The (slightly handwaving) explanation:
All video encodes the pixel colors in each frame using channels, either red-green-blue (RGB) or luminance-chrominance (YUV). Those 3 channels only define the color, not the transparency; for that you need a fourth 'alpha' channel. Since it's impossible to record alpha with a video camera, nor display it on a TV, all the 'consumer' video formats like H.264 stick with 3 color channels to make the files smaller and the processing easier. Video codecs that have the 4th alpha channel are either designed for a very niche application (such as Quicktime Animation, an ancient format created for doing graphic overlays like the lower thirds on TV news) or they are designed as high-end "mezzanine" formats for post production (such as ProRes 4444, or EXR image sequences) that have massive data rates to preserve every ounce of quality. You have to choose either a rarely-suported codec, or a monumentally huge file.
But, since you can't play a transparent video to an audience, whatever you want this for it's eventually going to be composited over something else then encoded to a delivery format, like H.264. Now all the transparent pixels are full of background, there's no alpha channel any more. A huge file that never leaves your computer is no problem.
If you are sending the video to someone else for compositing (e.g. selling lower third animations to a TV station) then ask them what they want, and supply it. If that means posting hard drives instead of uploading a link, so be it. They're paying.
... View more