A co-worker claims that if HTML5 is used to play a progressive download video on a mobile device, the entire video file has to download before it can start playing. I find that hard to believe, since there’s no RMTP streaming in HTML5… and if I play part of a video on YouTube, there is indeed a delay if I scrub to a later point in the video.
Can somebody confirm or debunk this story?
Missed the edit window and posted in the wrong forum… sorry, mods, can you move this to GQ, please?
HTML5 has yet to adopt a tag for streaming video which strictly limits its availability but that doesn’t mean there are not plug ins or work-a-rounds. So yes, you can use streaming video with HTML5 but not in it’s purest form.
This limitation is expected to cause problems as different browsers will adopt different way of making streaming video possible with HTML5
Monster Encoder is already an example of this. Much of the problem stems from the widespread acceptance of h.264 as an excellent high quality video encoder but it also subject to patents so major browswers want to avoid it as much as they can
OGC is the open source free equivilent of h.264 an of course the debate over whether it’s as good can take a few days of Googling to read the arguments over whether or not it’s true
The websiteDive Into HTML5 (dot) org is a great place to start learning about HTML5
It probably depends on the browser used. Depending on the video container format, the web server and the browser, it’s perfectly possible to do progressive downloads and skips using a plain HTTP server*. AFAIK, the main benefit of non-HTTP streaming protocols (or “extended” HTTP protocols, like youtube uses for flash) is the ability to automatically switch bit rates to match the transfer rate.
Maybe, but that doesn’t mean the whole intermediate file is downloaded during the delay. That just means the player is downloading a couple of seconds of video at the new point before it resumes playing.
I assumed the mention of the video taking time to start from a new position was evidence that the whole didn’t need to be downloaded. If it did, then the device would already have the video and have no need to buffer.
Thanks for the responses, everybody. I’m convinced that my co-worker is wrong, as he was unable to provide any cite.
I realize that this reply is a bit dated. It has been nearly 8 months since the last reply. Here is my experience.
I have uploaded a 2-hour home movie to archive.org. I uploaded mp4, ogv and flv versions of the movie. All three files are roughly 2 gig in size.
I have a web page that plays the movie using the HTML5 video tag with a Flash fallback for older browsers (pre EI9).
Using the same wireless Internet connection I tested the page with IE 8.0, Chrome 11.0, Firefox 4.0, Safari 5.0 and an iPad running iOS 4. With the exception of Safari, the movie started playing after only a few seconds of buffering.
My conclusion is that with the exception of Safari all of the tested do use true progressive downloading. It appears that the entire file must download before Safari will play using the HTML5 video tag.
Apple iOS 4 uses a true Safari browser. However, the videos are played using Apple’s Quicktime; which, does use progressive downloading.