The Venice project, a mysterious beta application from Kazaa and Skype creators Niklas Zennström and Janus Friis, has being variously hailed as the future of television, and the application which may finally bring to its knees the aging last mile bandwidth of the internet itself.
Retitled in mid January to ‘Joost‘, TVP is an IPTV application. That is, a program designed to carry high resolution video direct to consumers via the internet, rather than through satellite, cable or terrestrial broadcasts.
On the technical end, Joost uses both the UDP (to stream video direct to viewers) and TCP/IP (to share shows between users) protocols to create a hybrid Peer-to-Peer and Streaming, MPEG 4 H.264, (currently) free, on-demand TV network.
While Joost does live up to its promise to deliver full screen, uninterrupted streaming video at a watchable quality, a variety of potentially insurmountable challenges stand between the company and its goal of subverting broadcast television.
1. No Premium Content
Whilst Joost’s founders perhaps aim to ultimately provide licensed ‘A Plus’ broadcast content, current offerings are thin indeed, and there is little to suggest that the company has the industry connections or necessary expertise to lure rights holders into making such content available.
“Watch for sci-fi shows, rock videos, sports, comedy — anything with a testosterone angle. Deals are in the works with the three music majors, plus top US broadcasters and cable channels. For the rest of the world, there’s a modified PBS strategy: classic reruns, documentaries, and independent dramas.”
While there’s clearly a market for cheap, low quality television, creating a new distribution channel for low value content is unlikely to bring about the paradigm shift in multi-hour-a-day TV watching articulated by Joost’s founders in the latest issue of Wired Magazine.
As Zennström points out..
“You have to put together a whole consumer offering, a great instantaneous experience. A simple service that fills an obvious need and can be offered for free.”
Right now there’s no compelling reason to think that Joost can deliver on the ‘obvious need’ element of such aspirations. Kazaa and Skype were technological and infrastructural achievements, not content deals.
2. No User Content
It may be too early to tell, but early indications are that Joost, like the iPhone, will be a closed platform. Although the developers, unlike Mr. Jobs, welcome independent extension development, it appears they will avoid distribution of user uploaded content entirely. With online only content, and content production companies, from the Revision 3, TWIT, and Podshow networks to Channel101, growing in quality and diversity all the time, this has become the golden age of high quality, short format, web friendly media. At the other end of the scale, YouTube’s survival in the wake of massive takedown notices and competition from far more piracy friendly alternatives, signifies a huge unanticipated interest in the lower end of ‘user’ generated media; from inventive indy band videos, to post modern soap operas, to emerging comedians. According to Wired, the Swedes have no desire to tap this particular well spring of talent.
“Content that few people want to see — what Leiden engineers call “the too-long tail” — crimps a P2P network’s advantage.”
Whether you accept the sincerity of this reasoning or not (in fact, shows like ‘Diggnation‘ and ‘This Week in Tech‘ regularly beat the viewership of much cable television), if Joost’s founders follow their stated plan, they will fail to do for syndicated internet video what apple managed for the podcast, and in the process potentially waste their biggest advantage over ‘content providers’ like Apple, Microsoft and the major vertically integrated media corporations – rich, freely generated ‘sticky’ media.
3. Computers don’t deliver the TV experience
In contrast to active clip grazing or movie watching, computers are ill suited to the casual background parsing of TV. Theres a missing piece of the IPTV puzzle that Joost cannot in its current form solve. A link between the expanding, thinning, television and the computer. The key here is that connectivity must be bidirectional. It’s no use connecting your laptop to your Plasma via a composite cable (and in the process distracting your computer from any useful task), if changing channels necessitates a return to the keyboard. Technical solutions to this problem abound these days, with devices from companies like Slingmedia and Cisco prepared to carry high resolution audio and video to next generation HDTV’s. These are however, likely to be niche products when compared to the omnipresence of media centers in the form of the Microsofts Xbox 360, and Apple TV, both of which offer the potential to bring pay (and pay per view) programming to the television; in the form of all important big media licensed content. Arguably these players are in a far better position to ‘migrate broadcast television’s mass audience to the Web‘.
4. User Bandwidth
With grave doubts proliferating, as to the ability of the internet as it now exists to manage the load created by the growth curve of streaming video, and considering the effective and explicit bandwidth limits placed by ISP’s on their customers, it is by no means obvious that Joost can succeed in the mass market.
According to Joost’s detailed FAQ (beta customers only), in one hour’s use of the service “approximately 320Mb data will be downloaded and 105Mb uploaded”.
In Ireland, the broadband penetration poor man of Europe, common download and upload ‘allowances’ start out at around 10 GB download / 1 GB upload per month. That’s less than 10 possible hours of Joost watching, hardly enough for the program to replace broadcast television. Whilst this represents the worst peak of corporate bilking (leading to one of the worst broadband takeups in the EU), bandwidth limits are an uncomfortable reality throughout Europe and daytime and application targeted bandwidth throttling common in the US and UK.
5. Too Complicated
The Joost interface itself is a model of parsimony and slickness. It is also however, neither familiar nor obvious. Control of the program is via a video recorder / PVR interface metaphor, but in it’s current form more complex and ‘mystery meat‘ than either. Elements of the interface, like the channel library, used to manage channel subscriptions, and the ‘mychannel’ link used to open specific program items, remain unintuitive. A higher level interface design may prove necessary to ease the public into a world of fullscreen online television.
6. Advertising Specificity
Whilst Joost will offer advertisers a variety of viewer demographics – geographic, temporal, and viewing profile – this is not the kind of information which is necessarily most valuable to mass market advertisers. Whilst broadcast television ratings provided by companies like Nielson are deduced from representative samples rather than IP headcounts, they provide rich demographic information that includes income, employment and interests, for each data point. The difference, crudely put, is between ‘what kind of person’, an ‘what time and place’. Additionally, while Joost may allow advertisers to know which programming is viewed when, and which ads are ‘flicked’, this is a capability shared by existing PVRs, and one Zennström and Friis may well choose to leave disabled.
7. Visual Quality
Thanks to James Corbett, I’ve had a chance to sample the Venice Project over the past few weeks, and I can report that under XP at least, the visual quality is not up to the DVD standard claimed. Whilst the resolution may be as high, some channels have a washed out muddy look, more often associated with Flash video.
Check out a screen shot here (NOTE: This is a 2meg uncompressed BMP file!), a static camera, full resolution shot from the Green Day documentary ‘Bullet in a Bible’, you’ll notice it’s much lower quality than the reference images provided by Joost (check out a second screen grab here, from a different video, exhibiting higher quality). While this may be a function of the low end laptop I’m using, its a noticeable reduction in quality compared to watching a DVD or other H.264 content on the same machine.
For comparison purposes here’s another screenshot, a frame from an apple trailer played fullscreen in 480p, roughly equivalent to DVD resolution (again 2meg!). Perhaps image quality is increased on a higher speed connection? Our home machine is connected to Digiweb DSL XTRA (theoretical 3 meg down, 384k up), and about two miles from our exchange.
In summary, picture quality is certainly more impressive than rival streaming web formats, but not quite up to DVD. Does this matter? Well that depends a lot on point three. Sitting two feet from the screen its O.K., but I wouldn’t want to see a movie this way.
8. Not a Magic Bullet
Much like Zennström and Friis’s previous projects, Joost is not a radical advance in the state of the art, but rather a more reliable implementation. The technology to distribute high quality video via IP already exists, but is currently poorly implimented. What governs future uptake in this area is likely a mixture of legal (IP and regional distribution restrictions), economic (distribution costs), and psychological (ownership preference, tolerance of DRM) issues, rather than a best of breed technological race.
All in all there are many things to like about Joost, it’s interface, while likely too complex for casual users, is efficient and geek friendly; its expandability and image quality, almost instantaneous channel streaming, and the promise of ads as short as ‘one minute per hour’, are all commendable. However it’s ability to compete against free, burnable downloadable content, and branded services with a hook to the television, remains to be seen. There do exist however, at least two scenarios which could spell huge success for the upstart company.
1. Pay per view
If Joost can snag studio support, their service could provide a terrific (if lower resolution) alternative to ‘legal’ movie downloads, which if marketed correctly (read rented cheaply) could create a whole new market for lower resolution video – or cannibalize existing DVD sales.
2. Selling Out
The rapid skipless streaming technology, and piracy resistant (sic) encryption behind Joost, could make it an attractive purchase for use by a worried major TV network, suffering media giant, or net enabled set top box maker. Perhaps licenses to ala carte device or service tailored versions of the software and the backend network which supports it could be sold to multiple providers?
In the absence of either circumstance, Joost may well end up the Betamax of streaming video.