Make This Startup: HD Video Serving Service
Here I am, the night before Christmas, thinking about what I’m going to do with the high-definition camcorder I bought myself. Naturally, I am going to use it to share some HD video on this site to spice up reviews and so on. The problem lies in how to get that large amount of content from the server to your display.
Of course there are tons of video sharing and hosting services out there but not many support HD. Furthermore, the sites that do support HD video don’t let you embed that HD within your site. For example, after a brief email exchange with someone at Vimeo, it was confirmed that HD consumes too much bandwidth to let Vimeo users embed it on their site and currently only SD content is embedable. The rest of this post is the type of thing that goes through my mind when I have a realistically achievable idea.
Audience
Independent publishers looking for a bulletproof and quick solution to serving up high-definition content on their website. These users don’t care what is involved as long as it is reasonably easy, holds an uptime SLA and is fast. As such, they don’t mind paying bandwidth and membership/service fees.
These independent publishers could be anyone from professional bloggers to blogging networks and even larger publishing houses like newspapers expanding online. Example uses include publishing HD video interviews, product reviews, and anything else video is currently used for, in higher quality.
Value Prop
Reliable and fast high-definition video embedable within any site and the ability to fully-customize encoding, resolution, bit-rate, et cetera. Since HD content is too large resolution-wise to fit as a direct embed on most sites, content will likely be displayed in a lightbox format laid over the site when activated.
Technical Details
- Amazon S3 for storage and hosting
- Amazon EC2 for converting videos through ffmpeg
- Flowplayer or a homebrew flash player to display the resulting .flv files, with ability to degrade gracefully to SD or lower HD for users on slower connections
- Web interface for uploading videos, inputting ffmpeg video encoding parameters for more advanced users, managing video statistics (views, downloads, et cetera)
Business Plan
There are several routes to take for monetizing the service. When dealing with HD content, I don’t think anything can be offered for free. We’re talking about a lot of bandwidth consumption here. However, since the intended users are more professional than the average consumer they won’t mind paying one of the following:
- a) a flat rate membership fee with various tiers depending on their average traffic/bandwidth
- b) a DIY route with the user supplying their own Amazon AWS credentials and X.509 certificate for this service to use their own AWS account AND a standard monthly or yearly membership fee.
The latter (b) would be geared towards more tech-savvy users that feel better having their content easily accessible on their own AWS account. Furthermore, they could setup a CNAME to have the content appear to be served from their domain and remain brand-friendly.
The Time is Right
The latest beta of Adobe Flash 9 player supports the H.264 video codec for HD content in addition to HE-AAC for audio. This means less CPU is required to play superb 1080p full-HD content. However, since not all people can display such a large 1080p video on their screen without downscaling, this proposed service would be smart enough to detect capabilities and load the proper size or load a user-selected size (480, 720, 1080).
HD content can saturate an Internet connection by consuming over 2,000 kbps. Fortunately, the Internet populace is more prepared for this now than in past years. As a last resort users can always opt for an SD video or wait and buffer the HD version.
Check out Adobe’s HD gallery to see the capabilities of the latest flash player combined with superb HD content. Imagine watching video close-ups of the latest gadget and actually being able to discern fine details, contours and more, rather than a 400px wide fuzzy video.
Digital Web Magazine has a great article about HD Flash for the web and The New York Times had a recent article regarding upcoming Intel processors with the SSE4 instruction set that facilitate video compression and HD content viewing.
Development
A rough draft could get off the ground with ~6 months of development time and 4-5 hackers working off of $150k of initial funding, not including capital spent on potential partnerships with related companies using technologies this service could benefit from (ex, if Adobe had some great service to facilitate server-side video encoding).
Potential Snags
HD sounds expensive and it is. Lots of bandwidth, lots of storage and lots of processing power will be required from the get-go. Usually startups don’t have to deal with scaling until some time later, not so with any HD venture. Fortunately, this service can benefit immensely by outsourcing the hard stuff to Amazon, but it’s not free.
The prospective user-base is fairly niche but as HD becomes the de facto standard on the web, small publishers generally devoid of IT prowess will rely on this service. Proving to those publishers that the value prop of this service is right for them will be easy but finding the perfect price point will be tricky to say the least.
Acquiring funding might be a challenge as well since unlike every other startup out there now, the main website won’t be getting too much traffic. It would serve only as an interface for the publishers to upload videos and look over stats/settings. While the concept could evolve into a HD video-sharing site, that’s definitely not the initial purpose and would waste more bandwidth than necessary.
Thoughts
Do you think this has merit? The web is naturally moving to high-quality this and that, with video most-definitely taking the hot seat. If you were a small, independent publisher and wanted to serve HD video reviews, news, and so on to your readers, would you use this service?
I would use this service to get the word out. With so many people doing the same things, its hard to be unique :)
Great article Paul!
Talk soon….
I love the way you have structured the post. But I think having stayed in US and then going back to India and seeing you tube not working properly … I think HD on web for US would be what YouTube is for India.
I think practically - some sort of a netflix model would work easily - unless something special happens for bandwidth.
But I would ask this basically - Is there a need? If so what is the market? I certainly don’t see the market for 8MB photos version of flickr! Hope you get my point.
Niche market however will surely exist. Disney, Big well-known independent artist. you have to think of it in a bootstrap way. Not for common you and me :-). Of course, unless you are rich and in the US/South Korea.
Didn’t know Vimeo couldn’t embed the HD video content, that’s kind of a bummer. Sounds like a good idea since I’d bet many people are disenchanted by YouTube’s quality limitations.
Running Windows at least, watching HD movie trailers is a pain because Quicktime launches externally (and sporadically) and why bother with WMV-HD. If studios published in FLV format I would go there to watch new trailers for sure.
What camcorder to you get?
I manage technology for a medium sized church and we would be very interested in being able to serve HD quality video, both on demand and live (if possible, this may need to wait until multicasting matures). We currently use Amazon S3 for delivering some of our content and it works great but we lack the technical expertise to create a platform that would be convenient for other churches and similar venues to use.
just wanted to say that this is one of the most level-headed observations i’ve seen on the current state of HD - particularly from the cost/benefit analysis perspective. i would also encourage you to dig deeper into the scalability, and more importantly potential performance issues related with running a large-scale video serving business off the s3 platform, spoke-n-hub vs. the edge model of cdn, etc.
I agree that the time is right. However with HD in mind you may want to go down the set-top box route for distribution. This is because I certainly want to watch that content on my plasma and not on my laptop.
Secondly the bandwidth problem presents huge technical hurdles. Not to mention the net neutrality debate, where the ISPs will come after you (remember BBC iPlayer).
This service will need to exploit P2P like architecture, where content can be downloaded with minimum hops (if it exists). Neither Adobe or Amazon can help here. You will have rely on the likes of Kontiki or Limelight.
PS — What about IPTV players? Can’t they enable such an infrastructure?
SmugMug just recently added DVD quality videos to Power accounts and HD video to Pro accounts.
http://www.smugmug.com/help/video3
@hacintosh - “Power users can add video clips up to 2.5 minutes long (5 minutes for Pros).”
There’s a time limit and I’m not sure if that video is embedable on a website, just viewable on theirs, similar to Vimeo. Also it appears as though the video is just MOVs etc and not converted to HD Flash 9 for better loading times. I’m not 100% but that’s what I got from briefly checking out the site. Thanks for the comment.
I’ve been video-blogging in HD since September 2005 at http://techvideoblog.com
I’ve been posting my 500 HD videos in 1280×720 DivX 3.5mbit/s. 95% of computer screens support 720p and it looks great on those.
I think the new $200 720p camcorders from Kodak, Aiptek, Exemode are very interesting, hopefully one of these shoots good enough quality in terms of optics. If not I think the Sanyo HD1000, HD700, HD2 are very usable for HD video-blogging now. Samsung, Toshiba are also coming with some that look interesting. And of course the 1080i AVCHD from Sony and Panasonic. I’ve been using my Sony HC1 HDV since September 2005, so I am looking for a solid state HD camcorder upgrade, perhaps one of the $200 ones would do the trick for me, as long as the optical quality and encoding quality is satisfactory.
Someone beat you Paul. Stage6.com allows you to upload and embed high-def videos.
another one up for stage6.com, though you need the DivX Web Player. I kinda prefer it over most flash-based viewers any day. (stage6.com is owned by DivX)
Thanks for pointing out Stage6 Anthony. Unfortunately it uses the DivX player which is another thing people will have to install, where as Flash is already on most computers. Regardless, I made an account and am testing it out - but to me it’s just another file conversion I must make on my end before uploading. MTS > MOV > DivX AVI. I just uploaded a 4 second clip and it’s been 20 minutes and hasn’t shown up yet. I think I’ll continue my search for a service similar to the one outlined in this post.
Millions of people already have DivX installed. And the DivX Web Player is nice if you want to give people the option to progressively download, which means they can keep the file on their computer once they have finished playing it, or more precisely buffering it. Since they can set the default directory where all the DivX files should be saved to.
Stage6 sometimes is a bit unreliable, and it can be slow to have ones videos activated on it by its browser based upload feature.
Also bandwidth from Stage6 sometimes is not fast enough for HD content, especially to Europe, at least as far as I have tested it. So it’s just fine for DVD quality, but for HD quality sometimes it looks like the bandwidth provided is not fast enough. DivX says they have 10 million visitors per month on Stage6, so they must be using huge amounts of bandwidth, which can explain some of the occasionnal downtime or slow activation of new uploaded videos. But certainly the quality is better than anywhere else.
I guess I’ll have to make a specialty service for this. ;) I would consider making such a startup but, I wouldn’t know how big the audience would be. I wonder how many other people are looking for this kind of service. Also rates would get really complicated; If I had a $10 per month flat rate for unlimited bandwidth, I may make a killing or spend fortune.
I think Vimeo use Amazon S3 as it happens. My question, is it really necessary to embed video? Personally I have no problem clicking through to Vimeo / Youtube to watch a video. The Vimeo limits on videos (time and size), are terribly annoying, its not really suitable for anything other than sort clips- not high res dramas and documentaries. The issues with aspect ratios are enough to drive me away.
That said, personally I like the Quicktime delivery method on onegoodmove.org, having the option to download right there on the page but on anything other than the Mac at home, (read: work computers), Quicktime is a pain to use.
Paul,
Nice post. I have been working on a similar idea for the past 8 months. Just last night (here in Atlanta) I had a similar conversation with the guys at Konsole Kingz (see: http://www.konsolekingz.com) about a start-up HD distribution platform. They have a deal with MSFT and the the Xbox team - they produce most of the urban HD content on the Xbox Live network.
In our conversation we agreed that:
1) There are enough HD box (i.e. Sony PS3s, Xbox 360s and Tivo HDs, etc.) in the world to support a low overhead startup.
2) Openness of these boxes is key - I think that Sony has the right idea - having a browser on the box. If they would only implement the same RSS reader on the PS3 that is found on the PSP, the world would be a much better place for HD content.
3) There is a ton of content already in the wild that simply needs to be converted to the proper format, wrapped in a little meta data and published in a user friendly way. For proof, head over to the iTunes podcast directory and peep out the growing amount of HD content. There are also dozens of content start-ups that are producing wonderful HD - think MariposaHD - It’s so well done that it almost drove me to buy a new PC just to watch it.
4) In the end, RSS is everything and Torrents might be a waste of time. The ability to subscribe to content, and have a slow, direct download, is much better than hunting for something interesting (think Pirates Bay, Vuze and Stage6) and waiting for the peers to feed you. This is why the iTunes podcast directory is so popular globally. If this were not the case Vuze would be the answer - trust me it’s not.
In the end a graceful directory, workable publishing standards and amazing content will win out over any other (YouTube like) model.
Sites to sample:
http://www.ps3stage.com - example of a Stage6 hack that proves how painful a “wack” directory and user experience can be. Great way to find HD Divx content for the PS3 if you have time to hunt and peck.
http://www.MariposaHD.com - one of the best of broadband HD content. howerver, they are at the mercy of the torrent crowds sadly.
http://www.konsolekingz.com/blog/ - doing a great job in focusing on their niche, and MSFT has brought them an international audience via their very closed Xbox Live network.
Paul, if you would like to talk about this in person hit my up at jharris9999 at gmail.com. I might even have a few bucks to toss into the pot.
All the best and happy New Year,
James Harris
CEO and Chief Storyteller
Elemental Interactive / ListenShare
Atlanta, Ga
Sorry, the link is http://stageps3.com/ not http://www.ps3stage.com
James
This might be what you are looking for: http://www.ooyala.com/
They were the winner of the Amazon Web Services challenge: http://www.businesswire.com/portal/site/home/index.jsp?epi%2dcontent=NEWS%5fVIEW%5fPOPUP%5fTYPE&newsId=20071206006270&ndmHsc=v2%2aA1196946000000%2aB1197040954000%2aDgroupByDate%2aJ2%2aN1000837&newsLang=en&beanID=202776713&viewID=news%5fview%5fpopup
It’s a valid concept that is only just now becoming feasible. I would wonder if a plan could be put together targeting a niche with a free offering to keep the offering focused, increase the CPM of any ads if ad-supported, and fend off competition from any of the big boys just upping their resolution.
I’m not so sure of the AWS route - you already mention that it would need to scale early, and that bandwidth costs would be an issue. For all of its benefits, AWS pricing would chew through dollars much faster than comparable capacity at a colo.