Web video streaming: Which is the best?
A new study examines five Web video streaming technologies: Flash, QuickTime, Real, VX30 and Windows Media. Researchers evaluated “image quality, streaming quality, accessibility, encoding and Web-authoring efficiency, and cost.”
Perhaps more interesting to educators and online journalists — the report includes tutorials for using each of the five technologies.
Edgar Huang and Clifford Marsiglio, researchers at Indiana University-Purdue University Indianapolis, conducted the study in May and June 2006 and self-published the results in August 2006. The findings rank the barely known VX30 first, Flash second, and WMP, Real and Quicktime third, fourth and fifth, respectively. Good news for Flash (see VX30/Zentu 4 Flash), bad news for Quicktime! And 36 percent of the computers in the study were Macs.
Our study shows that VX30 has the best overall performance while Flash outperforms other streaming technologies in terms user accessibility, perceived image quality, initial buffering delay and rebuffering. [more]
Update: It’s important to understand that VX30 is a codec, and Flash video (for example) can be encoded with more than one codec, e.g. On2 or Sorenson.
Update 2: VX30/Zentu may be controversial.
Technorati tags: online | video | technology | streaming | Flash
I feel like you were baiting me with this post. QuickTime has nothing to worry about, it has the iPod.
Doesn’t, “For easy and fast access and interaction, we have decided to publish the study at this permanent Web site,” really mean, “We couldn’t get this published in a refereed journal (or unrefereed technical magazine) so we’re putting it online”?
It’s interesting to note that they streamed the Windows Media and QuickTime streams off a Real Helix server rather than a Windows Media Server or QuickTime Streaming Server, respectively. I wonder how performance might have been different on their native servers?
I could poke a whole lot of other holes in it (QuickTime can stream over http on it’s native server, for example), but I don’t have time this morning.
Note: the comment author is a former copy editor at a refereed academic journal.
August 29, 2006 at 1:24 pmI am hoping for any relevant comments from knowledgeable folks like Craig! As for the refereed journal thing … well, sometimes you do research because you are curious, and you might just want to share the whole enchilada with the public. I prefer to think that was the motive here.
I would never “bait” anyone … well, almost never …
August 29, 2006 at 3:36 pmHi Craig,
Glad to see cynicism is alive an’ kicking! As for publication, we have been contacted before and after about publishing these results in both pop and academic journals. We may still publish more traditionally based upon these results and others like them, but this was an experiment that we wanted to do. This was a project based around a ‘New Media’ institution and as such, let’s get it online and include ways for others to edit, append and comment. Comments from one editor of an academic journal in this area were very positive. I do miss the peer-review before publication a little, but the information is sound.
Personally, I don’t publish to see my name on some haughty-taughty rag anymore. Its nice, but honestly, forwarding a quote from the NYT or Time Magazine to my parents actually seems more rewarding. Besides, have you seen the folks they let work on academic journals these days???
The idea of the study was to review several well-known technologies being run on their native streaming engines. Generally educational facilities buy into the entire package if they are going to make a commitment to a single technology. This was the assumption we made going into the study. The fact that any of the five technologies can be served over standard http is irrelevant – the fact that Zentu VX30 is designed to run this way from the start, however, is relevant.
As for VX30 being a Codec and Flash being a container technology, yes, this is correct. In the context of the study, as mentioned above, we were measuring the entire package being sold. The VX30 has a larger user base in commercial business than one would generally know. It really is a sneaky technology, especially the one we chose to use based around Java technologies, thus deeming it a ‘playerless’ technology. Most operating systems by default come with some version of Java these days. The same cannot be said for Flash, WMV or QuickTime. At one point, it looked like Real was going to partner with enough OS Manufacturers for this to happen, but greed on all parts overtook this occurrence.
The Codecs used on the individual container technologies were almost always the most advanced technology available, with the exception of Windows Media of which we had to scale back to ensure that one could play these on ‘alternative operating systems’ (insert smiley here). I use a Mac as my daily driver and the latest WMV-9 wouldn’t work. The latest Flip4Mac technology seems to play this technology ‘occasionally’. But this was released after the study.
And as for VX30’s ‘controversial nature’, trust me – this was a known. I support F/OSS application development and have released a few GPL based applications myself. The developer of this technology has made himself no friends in the community. But this is science and we all know scientist have no morals, so we proceeded blindly. Sadly, Dr. Huang wouldn’t allow me to push this paradigm further by utilizing videos of monkeys pulling kittens tails. There are just some things that even immoral science types won’t do.
But as for QuickTime, my biased nature still believes this to be a great technology. It uses published standards and utilizes the open technology that best serves its market. As of the results of this study, it might be said closed is better – but I still have hope. Using free tools for encoding and transcoding, QuickTime seemed anecdotally the best served by having these standards available. Utilizing some of the free tools for others, it was a crapshoot to see if the technology would work until I tried playing it back on another machine. There is something to be said about openness that I still believe in – even if their most famous technology in play right now is wrapped in DRM that hides the openness away from the light.
–
All in all, we welcome almost all comments on this topic. Heck, I think we’d be even willing to entertain cynics that seem to live only to poke holes in arguments, but I’d rather have folks join in collegially and make the experiment work so that this information can be shared ‘for the greater good’.
If anyone is interested in helping expand this knowledge, we have a Wiki available but our email addresses are published for all the spambots in the world to take advantage of as well. We will periodically revisit the static content and update it with this information where we find appropriate – and this in itself is a reason to eschew traditional publication. No sense someone picking up a 4-year-old article from a dead-tree rag and taking choose from a transient and nonstatic technology study and basing their decision around this.
August 30, 2006 at 12:14 pm