My immediate thought on this is whether multicast could mitigate much of the bandwidth issues with (linear broadcast) radio-over-3G, which is essentially two problems lumped into one.

You’ve got the problem of how to get enough bandwidth to sustain the stream from mobile device to base station (multiplied by however many users in a given cell are listening to a stream), and you’ve got a secondary problem of getting the stream from the server to the base stations.

IP multicast, assuming an amenable network architecture (and mobilecos tend to be a little better than traditional ISPs at leveraging this stuff to squeeze things through their networks), would solve the second problem quite neatly. It might require broadcasters to colocate a multicast relay server within the mobileco’s network, but cost-wise that’s peanuts compared to the various alternatives.

(I don’t actually know if there’s an over-the-air equivalent of multicast in 3G or LTE packet data, though you can guarantee that if lots of broadcast-type data applications start become prevalent, a future version of the standards will include one if it doesn’t exist today).

None of this is to say I think broadcast-radio-over-3G has a bright future (it’s too early to tell, I think), just that technically it’s entirely feasible on not very much of a budget…

I don’t reckon the real issue is the second problem; rather more, the finite bandwidth available on each base station.

http://james.cridland.net/inventions/mobdab.html is a little modest invention that fixes that.

Is it likely that the operators performing the final delivery would be able to charge the broadcaster?
That has not happened yet for regular internet traffic has it?

I remember UK ISPs complaining about the huge increase in broadband traffic due to BBC iPlayer usage – but requests for payment from BBC (except for its own direct network connctions) have been batted back.
http://news.bbc.co.uk/1/hi/7336940.stm

As the listener pays – via their phone contracts then it will be self-limiting … with mobile operators competing on tarrifs.

I think the “whether the network would cope with all this” is a – if not the – key point.

I took a walk to work this morning. It’s five miles, from a village to a county town with a population of about 100k. Sure, it ain’t London.. but it’s a not untypical way of life for tens of millions of UK residents.

I was listening on my iPhone as the battery in my FM radio had died. I got rock solid 2G signal the whole way (“full bars”), with strong 3G for the last bit. Even on a modest 48k wma stream, the 2G just can’t cope for anything more than dxing anoraks (or middle managers who want to hear their output!). Even 3G fell short of perfect. And this was at 6.30 this morning with presumably not much mobile traffic competing for the cells in the middle of my field.

Sure.. you can argue that 2G isn’t really meant for streaming. But the 3G just isn’t there outside the cities, and the roll-out of sufficently strong 3G to provide universal, or even half-decent, coverage would make these costs hugely higher.

DAB has already seen a big concentration of service in urban areas. This is another format which won’t fly very easily away from the cities.

[...] innlegget “Radio broadcasting over 3G – impossibly costly?” av James Cridland for [...]

It is very strange that IPTV does not have this problem, maybe theres something wrong with the IP Radio?

Bandwidth requirements will kill __________________

How many times have we heard this sentence, and how many times has it actually become true? Is there ANY technology that has been put out there, adopted heavily, and then retracted because of bandwidth requirements? Ever?

Andrew, listening to the radio via 3G is certainly not ‘adopted heavily’; and this research would tend to add considerable doubts about whether it would ever be so.

Are there any technologies which have not made it to heavy adoption because of their bandwidth requirements? Absolutely…

The cost and performance problems of “broadcasting” over internet protocols may be mitigated somewhat by the fact that it is not exactly analogous to broadcasting from an antenna. Internet broadcasts are not streamed so much as buffered (although that buffering window may get very small), and need not be exactly real-time. If a broadcast can be considered as a delivery of a bunch of small files, but delivered only to those who need them and only when little delivery windows are available, and they don’t have to be delivered in exactly real time but only need to play without audible gaps, then buffering efficiencies emerge. If it turns out those same small files are used by many many listeners at approximately (but not exactly) the same time, then network caching further localizes the delivery problems–the internet has gotten really really good at the problem of efficiently delivering a whole lot of small packets.

Costs may indeed be too high (with our service I’ve calculated they won’t be) but the comparison against traditional real-time broadcasting must take into account the different efficiencies of each delivery mechanism.

[...] while back, I linked to a French report that said that listening over mobile phones could be an impossibly costly broadcast medium for [...]

Mr Santa Maria,

As always I’m loving your work. You’re at the bleeding edge, fine sir.

Many thanks for the inspiration that you give to us all.

I appreciate your point of view and I think that it points to an uncommon answer to this question of whether 3G radio will take off. In my opinion there’s more to business than just “will it work better” or “will it improve the experience”. One question that I think we should really be asking is “is the effect big enough to warrant the change”. I personally feel that a change in the structure of radio as a service and experience on the whole will be what decides the fate of this issue. There must be more than just a cheaper or more convenient solution, this just isn’t enough to warrant this large of a change in the norm.