* Posts by Steve Green

28 publicly visible posts • joined 9 Nov 2007

BBC chief acknowledges DAB flop & internet radio

Steve Green

Pure Digital nobbles its Wi-Fi radios

@JeffyPooh,

"Stand-alone Internet Radio appliances ($100-$500) are slow to boot, ponderous to program a station, and generally a nuisance."

Nope. Pure Digital Wi-Fi Internet radios are (deliberately) slow to boot, (deliberately) ponderous to program a station and generally a nuisance, but that's because Pure Digital are as much opposed to Internet radio as the BBC and the rest of the DAB industry are.

If you buy a Roberts Wi-Fi radio though, it tunes into an Internet radio stream very quickly (whereas a Pure Digital Wi-Fi radio takes about 15 - 20 seconds before audio starts playing - Pure's marketing director Colin Crawford told me this long delay was to "make it easier" for listeners - bollocks, it's to put people off listening via the Internet, because their BBC masters don't want people to do that), navigation is simple, and they're overall good to use.

Pure Digital Avanti Flow internet radio and iPod speaker

Steve Green

Internet radio vs DAB quality

After writing 2 paragraphs describing the audio quality on DAB, how come Internet radio only received this measly, rather sneering sentence?:

"Bit rates can be an issue, but the Avanti presented internet radio in perhaps the best quality we've yet heard from this type of unit."

I take it you're not aware that around 95 UK commercial radio stations (i.e. ones on DAB and/or FM), including most of the biggest commercial stations, are using 128 kbps WMA or MP3 for their Internet radio streams, and the BBC has recently launched 96 kbps WMA streams for its national stations that are specifically for Wi-Fi connected devices? Both 96 WMA and 128 kbps WMA or MP3 provide significantly (96k WMA) or far (128k) higher audio quality than 128 kbps DAB due to DAB's use of the 1980s-designed MP2 audio codec? There's also over 4,000 Internet radio stations on Shoutcast that use bit rates of 128 kbps or above with the MP3 or AAC+ audio codecs.

Also, in a previous review of a DAB radio that could also playback MP3 files via SD card, you claimed that DAB sounded better than MP3. Considering that MP3 is a far better audio codec than MP2, could you explain how DAB could sound better than MP3? If DAB really did sound better than MP3 files, your MP3 files must sound absolutely dire, in which case, I suggest you start encoding your MP3 files using the Lame MP3 encoder rather than whichever 1990s-vintage MP3 encoder you seem to be using at the moment, or use a bit rate higher than about 80 kbps, which is the MP3 bit rate that provides approx equivalent audio quality to 128 kbps MP2 used for DAB stations - i.e. at the bit rates that people typically use MP3 at (128 to 192 is typical, but a lot of people use higher than that today) to encode their own files, MP3 absolutely murders DAB in terms of audio quality.

Steve Green

@Damien Cahill

"The Logik IR100 which I have, works great, but after 15 feet it loses the will to chat to my broadband router. Any review for Roberts 201/202 or the Pure Evoke Flow fail to say whats how far removed from the router will the Wi-Fi reception work."

I can vouch for the Pure Evoke Flow and Roberts WM201 and WM202 all working about 20 feet or so from my wireless router, and the signal has to travel through at least 2 walls to get to the Wi-Fi radio.

The Roberts WM201 is probably going to be the best in terms of range, because it uses an external Wi-Fi antenna, whereas the others, including the Pure Avanti no doubt, use an internal antenna.

In the ditch with DAB radio

Steve Green

@Luther Blissett

Don't worry about the 2017 date - it's completely unattainable, because they've only set a target for all cars to come factory fitted with DAB by 2014. There are 30 million cars, and car sales are around 2.3m per year, so they'd only have 3 years of full car sales with DAB factory fitted - that's only 6.9 million cars out of 30 million.

Basically, the 2017 figure was simply a fabrication because they didn't want to admit the truth that it's giong to be around 2022 - 2024, because that's just too far for anybody to care about.

Steve Green

DAB+ will replace DAB

"the UK is unlikely to make such a switch [ to DAB+] because the majority of the 7 million DAB receivers sold to date are not upgradeable."

Ofcom estimates there are around 120m - 150m FM devices in-use. So it's common sense that DAB+ is bound to replace DAB over time as the number of DAB+ receivers first outnumers, then vastly outnumbers the DAB-only devices. It'll be a gradual process, but the end result is inevitable.

For a stereo station using 128 kbps MP2 on DAB, they could reduce that station to mono at 64 kbps and broadcast a DAB+ version of the same station at higher quality than the MP2 version without requiring any additional capacity at all, so the transition isn't even difficult.

"BBC’s national DAB multiplex presently costs £6 million per annum for a network of 96 transmitters that cover 86 per cent of the UK population. To increase coverage to 99 per cent would require 1,000 transmitters, increasing the cost significantly to £40 million per annum."

I've been provided with info under the Freedom of Information Act that the £40m estimate is for 95% population coverage, so it's going to be far higher than £40m per annum to achieve FM's 99% coverage.

A rule of thumb is that the transmission costs are roughly proportional to the number of transmitters. So using the above figures, that would work out to be about £60m per annum.

In other words, we're going to have to pay through the nose for the privilege of having the choice of stations and the audio quality of those stations being severely limited just so that the BBC and the commercial radio groups can stop us listening to anything they don't produce.

The most outrageous thing about the DRWG report, IMO, is the fact that the BBC obviously has a power of veto to stop these blatantly anti-consumer recommendations being made, but the BBC went along with them.

Then again, this didn't come as any surprise, because the BBC has acted disgracefully on DAB since day 1.

Steve Green

Re: I had a DAB in my car

"The sound quality is actually a big improvement over FM. I know there are loads of hi-fi prats that go on about how awful it is but they're idiots who spend a grand on a cd player because "it's got a lower bit error rate""

DAB's sound quality is miles worse than on FM. I don't know which DAB car stereo you had, but DAB receiver modules that implement FM are absolutely diabolical on FM (they're made by DAB companies, so there's a slight vested interest in making them sound terrible on FM...), so I wouldn't go off what sound is coming out of a DAB car stereo - try comparing the sound quality with FM on a normal car stereo instead.

"Just go DAB+ or come up with some sort of standard that is low bandwidth enough to piggyback on the FM signal."

Correct re DAB+. But all the systems that piggyback on the FM signal are a pile of doggydoo - in fact, they're even worse in terms of allowing protectionist practices than the DRWG report wants to happen.

New alliance tunes into the 'net

Steve Green

The BBC hates Internet radio

The BBC would much prefer it if Internet radio didn't exist at all - because it views Internet radio as a threat to its beloved DAB system.

The BBC is only getting involved in this so that it can influence Internet radio - i.e. hold it back if it gets going.

The BBC's "management" of its Internet radio streams over the last 5 years for instance has been grossly incompetent. They were using Real Player G2 codec with a bit rate of 32 kbps until last year, and they 'transcoded' the audio rather than spending £5k per annum on a leased line to avoid transcoding - at a time when they were spending £6m (now about £9m) per annum transmitting DAB.

Overall, the audio quality was unlistenably bad, which was just how the BBC wanted it, because it stopped pesky consumers from deserting their beloved DAB platform.

Fixing the UK's DAB disaster

Steve Green

@Julian Regarding the year-on-year sales graph

Re the number of FM devices: the 120m - 150m figure I quoted was Ofcom's estimate of the number of FM devices that are *in-use*.

Re growth: the trend of that graph is downwards all the way from Jan 2006 when the BBC stopped showing its TV ads, so if DAB didn't receive a load more TV adverts the growth would fall to zero and then it would turn negative. Negative growth means annual sales would fall each year.

Steve Green

@Steve Todd

"Want to dial down the BS? See http://www.digitalradiotech.co.uk/articles/Annual-DAB-sales-50-below-forecast.php on your own web site and dated October 2007. 6.5 million was the number of sets sold as of the start of 2007."

Look at the top of the graph you're referring to, it says "by year end". So you're claiming that they were the sales at the start of the year, when they were actually the sales at the end of the year.

"You've proved that you can reduce transmission charges, not that the current charges are uneconomic."

Don't take it from me, take it from GCap:"The Board of GCap Media believes that DAB, with its current cost structure and infrastructure, is not an economically viable growth platform for GCap Media."

DAB's current cost structures are that it costs over £1m per annum to transmit a 128 kbps station nationally on DAB. That's why theJazz, Capital Life and BFBS ALL closed down on the Digital One multiplex two days ago, and that's why Planet Rock has just got one month to find a buyer or it will also be closed down.

"The public are currently buying about 2.5 million DAB sets per year. More than 50% of all radio sets on retailer shelves currently have DAB. Sales per month are still increasing. That doesn't much look like the public resisting to me."

DAB sales were 2.07m last year, not 2.5m. And there's a graph of DAB's sales growth on the first page of this article - the problem is DAB isn't growing quickly enough - 18% year-on-year growth last year overall, which is diabolical when there's 120m+ receivers that need to be replaced. DAB is never going to get to the finish line, basically.

Steve Green

@Steve Todd

"Mr. Green's other reality distortion number of 6.5 million sets. Elsewhere, on his own website, you find that more than 10 million sets have been sold to date)."

DAB sales were 6.5m the last time I saw any figures on it, which was post-Christmas. You definitely will not find me saying 10m DAB sales anywhere on my website.

"Transmission charges are another area where the numbers seem dubious."

A multiplex costs £X to transmit per year, so if 3x times as many stations transmitted on the multiplex the transmission costs per station would be x/3. Not that I'm suggesting that they should cram 3x as many stations on though, because they should improve the audio quality first.

"As the adds are padding they cost the Beeb next to nothing"

That's not the point - the point is that DAB has failed to sell well despite having masses of TV advertising lavished on it, and if it were a commercial format the plug would have been pulled by now because you can't just keep throwing money at something when the public are resisting buying it.

"Lack of stations? There are currently over 300 stations broadcasting on a mix of national, regional and local DAB multiplexes."

People can only listen to the radio in one place, so how many transmit across the entire country is irrelevant.

Steve Green

Re: Four times?

See page 48:

"Increasing coverage further to levels similar to those of FM radio may cost the BBC up to £40m per annum, as the number of transmitters would need to be increased to approximately 1000."

Then add £3.6m for the BBC's local DAB stations - total = £43.6m. Then look at the table, which shows that FM costs £12.3m per annum, and that's for all of the BBC's FM stations.

And as for the other thread, I read that, and as I recall it most comments were against DAB.

Steve Green

Regarding the year-on-year sales graph

I've just received an email about my use of the year-on-year sales graph, and people are calling me "a liar" for using it, so here's the reply I've just sent to the person by email, which explains why it DOES show a failing format:

The text in the article that's relevant to the graph is this bit:

"One thing is certain: the UK can't carry on the way it's been going.

Digital radio's year-on-year sales growth fell off a cliff the moment the

BBC stopped its "DABaganda" TV ad campaign in the run-up to Christmas 2005,

and sales growth has continued to slide ever since."

which says "Digital radio's year-on-year sales growth fell off a cliff".

The relevance of the year-on-year sales growth graph is that there are

120m - 150m devices that contain FM, and there have only been 6.5m DAB

receivers sold so far. Therefore, unless there is a reasonably high level of

sales growth, DAB will NEVER reach the target number of receivers sold. For

example, if sales growth falls to zero, as the trend of that graph suggests

will happen - unless the BBC kindly donates a loads more air-time for DAB

adverts - the annual sales wouldn't increase year-on-year, and annual sales

currentlys stands at 2m, so it would be 2m forever more if there was no

growth, so it would take 57 years for DAB sales to reach 120m.

Now, I don't know about you, but I happen to think that 57 years is too long

to wait for DAB to replace FM."

How the BBC plans to save your ISP

Steve Green

Spelling

"There is only one sodding letter O in the word lose. Its not a hard word to spell. Look, Its lose, as in lost and loss. Loose is an entirely different word! I can't understand why so many people seem to struggle with it."

There is a sodding apostrophe in "it's". It's not a hard apostrophe to put in. Look, it's it's, as in it is and innit. Its is an entirely different word! I can't understand why so many people seem to struggle with it.

Roberts WM-201 Wi-Fi internet radio

Steve Green

Listen Again

"1) Does it do full "Aunties 7 Day listen again" - or just the live streaming of whats currently playing?"

Yes, it does the BBC's Listen Again.

DAB: A very British failure

Steve Green

Yes, you are a luddite

@Mark W: You said: "As someone who used to work as a sound engineer in the recording industry, I wouldn't be considered a 'Luddite' as per Steve Green's post. In fact, the person with the narrow mind is indeed Mr Green. Look outside your little world and see the real people out there."

and you even have the audacity to go on to say:

"Unfortunately us audiophiles out there are the minority"

Here's a dictionary definition of "audiophile":

http://www.merriam-webster.com/cgi-bin/dictionary?book=Dictionary&va=audiophile

"a person who is enthusiastic about high-fidelity sound reproduction"

And yet you are accusing me of being narrow-minded because I think teh audio quality on DAB is crap? Do me a bleeding favour.

The analogy with the luddites is perfectly valid. There are people who want DAB to continue just as it is, and you seem very much to be of that persuasion

The problem with DAB is that it is using an audio codec and error correction coding that both date back to the 1980s - the MP2 audio codec is very similar to the MUSICAM codec, which dates back to the early 1980s!

MP2 was DESIGNED to be used at bit rate levels of 192 kbps or higher, and yet *98%* of all stereo stations on DAB are using it at 128 kbps. Anyone who knows the first thing about compressed audio knows perfectly well that if you use an audio codec at below its "sweet spot", i.e. below where it is optimised to be used at, the quality descends rapidly. If you don't believe me, here's what Karlheinz Brandenburg (the co-inventor of MP3) has to say about it:

"Lower bit-rates will lead to

higher compression factors, but lower quality of the compressed

audio. Higher bit-rates lead to a lower probability

of signals with any audible artifacts. However, different

encoding algorithms do have ”sweet spots” where

they work best. At bit-rates much larger than this target

bit-rate the audio quality improves only very slowly

with bit-rate, atmuch lower bit-rates the quality decreases

very fast."

DAB stations are being transmitted at bit rates waaaaay below MP2's sweet spot.

Would you encode an MP3 at 64 kbps or 80 kbps? That's the equivalent quality that DAB stations are using. NOBODY would dream of using 64 kbps or 80 kbps to encode their MP3 files, so why do we have to put up with this crap quality on DAB?

Last year I looked on Kazaa just to see what bit rates people were typically using nowadays to compress their MP3, and about half were using 128 kbps and half were using 192 kbps. 128 kbps MP3 is in a different league to 128 kbps MP2, and 192 kbps MP3 is on a different planet to 128 kbps MP2.

Accepting DAB is simply accepting mediocrity in the extreme, and anyone who does so deserves to be called a luddite. End of story.

There were 5,770 Internet radio streams on shoutcast.com using bit rate levels of 128 kbps or higher with the MP3 or AAC+ audio codecs the last time I looked. The vast majority of these will be providing far higher quality than DAB provides, so please justify why small-scale Internet streams are providing far higher quality than is being provided on DAB.

The BBC has been trialing the use of multicast for years now, and it's apparently going to be launched this year. That will allow all the biggest UK radio stations to use bit rates of 128 kbps - 192 kbps with modern audio codecs, and it's scalable to as many listeners as you like, because that's how multicast works. And the bit rates of multicast and of Internet streams in general will simply continue to go up and up because bandwidth is getting cheaper and cheaper. DAB cannot compete with Internet radio, so anyone who's trying to justify DAB when Internet radio is far better in every way (apart from some mobile reception issues - although have you seen how low the price of mobile broadband is now? and it's only going to get cheaper), is a luddite. Sorry, but there you go.

Steve Green

I just love uneducated posts

@Mark You said: "LOL, I just love uneducated posts like yours" and "and I will challenge you to tell the difference bewteen a CD and a high bitrate MP3 in a blind listening test. There is nothing wrong with DAB audio quality, a real music lover would be glad to hear the music, and not worry that it's in 128k MP2 format, which falls into the adaquete category."

You mention "high bit rate MP3", but you then go on to mention 128 kbps MP2 in almost the same breath, when MP3 is a vastly superior (yes, it is) codec to MP2, AND MP3 was designed to be used at 128 kbps, whereas MP2 was designed to be used at bit rate levels of 192 kbps or higher.

Your view is as uneducated as it is possible to be on the subject of audio codecs.

If 128 kbps MP2 were A-okay, as you say, why does the BBC use 256 kbps MP2 for the audio on its BBC 1, 2, 3, 4 TV channels? DAB is a radio system, isn't it, so why is TV providing far higher quality than radio?

Face the facts: DAB was designed in the late 1980s, it was incompetent to adopt the DAB system in the UK, and anyone sticking up for it is a luddite, because (a) DAB+ should be replacing it, and (b) Internet radio already murders it in terms of audio quality, number of stations and rock-solid reception quality, and in years to come it will crucify DAB and DAB+.

Steve Green

DAB is the luddites' choice

Some facts for the luddites to ponder regarding DAB:

DAB was designed in the late 1980s, and the DAB system we're using in the UK is still using exactly the same technologies that were chosen for it in the late 1980s.

The audio codec used is MP2 (MPEG Layer 2), which is meant to be used at 192 kbps or higher but 98% of stations in the UK on DAB use it at 128 kbps. For comparison purposes, the BBC uses MP2 at 256 kbps for its BBC 1, 2, 3, 4 TV channels, and yet it only sees fit to use 128 kbps for its music stations apart from Radio 3.

The (convolutional) error correction coding is extremely weak, which is why so many people suffer from "bubbling mud".

The AAC audio codec was standardised in 1997, which was 5 years before DAB was properly launched in 2002, and Reed-Solomon error correction coding was invented in 1960 and is used as the error correctino coding on CDs. If these two technologies had been adopted as part of an upgrade of the system prior to DAB being properly launched in 2002 we wouldn't have a problem with sound quality on DAB and those that suffer from poor reception quality would receive a far more robust signal for the same transmitter power levels that are used at the moment.

DAB+ has now adopted AAC+, which is a slightly more efficient version of AAC, and it has adopted Reed-Solomon coding, so it does now solve DAB's technological problems.

The reason why DAB+ was designed was that country after country was refusing to use DAB, and if DAB wasn't upgraded then the UK, Denmark and Norway would become stranded as being the only countries using the old DAB system, and everyone else would adopt one of the mobile TV systems to carry digital radio instead, like France just has.

Just to show exactly how incompetent a decision it was to adopt DAB without upgrading it first, consider that AM radio has been around since the 1920s and FM was first broadcast in the 1940s, and they're still both being used today, and yet just THREE YEARS after DAB was properly launched in the UK WorldDAB decided that it had to design DAB+ in October 2005. How ridiculous can you get? And FM provides higher audio quality than DAB!!

And as the excellent article comments, Ofcom only applies it "light-touch" regulation remit when it suits its own agenda to do so, and in the example of DAB+ it's stopped Channel 4 from using DAB+ for some unknown reason. Oh, hold on, Ofcom's Director of Radio, Peter Davies, was part of the "strategy" team at the BBC that made the fateful decision to use the 1980s version of DAB, so our hallowed regulator seems to have a massive personal vested interest in this issue.

As the article also suggested, the future of digital radio is on the Internet, and DAB will become the choice of the luddites and technophobes ONLY. Well, perhaps a few other people without any sense may also use them.

GCap Media, the UK's largest commercial radio group, has been running 128 kbps WMA streams for all of its stations since January last year, and there are around 5,700 streams on shoutcast.com alone that are using bit rate levels of 128 kbps or above with the MP3 or AAC+ audio codecs, which are far superior to MP2, as is WMA, and hence all of these statinos provide far higher quality than DAB provides.

The BBC has been trialing multicast for live streams for a number of years, and it will apparently be launching it this year. The bit rate levels are also between 128 kbps and 192 kbps, with modern audio codecs, so again teh audio quality will be far higher than DAB provides.

By the time of the London Olympics, apparently we'll see a dozen or more HDTV streams of different sporting events delivered via multicast, and as HD uses bit rate of 10 Mbps+, the bit rates of radio stations using multicast can obviously be far higher than teh current 128k to 192k.

As I say, DAB will be the preserve of luddites and technophobes, because why would anyone want to listen at lower quality to a narrower range of stations (all DAB stations have Internet streams) and possibly with dodgy reception quality?

The problem with Internet radio up to now is that the BBC has been deliberately limiting the audio quality of its Internet streams, and given its massive bias towards DAB, I would say that has obviously been to help the failing DAB system. The BBC has just launched the iPlayer TV streams which are usign bit rates of 550 kbps serving up to 660,000 people per day, and yet until around September last year the BBC was only using bit rate levels of 32 kbps with the dire Real G2 audio codec for its Internet radio streams. The BBC also butchers the audio for its Internet radio streams by receiving its own stations off-air via satellite at the location where the Internet radio servers are (even though it has fat pipes to Maidenhead where the servers are located) where the audio is re-compressed to Real G2. This is terrible audio engineering practice, and combine that with the ultra-low bit rates until late last year and the use of a dire audio codec, it's no surprise that the BBC's Internet radio streams were completely unlistenable - and now the majority of the public is under the misconception that Internet radio provides low audio quality, so job done by the BBC.

However, the BBC can only go so long before it becomes so much of a laughing stock that it will be forced to provide good quality - multicast is an obvious example where it will be inevitable that the BBC will provide high quality Internet radio streams.

Beeb confirms iPlayer streaming dominance

Steve Green

Re: @Chris Williams

Chris,

I'll admit that you've been proved right and I've been wrong regarding how the public has used streaming vs downloading. In my defence though, I was assuming that the picture quality of the streaming version would be a lot worse than it actually is - the reason I thought it would be crap was due to the fact that the BBC has recently increased the bit rates of its live radio station and Listen Again streams to 64 kbps (using the ancient ATRAC3 codec) and they sound diabolical, so I wasn't expecting the BBC would lavish so much bandwidth on the TV streams, to be honest.

If the picture quality were the same on both the streaming and download versions and there were absolutely no problems with stream buffering then I think it's pretty obvious that everybody would use streaming because it avoids having to wait for the programme to download.

However, I still think that the download version is complementary to the streaming version, and I'm completely against your proposals that it should be scrapped. It does provide higher picture and audio quality than the streaming version, and it always should do because it costs the BBC far less in bandwidth terms - I'd bet that 95%+ of people don't realise that the quality is better on the downloading version. It also avoids the problem of buffering, and on a number of occasions when I've tried it the streaming version has suffered horrendously from buffering. I don't think it will be feasible to stream HD content for a few years yet, whereas it could be offered on the P2P version today if they wanted to. And the speed of downloading programmes should also improve over time as more people install the download app so the number of people sourcing programme files will increase.

For the reasons of better quality and the avoidance of buffering, I will always favour downloading over streaming, so I'm completely against your proposal that the download version should be scrapped.

I also think that your dislike of the download version was at least in part due to it being Windows-only, and you were therefore sticking up for the smallish minority of people using Linux and Mac, so I don't think it's fair for you to now say that you're against the download version because only a smallish minority of people are using it - if you're going to stick up for one minority then you should stick up for minorities full stop. The BBC is a public service broadcaster, so it should provide for what everybody wants, not just the majority - if you're a Linux user, say, you shouldn't need to be told this!

Interesting news about the BBC building a Content Delivery Network.

I have a couple of questions for this Anthony Rose bod: when are the multicast TV and radio station streams going to launch? And when is the BBC going to change from using the prehistoric ATRAC3 audio codec for the live and on-demand radio streams?

Cheers,

Steve

Steve Green

MPEG-2 vs MPEG-4

It can't be MPEG-2 because they're using a bit rate of 550 kbps altogether for the video and audio, and at those bit rates it would look worse than the worst YouTube video you've ever seen if they were using MPEG-2. They're definitely using highly efficient codecs for both video and audio, so I reckon it's very likely that they're using H.264 video and HE-AAC audio - both of these codecs were added to Flash last August, see: http://tinyurl.com/2l9n5q.

Steve Green

Multicast

Referring to:

http://support.bbc.co.uk/multicast/

"The trial is still running but as a closed user group (password to access available on request, at our discretion) whilst we prepare for the next phase - inclusion of the live, multicast, channels in iPlayer (working title)."

That's referring to the live multicast streams of the BBC TV channels and the national radio stations that are expected to launch later this year, and will be included under the umbrella of the iPlayer. See:

http://www.bbc.co.uk/multicast/

Steve Green

Re: @Chris Williams

Firstly, you've never mentioned (to my knowledge, and I read all articles about the iPlayer) that what motivates your dislike of the download version is the licensing costs issue - my impression has been that you jumped on the anti feeling towards it due to it being Microsoft and it not supporting Linux and the Mac. (for the record, I'd prefer it if the download version supported Linux and Mac, but it's no reason to stop Windows users from using it)

Secondly, have you actually looked into the issue of licensing costs? Licensing costs saved on the download version is extra licensing costs paid for the streaming version to Adobe for Flash and licensing costs for the use of whichever codecs the streaming version is using, probably MPEG-4 H.264 and HE-AAC.

Also, because Microsoft wants people to use WM it provides cheap licensing terms - it gives WMV encoders away to broadcasters, for example, so that it can keep people using WM, and there was a load of arguing about the licensing cost of H.264 a couple of years ago after Microsoft lowered the licensing costs to use WMV to very low levels - a sort of loss leader type affair. So I'm not convinced that the licensing costs would be any lower for the streaming version than for the download version. Also, Internet bandwidth costs for the BBC will obviously be far, far lower for the P2P version than for streaming, because users are providing the bandwidth rather than the BBC having to - and it ain't half lavishing bandwidth on this streaming version, because I'd imagine they're providing about 10 - 20 Gbps of bandwidth for it, which ain't free.

As for the use of the word "bias", I'm sorry but I think it's self-evident that you're biased against the P2P version - if you don't think you are, try re-reading some of your previous articles with an open mind, and then tell me that I'm wrong.

As for multicast, I repeat that it cannot be used for the streaming version of the iPlayer, because it is a live streaming technology - that';s why the BBC is trialing it at the moment with all the live streams for the TV channels and radio stations. Multicast only makes things more efficient because different people are watching the same stream. But people don't watch the same stream on the iPlayer - if someone starts watching Eastenders say 5 milliseconds later than someone else watches it the two streams are completely different, so unicast *has to* be used. Ironically, the way to make live streams more efficient is to use P2P!

Steve Green

MPEG-4

Does anyone actually know for sure that MPEG-4 isn't being used on the streaming version of the iPlayer? Flash added MPEG H.264 for video and HE-AAC for audio just before the streaming version of the iPlayer launched, and I've read elsewhere that the streaming version was going to use H.264.

Steve Green

@Chris Williams

I'm with Vlpes Vulpes on this. For some reason you seem to have an attitude problem. Basically in each and every article you've written about the iPlayer so far (maybe with the exception of this one) you've asked for the download version to actually be pulled. WTF would be the point in doing that? You seem to act like some spoilt kid who hasn't got a football, so he goes over and robs someone else's football so they can't play with it - it's pathetic.

The download version of the iPlayer provides higher picture and audio quality than the streaming version, doesn't harm any Linux or Mac users, by defintion won't have any buffering problems, it benefits from scale rather than being hindered by it, and there'll be a long wait for streamed HD content whereas it's feasible today on the download version.

It is COMPLEMENTARY TO the streaming version of the iPlayer, it is not an either or situation, but for some fking reason you've got a chip on your shoulder and you actually want it to be taken away from people who want to use it rather than the streaming version.

And not content with wanting something useful to be taken away from people who use it, you come out with a load of nonsense to back up your bias. Yesterday you wrote that the spike in traffic to the iPlayer website was DUE TO the release of the streaming version, which shows you have zero understanding of the reason for advertising and why TV advertising campaigns cost millions of pounds. And in a reply to a previous article you claimed that multicast could be used for the streaming version of the iPlayer to make it more efficient in bandwidth terms, when in reality multicast is a LIVE streaming technology, so cannot be used with on-demand streams - that was just one of a number of inaccuracies.

If I were you I'd concentrate less on bias, and just get your facts right and stop trying to be a smart arse.

Beeb's iPlayer reaps streaming traffic dividends

Steve Green

@Chris Williams

"The BBC's Flash-based streaming service has gifted a massive [14x] traffic boost to the iPlayer site since it went live in mid-December, independent figures have revealed."

So you're seriously claiming that the streaming version being launched led to the traffic to the iPlayer website going up by 14x??

Why do you think TV advertising campaigns cost millions and millions of pounds? It's because TV advertising campaigns multiply sales - it is not an excuse for the advertisers to donate money to ITV et al. The spike in traffic to the iPlayer website will have been almost completely due to the TV advertising blitz over Xmas. If the streaming version had launched before Xmas and there had been no TV adverts the traffic to the iPlayer website would have hardly gone up at all.

Also, do you actually use the streaming version of the iPlayer? I've found that it buffers so badly that it's unusable at peak times. And there's nothing wrong with my Internet connection, because to check that it's okay when the iPlayer is unusable I've tried downloading from a fast server and it manages to download at about 6 to 7 Mbps.

"We've written at length on why the desktop client is a non-starter for the mass market. With the case for streaming now proven, the BBC would do well to silence the interoperablity grumbles for good by killing it off."

So you think it's a good idea to kill off the download version that provides higher picture and audio quality than the streaming version, works perfectly well, by definition has no problems with buffering, will scale well as more people use the iPlayer, and is likely to carry HD content in future? And in its place you want the BBC to only use a streaming version that suffers horrendously from buffering problems, scales abysmally due to it using unicasting (and cannot use multicasting), provides poor picture and audio quality and will probably be able to carry HD content by about 2046 once the BBC can afford the Internet bandwidth required? Good call.

Flash-based iPlayer is go

Steve Green

Re: audio

@Elisha: did the BBC forget to use these new audio codecs that Flash supports on the streaming version of the iPlayer then, because the audio quality is diabolical.

Steve Green

Re: Streaming vs downloading

I'll answer each point in turn:

I tried an iPlayer stream yesterday and it buffered a bit, and as I said, I tried it this morning and it buffered about 10 times in the first 3 minutes - and it then stopped saying there was some problem and to try again. But these might be initial teething troubles anyway.

Okay, I'll take it back that you were mud chucking for the sake of it, but I stil think you're downplaying the value of the P2P version of the iPlayer.

The actual number of people actively using the iPlayer isn't much of an issue at the moment, IMO. If in a year's time and only a few thousand people are using the download version then you'd have grounds to say that it's a waste of time, but I doubt that will be the case. You could just as well use the same argument as you did to be sceptical about Bittorrent, because no doubt there's some X to 1 ratio of how many have downloaded it and how many actually regularly use it.

On the "if you want quality" issue, yes, people should always use a PVR first, but I use the iPlayer for when I've forgotten to or wasn't able to record a programme.

Downloading on Bittorrent is, IME, very, very slow compared to downloading programmes on the iPlayer. On the iPlayer there will be a single file for each programme, whereas on Bittorrent there would be many different recompressed recordings of a programme. For example, 1 user starts recording at 8.29, another starts a few seconds later and so on, and they're different files altogether. On the iPlayer it is far more likely that there will be a lot more sources with the file you're looking for compared to on Bittorrent, hence the download speeds are far quicker. And this should improve dramatically over time as more people install the iPlayer, whereas lots of people have already installed Bittorrent clients.

Quality doesn't matter on YouTube because they're very short clips, but I don't think I'm alone in not wanting to watch TV programmes with cr&p picture and audio quality.

Multicast is a technology for the distribution of live streams, so it cannot be used for on-demand programmes on a catch-up service.

Fair point about Kontiki being dreadful, but I'm sure that a lot of the work they've done for the PC download version can be re-used for a set-top box version.

Steve Green

Streaming vs downloading

What Chris Williams's original article slagging off the download version of the iPlayer ignored, and what has proven to be the case now it's launched, is that the picture and audio quality of the streaming version are miles worse than on the download version, and there's the issue of the stream buffering which is obviously not an issue for downloaded files - I've just tried watching a few minutes of a streamed programme and it buffered about 10 times in the first 3 minutes.

If you look at if from the point of view of how much bandwidth the BBC needs to provide these services, the streaming version uses unicast distribution, so they've got to distribute the streams in parallel, whereas on the P2P download service the users are doing most of the distribution for them, so the quality will always be better on the P2P download service than it will be on streaming, and HD content should be provided via downloads long before it'll be available on the streaming version.

The original article also suggested that the P2P service was basically a complete waste of time, and that they'd have been better to bypass downloading to PCs altogether and go straight to a set-top box version, but I'd have thought that a set-top box version would just contain pretty much the same P2P software that's used for the PC version just put inside a set-top box, so the P2P version is hardly a waste of time.

Another thing with P2P is that the more users there are the quicker the download times should be, so that'll improve as more people sign up to it, and at off-peak times I've seen 30-minute programmes download in 5 minutes - and speeds should increase over time due to broadband connection speeds going up as well.

I'm all for slagging off the BBC for taking years to develop the iPlayer and spending a small country's GDP on it and not porting it to Mac or Linux, but some of the criticisms of the download version just seem like indiscritimate mud chucking for the sake of it.

Music DRM 'dead by next summer'

Steve Green

@Steve

You say that 320 kbps MP3 is good enough for the vast majority of people, but are the majority of music download services offering quality as high as 320 kbps MP3? DRM-protected AAC on iTunes is a measly 128 kbps, and the large majority of music download services don't use quality as high as 320 kbps MP3 from what I can gather.

Also, I think lossless will become more and more popular over time (will anybody need to use lossy audio compression when we've all got 4TB+ hard drives??), so I think you should look at offering lossless at least as an option for people in the near future.