The thing that is amazing about imagemagick/graphicsmagick is the fact that they can open 100+ formats. That is not something to be taken for granted.
Other than that... they feel kind of old. Video++ feels really new, I'm excited to try it out, but I will miss support for all those weird formats I love so much.
I don't see any video examples though - or is that all handled using opencv?
I've found opencv's png/video decoders to be extremely picky. Opencv doesn't open some files that other libraries will and because of that I've been extremely hesitant about using it as my main image IO.
The video++ project is still very young. It's goal is not to re-code all the image formats, but just providing nice abstraction to build fast image processing apps. You can easilly load whatever format you like and pass the buffer and its dimensions to vpp::image2d to process it. I am actually using opencv to open images or grab images from videos, and then process them with video++ via the opencv bridge.
If video processing is what you want to do, I can very much recommend Avisynth[1] (and while it's Windows software, it runs quite well in Wine). For a more natively cross-platform alternative, there's also Vapoursynth[2], but I personally prefer Avisynth's DSL much more over Vapoursynth's Python scripting, not to mention it's generally got more filters available for it.
Other than that... they feel kind of old. Video++ feels really new, I'm excited to try it out, but I will miss support for all those weird formats I love so much.
I don't see any video examples though - or is that all handled using opencv?
I've found opencv's png/video decoders to be extremely picky. Opencv doesn't open some files that other libraries will and because of that I've been extremely hesitant about using it as my main image IO.