steps and choices to take, therefore there is not "the one" advice to give here.

First of all, afaiu you want to implement the described video processing on Android?

If that is the case, you will have to create or use native bindings to Java runtime (JNI i.e Android NDK) in order to access FFmpeg or OpenCV or any other C lib itfp. My first advice is therefore, you should check how feasible that is (existing bindings working for your target JRE etc).

Further, i find the description of what you want to produce quite unspecific. I guess parts of what you want to do with either sourced videos or images can be achieved with ffmpeg or opencv based processing.

In the end you just need to start writing some kind of processing steps with either. Once it gets more specific in details you can probably ask more specific questions on how to use these libs or even other tools.

READ MORE
2 upvotes·9K views
Avatar of Steve Barnes

Steve Barnes

Lead Software Tools Engineer at Leonardo UK