iPlayer TV streams bandwidth could be 120-times higher in 5 years

4th September 2008

People in the TV industry are predicting that on-demand viewing will account for 40% of all TV viewing by 2013 (the headline of that article is confusing — the 40% figure they're predicting is in the second paragraph). So although might already look high , it's actually small-fry in comparison to what it would be in five years' time if on-demand viewing does hit 40% of all TV viewing.

In the UK, 51.4 million people watch TV, and the average viewer watches 26.4 hours of TV per week. So, assuming that these figures remain at the current level, on-demand TV viewing in 2013 would amount to:

On-demand TV hours viewed per week = 40% x 26.4 hours per week x 51.4 million people = 542.8 million hours per week

If the BBC iPlayer accounted for, say, 20% of all on-demand viewing in 2013, and assuming that the average bit rate of an iPlayer TV stream will be 1 Mbps (that's likely to be a conservative estimate, because there should be a lot of HD stuff available by then, and HD streams use bit rates in the 8 Mbps+ range), the bandwidth for the iPlayer TV streams in 2013 would be:

iPlayer TV streams bandwidth in 2013 =

That's about 120 times more bandwidth than the iPlayer TV streams were using in May this year, and it equates to the bandwidth increasing by a factor of 2.6 each year for the next five years (or 8.312% each month), assuming that the rate of growth were constant over the next five years (the rate of growth wouldn't be constant in reality, because the growth would likely be higher earlier on and it would tail off later, but assuming that growth is constant makes the analysis far simpler).

The following table and graphs show what the iPlayer TV streams bandwidth would be over the next five years using the above assumptions:

iPlayer bandwidth estimate from 2008 - 2010

iPlayer bandwidth estimate from 2008 - 2013

Comment on the accuracy of these estimates

There are a number of assumptions made about what will happen in the future in the calculations above, such as that the average bit rates that iPlayer TV streams will be using, and the TV industry is only guessing about the fact that on-demand TV viewing will make up 40% of all viewing by 2013. However, the important thing is the magnitude of the figures, and I don't think they will be far off. For example, if on-demand only made up 30% rather than 40% of all TV viewing, then that would reduce the bandwidth by 25%, but you would still be talking about the bandwidth being 90 times what it is today.

I actually think that the average stream bit rate will be a lot higher than the 1 Mbps figure I've assumed by 2013. The BBC has already launched new higher quality iPlayer TV streams that are using 800 kbps, and the BBC has said that it wants to let people start watching the iPlayer on their TV sets, and once that becomes popular the bit rates will need to be pushed up again because TV screen sizes are a lot bigger than computer monitors, so the picture quality would be poor if the bit rates stayed at 800 kbps. Against that, there is the effect that improvements in the efficiency of H.264 video encoders will have (this reduces the bit rate required to provide a given level of picture quality), but that would have a much smaller effect on the required bit rate than the move from people watching on computers to TV sets.

Also, as mentioned above, the bit rates used for HDTV streams are in the region of 8 Mbps+, so if the BBC makes a lot of programmes available in HD-quality by 2013 then the average bit rates could be far higher than the 1 Mbps I've assumed — and I would expect the BBC would want to provide HD iPlayer streams by 2013 because BT and Virgin will already have rolled out superfast broadband to millions of homes, and people would want to be able to watch HD quality stuff on their shiny new fibre-optic broadband connections.

The least predictable variable is what percentage of all on-demand viewing the BBC iPlayer will account for by 2013. On-demand viewing includes timeshifting programmes on PVRs such as Sky+, as well as watching the true video-on-demand services such as the iPlayer, 4oD and ITV.com Internet services and Virgin On Demand. But PVRs haven't sold in the very large numbers that some people thought they would, so my estimate of the iPlayer accounting for 20% of all on-demand viewing seems reasonable.

Finally, I doubt the number of people watching TV or the number of hours per week that people watch will change much over the next five years, because people already have the Internet, DVDs, games consoles and so on.

Overall, my gut feeling is that the above estimate of iPlayer bandwidth will most likely be on the low side, mainly because the average bit rates could easily be a lot higher than the 1 Mbps I've assumed, and this would probably more than offset any reductions in other variables.

Bookmark with:
Digg	del.icio.us	Reddit	Facebook
Google	Stumbleupon	Slashdot