Internet video’s performance/quality vise

by Robin Harris on Wednesday, 5 December, 2007

Internet video is about where film was 100 years ago
I was talking to a company who will be announcing a video infrastructure solution when the CEO mentioned something he called the “video performance/quality vise.”

Here’s the problem: a video stream requires both capacity and bandwidth. Higher quality video requires more bits per second and more capacity. Bandwidth and capacity both cost money.

So as Internet video quality rises, the financial cost to provide the video rises too. An HD video stream is 4 Mbit/sec.

500,000 channels and somethin’ on
As cute as YouTube, et. al. are, they suck. Movies are small, picture and audio quality awful, and viewing options limited – like films 100 years ago.

Bandwidth limitations are part of the problem, at least here in the US. But those are being addressed, however slowly.

What happens when Internet video becomes competitive with broadcast TV in quality? Popularity will soar. As TiVo has shown, people love choice. And the Internet will have the most choice.

The price/performance/popularity vise
Digital Fountain’s raptor codes will change the Internet landscape for video. High quality video will drive be much more popular, just as long-form movies took film to the next level.

Bandwidth costs are dropping fast to pennies a GB. So infrastructure costs – especially storage – are critical to Internet video’s commercial success. The more popular it gets, the more storage will be needed. It is a huge opportunity.

The StorageMojo take
Massive data storage is still a very young technology. The ultimate cultural impact will be more profound than film because of the many-to-many nature of the Internet and the low barriers to entry. Should be fun!

Comments welcome, please. I don’t think the firm wanted me to mention their name, so I haven’t. If we get that cleared up I’ll update the post. Or maybe wait a while to write about them.

Update: Joe, thanks for catching the 4Mbit mistake I made. I corrected it above.

{ 7 comments… read them below or add one }

Tracy Reed December 5, 2007 at 8:35 pm

My computer is way faster than 10 years ago but I still have the same speed cablemodem, same cablemodem price, and pay about the same as well. We really need to do something about increasing the last-mile bandwidth in the US.

Wes Felter December 6, 2007 at 1:02 pm

Oh man, you bought the Digital Fountain PR? They’ve been pushing FEC for years and I’ve yet to see any impact. I doubt video will change that. Real, Apple, and MS spent years developing optimized protocols for video, and what took over the market? HTTP. Even in the rare chance that the market decides to adopt FEC, it will probably be a free, non-patented variant like LDPC or good old-fashioned Reed-Solomon. Worse is better.

random_graph December 6, 2007 at 5:57 pm

You said it Wes, FEC is certainly a cool thing, but DF has been around almost as long as online video, the scalability & quality problems have always existed, and DF certainly hasn’t changed anybody’s landscape! Bless their heart though for pursuing some interesting stuff….them and the wavelet CODEC startups. Unfortunately, existing protocols and codecs have demonstrated an uncanny ability to be “good enough”, and content owners mostly lean towards being risk averse.

IMHO your claim about storage cost is worth some clarification. It’s not about popular content, it’s about *unpopular* content. For popular content, amortize the cost of storage over the Opex of distribution, and cost of storage gets very SMALL. The place that declining cost-per-GB really brings elasticity is the long-tail. When the fixed cost of onlining an asset drops below the level of marginal distribution value (either tangible or intangible), then you encode it and put it on disk. Witness http://www.archive.org/details/movies.

Robin Harris December 6, 2007 at 9:18 pm

Tracy – American’s have been arguing about whether the federal government should drive infrastructure improvements since about 1790. The railroads were given huge tracts of land to build the transcontinental lines at no other cost to the taxpayers. AT&T was given a monopoly for 70 some odd years. Thank goodness the feds built the freeway system.

In the late ’90s we gave the telco’s a couple of hundred billion in tax breaks to take broadband everywhere. They took the money and did – nothing. So when they whine about net neutrality and how hard it is being a telco it is just business as usual.

Wes – yes, I bought the PR. I’m late to their game and don’t have their history, though the fact that they are 6 or 7 years old did give me pause. Nice demo though!

Random – you are absolutely correct on the economics of unpopular content. Thanks.

So it is just a matter of bandwidth to the home increasing enough to support higher quality, eh guys? OK, telcos, stop whining and start building bigger pipes!

Robin

Wes Felter December 7, 2007 at 1:59 pm

These digital video people are just lazy. :-) I recently discovered that YouTube encodes at 350 kbps and was shocked. Who has broadband that’s less than 1 Mbps? Likewise, the people who can actually watch HD video have fast enough connections. (Obviously you shouldn’t ‘t force HD content down the throats of people who don’t care, but I don’t think any sane person is suggesting that.) For example, I can watch 720p trailers while they are downloading on my PS3.

Let’s end the blame game and just get video out there – 1 Mbps SD for the masses and multi-Mbps HD for the early adopters. If there is anything that concerns me, it’s how many ads will be needed to pay for 10x the bandwidth.

Bill Todd December 7, 2007 at 2:27 pm

Hmmm – since I corrected your misunderstanding about Digital Fountain a couple of weeks ago at your ZDNet blog, there’s little excuse for you to have repeated it here.

Their technology is not about video on demand but about broadcasting the *same* content simultaneously to a whole lot of people (where different recipients may suffer loss of different packets and could overload the *server* and its immediate environs – not the network at large – if they all asked for retransmissions). For transmission to a *single* recipient the technology always requires *more* bandwidth than traditional retransmission of lost packets would, and offers little in the way of jitter-avoidance as well (since the way you avoid jitter in either case is to delay the display on the recipient such that your buffering will ride out any likely transmission glitches).

– bill

Joe Miller December 10, 2007 at 12:55 pm

When you say “an HD video stream is 4MB/s”, do you really mean 4 megabytes/sec or 4 megabits/sec?

4 megabytes/sec = 32 megabits/sec.

That’s a lot of bandwidth and a lot of storage, but is not necessary for HD video, as some corners of the internet are showing already (see the forums at doom9.org)

First, let’s review the most popular codecs for HD video, in particular the 3 video codecs allowed on Blu-Ray and HD-DVD discs: MPEG2, VC-1, and H.264.

MPEG2 on blue-laser discs is pretty much the same as the codec used on standard DVD. The compression ability is not great, but CPU requirements are low. So, to fit HD video inside MPEG2 will indeed take more space. I would say at least 10-20Mbps to get a “decent” stream of 720p. Maybe 25-30Mbps for 1080p

VC-1 (WMV-HD) and H.264, however, offer a much more compelling storage/bandwidth ratio. They do this at the expense of CPU, of course. So people will need higher end CPUs or a newer graphics card with offload capabilities to watch these kinds of videos.

Now, if you search the net, you will find a lot of people re-encoding their BluRay or HD-DVD titles from the massive 15GB-50GB discs down to 4GB and 8GB. They do this by shrinking the 1080p content to 720p and lowering the bitrate.

720p would be a great format to deliver HD streams on the internet. 1080p is currently overkill, but we could grow into it.

In my own experience, it’s not all that challenging to compress a 30GB 2 hour movie from MPEG2 1080i or 1080p down to a 4GB 720p H.264 video with a bitrate of between 3-6Mbps. What you end up with is an HD movie in the same storage footprint as a standard-def DVD. The only catch is that you need some beefy CPU horsepower to play it back smoothly. The quality is still very very good and an enormous improvement over youtube and standard DVD.

3-6Mbps is probably reasonable for cable modem subscribers (and bored office workers), but some DSL modems may have trouble.

Long story short, I don’t think it’s a question of only capacity/bandwidth, but also of CPU (and software). We could be providing more HD video streams today at relatively low bandwidth/capacity levels, but the software and CPU power to play these back may not be all that widespread yet.

Leave a Comment

Previous post:

Next post: