Response to DAB article on The Register

After The Register published an article that was critical of DAB by yours truly, they then published another article in which Nick Piggott, who works on DAB at GCap Media, was allowed to defend DAB, so here I'm going to set the record straight about some of the things he said.

I could say a good few things about this, but as GCap launched 128 kbps WMA streams for its stations -- which are at higher audio quality than the same stations are at on DAB -- at the beginning of last year, I'll let this lie.

Sorry to have to break this to him, but DAB+ is about as beautiful as Amy Winehouse first thing in the morning, and it's about as elegant as a hippo riding a unicyle. And I say this as someone who proposed that DMB should be used long before DAB+ was designed (DAB+ is just DMB without video support), and I support the use of DAB+, so I'm hardly biased against it.

DAB+ is basically a botch job that has allowed the broadcasters to make use of their existing transmitter networks -- if you had a clean slate you would never choose to use anything like DAB's 1980s-based transmission scheme as a base on which the DAB+ elements are built on top of. The DAB parts of the system seriously limit how good a system DAB+ is.

In reality, development of AAC began in 1994, it was standardised in 1997, and the BBC undertook listening tests that compared AAC with MP2 in 1996 and 1998.

Both listening tests showed that AAC was twice as efficient as MP2. And here's what BBC R&D engineers were saying about AAC in a brochure for an R&D open day in 1999:

AAC was state-of-the-art in 1999, whereas MP2 is based on the MUSICAM audio codec, which was designed in the mid 1980s. MP2 was nowhere near state-of-the-art in 1999. They had the opportunity of designing a system in the late 1990s that would be very similar to what DAB+ is today, but they screwed up, it's as simple as that. The whole fiasco could have been avoided if they'd just have looked around at the technologies available to them, but they just took the blinkered approach, probably because they didn't have a clue about the other technologies available at the time - but the experts in the BBC R&D department did know, but they will simply have been ignored by the cluless BBC executives who thought they knew best.

It's true that bubbling mud should no longer be a problem, but bubbling mud is actually caused by the error correction coding on DAB being too weak rather than it being the fault of MP2. The audio quality problem is in addition and separate to the bubbling mud problem, because the bubbling mud is a reception quality issue, whereas the problem with the audio quality is due to the broadcasters deciding to use bit rate levels that are too low to provide good audio quality - i.e. the audio quality would still be dire even if you received a DAB signal that has perfect reception quality.

That's true in the very short term, but in reality it's pure spin. Assuming DAB does reach high penetration levels (i.e. it doesn't actually completely collapse in the meantime), the commercial broadcasters will launch a lot of new stations in future, because commercial radio saw high growth in the 1990s due to the expansion in the number of commercial stations, and they've always viewed DAB as a way of replicating that growth today. Also, Ofcom estimated that 90 out of the 326 or so analogue radio stations that are transmitting today won't be able to transmit on DAB (how on earth did they fail to notice this problem when they were doing the initial "planning"?) due to either the local DAB multiplex in their area being full or they just can't afford the transmission costs. They could afford them on DAB+ though, and there would be space to fit them on (much less space is needed to accommodate a DAB+ station due to the lower bit rates required), so over a quarter of all analogue stations need to be fitted on in addition to any new stations the commercial radio groups will launch. So they will definitely reduce their transmission costs per station by using DAB+, because the multiplexes will inevitably fill up, and it was simply disingenuous of Nick Piggott to suggest that they wouldn't save any money - they won't immediately, but they will save in the future.

This hasn't got anything to do with inventory, this is all about the DAB receiver manufacturers profiteering by selling non-DAB+ receivers at the expense of consumers, because the manufacturers want to sell people two radios instead of one.

There's a transcript of a presentation given by Colin Crawford, Pure Digital's Director of Marketing, at the bottom of this page. He tried to make out in that presentation that DAB+ is very expensive for them to adopt, and yet according to his own figures in the presentation, if you remove the audio codec licensing costs (and adjust for the overly high price he said for memory chip costs), it only costs the manufacturers 40p extra to make a DAB+-upgradeable radio - the audio codec licensing costs would be paid later when the owner chooses to pay for the software upgrade to enable DAB+.

Frontier-Silicon designed its new DAB+-upgradeable modules to be smaller than previous modules, and they produced reference designs to allow manufacturers to easily convert existing models to DAB+ -- the product manager of the Venice 5 DAB+-upgradeable module told me in October/November 2006 that he expected that the large majority of all DAB receivers on sale by the end of 2007 would support DAB+, because he expected the manufacturers to convert their existing models over to the new module. There's only about 10 or so models out now, albeit that there should be a wide range of DAB+ models out later this year. They've had loads of time, but they're simply delaying building them to profiteer.

If there were a wide range of DAB+ receivers on sale then people could continue buying "DAB" radios. And there's no excuse for dishonestly claiming that DAB+ will never be used in the UK, which is what the broadcasters have tried to do over the past year - including the BBC.

GWR and then GCap have continually promoted DAB since day one in their radio jingles by saying that it provides "superb digital quality sound". The DAB industry has been pushing everyone to buy DAB, and there has never been any promotion of the fact that radio via digital TV (and now online streams) provides higher quality than DAB, and in my opinion, based on emails I receive via my website, I think the vast majority of people are under the impression that DAB provides higher audio quality than the other digital platforms, which is due to the way the broadcasters have promoted it.

I calculated in 2004/5 that DVB-H transmitted in Band III (the frequency band that DAB uses) would be 12-13 times cheaper to transmit than DAB is. They're only ballpark figures, because it simply uses the assumption that transmission costs are linearly related to transmission power, which is sometimes used as a rule of thumb. The reasons why DVB-H would have been so much cheaper to carry digital radio (it does rely on full multiplexes to get the lower transmission costs per station, though, but you wouldn't design a system to run with half-empty multiplexes) are that AAC+ could be used, so that reduces the transmission costs by a factor of 3, and DVB-H uses far stronger error correction coding than both DAB and DAB+ use, and using stronger the error correction coding reduces the transmission powers required, in this case by a factor of 4, so the overall transmission costs per station works out to be the product of 3 x 4 = 12 times cheaper than they are on DAB.

There's also a DVB-H2 system coming out in the next year or two, which will use MIMO antenna array technology, which could potentially be 100-times cheaper to transmit than DAB is, because MIMO is extremely effective in terms of reducing the transmission powers required to provide robust reception quality, and DVB-H2 is likely to use the LDPC + BCH coding that looks like it will be used on the DVB-T2 system as well, and LDPC coding is one of the forms of error correction coding that perform within a fraction of a decibel of Shannon's limit, which is the theoretical limit beyond which no communication system can pass, so this would also provide a significant reduction in transmission powers as well.

What Nick Piggott failed to mention about DVB-H is that a couple of years ago there was a mobile TV system format war between DMB (a DAB-based standard) and DVB-H (a DVB-based standard), and of course the UK DAB industry was very much supporting DMB, so he's clearly still loathed to admit that DVB-H is the far better system of the two, and it would therefore have been a far better system to use for digital radio as well.

Oh yeah, I bet people will be queueing at the shop doors to get their hands on a Bluetooth headset with built-in DAB receiver -- Steve Jobs must be quaking in his boots that the nanoDAB will be so sought-after that the iPhone will simply stop selling altogether.

Bookmark with:
Slashdot	Digg	del.icio.us	Reddit