| BBC has deliberately delayed launching higher quality live Internet radio streams to help DAB14th June 2009 Dear James, I would like to complain about the BBC deliberately degrading the audio quality of the live Internet radio streams for the last few years due to the BBC being extremely biased towards DAB and extremely biased against Internet radio. The complaint is structured in two parts: the issue regarding the BBC deliberately delaying the launch of the higher quality live Internet streams for the past year, followed by examples of the way the BBC has deliberately degraded the audio quality of the Internet radio streams for the last five to six years. The BBC's live Internet radio streams were meant to improve in quality in June last year, immediately after the quality improvements had been implemented for the on-demand Internet radio streams. Yet an entire year has passed since then but the BBC is still delivering its live streams at low quality (using 64 kbps Real Player) to the vast majority of users. You repeatedly said on the BBC Internet blog and the BBC Radio Labs blog last year that the BBC wanted to deliver the live Internet radio streams at lower quality than the on-demand radio streams (relevant quotes from the BBC blogs are copied below). You attempted to justify this course of action by saying that live radio is also available via DAB, FM and the digital TV platforms. However, such a "justification" is being biased against the live Internet radio streams by definition, because it is either technically possible and economically feasible to deliver both the live and the on-demand streams at the same level of quality or it is not, and the obvious fact is that it is possible to deliver them at the same quality, therefore wanting to deliver the live streams at lower quality was blatantly being biased. The BBC later tried to distance itself from making these biased remarks, but due to the delay in launching the higher quality streams, the live streams have been delivered at lower quality than the on-demand streams anyway, so for the last year the BBC has done what it said it wanted to do in the first place. The BBC has deliberately delayed launching the higher quality live streams, and this is down to the BBC being extremely bias towards DAB and biased against Internet radio. The reason why the BBC is so biased against Internet radio, and biased against the live streams in particular, is because Internet radio is widely accepted to be the main long-term "competitor" to DAB in terms of live radio listening, and as the BBC's radio stations would face far more competition online than on DAB the BBC thinks that it would lose listeners if Internet radio became popular. The following quote from the "Myers Report" (a report for government about the regulation of local content on commercial radio) sums up the radio industry's view of Internet radio perfectly:
The BBC has therefore adopted the long-term strategy of trying to stop Internet radio becoming successful in order to minimise the number of listeners it would lose. In other words, the BBC has adopted a protectionist strategy which revolves around pushing everyone towards DAB and away from Internet radio, even though the BBC knows perfectly well that this strategy is blatantly against the interests of licence-fee payers. The BBC also knows perfectly well that it is supposed to be "platform neutral", i.e. it is not supposed to give preference towards one platform over another, but for many years now the BBC couldn't have been more biased towards DAB if it tried. There is further evidence to support the premise that the live streams have been delayed deliberately. For example, the following is a quote from a BBC blog about the Internet radio streams from July last year:
However, the BBC subsequently launched new higher quality 800 kbps H.264 iPlayer TV streams on the 12th August. These iPlayer TV streams were launched whilst the Olympics was in progress, which therefore disproved the claim that a BBC-wide code freeze was in place whilst the Olympics was on. I therefore suggest that the Olympics being on was simply used as an excuse to delay launching the higher quality live Internet radio streams. Another issue that strongly suggests that the BBC has deliberately delayed launching the higher quality live streams for as long as possible is the time it has taken the BBC to "test" the live streams. The new AAC/AAC+ live Internet radio streams finally began testing in February this year, yet it is now June and the streams still haven't launched. Performance testing of Internet streams consists of accumulating a sufficient amount of reception data from which statistically significant inferences can be drawn about the performance of the streams. Anyone with a basic knowledge of statistics knows that you can have a high degree of confidence (e.g. 95% or 99% confidence) in results that have been inferred from a surprisingly small amount of data, and the BBC would have reached such a level within the first week of these tests being carried out in February. Certainly the results would not have changed by any significant degree once 2 - 3 weeks of data had been acquired, therefore the last three and a half months of "testing" has served no purpose whatsoever. Similarly, listening tests to gauge the audio quality at different bit rate levels literally only take a few hours to complete, so this couldn't explain the four month "testing" period either. The BBC has therefore obviously chosen to string out the "testing" process in order to deliver the live Internet streams at low quality for as long as possible. Arguably the most damning indictment of all is the fact that if the BBC really wanted to improve the audio quality of the live Internet radio streams, it could very easily have done so by simply increasing the bit rate levels used for the existing Real Player live streams -- for example as a stopgap measure until the AAC/AAC+ streams were ready to launch. The fact that the BBC couldn't even be bothered to do something as simple and obvious as this at any point over the last twelve months speaks volumes about the BBC's attitude towards the audio quality of the live Internet radio streams. People spend around 17 million hours per month listening to the BBC's live Internet radio streams, yet the BBC has deliberately kept the audio quality of the Internet streams as low as possible for as long as possible. Furthermore, as I will discuss below, the BBC's actions over the past year are far from being an isolated incident, because they are in fact simply a continuation of the disgraceful way in which the BBC has deliberately mis-managed its Internet radio streams for the last five or six years. BBC's Internet radio streams were delivered at diabolical quality for several years to help DABThe BBC could have used AAC+ for the Internet radio streams from January 2004 onwardsRealNetworks added support for the AAC/AAC+ audio codec to Real Player in January 2004, and AAC+ is far and away the most efficient and therefore the best audio codec to be used at the extremely low bit rate levels (32 kbps) that the BBC was using for its Internet radio streams until mid 2007. Therefore, even without increasing the bit rate levels it was using, the BBC could have drastically improved the diabolical audio quality it was delivering its Internet radio streams at simply by switching from using the dire Real G2 codec to using AAC+ instead. The BBC has therefore been using completely the wrong audio codec for its live Internet radio streams for the last five and a half years, and the BBC is still using the wrong codec to this day. I find it impossible to believe that people working on digital radio at the BBC haven't been aware that they could have used AAC/AAC+ for the Real Player streams for the last five and a half years. For example, Virgin Radio launched 128 kbps AAC Real Player streams in April 2004, and I don't believe for a second that the BBC's digital radio team weren't aware of the existence of these streams. I therefore conclude that the BBC deliberately chose not to use AAC+ in order to deliver its Internet radio streams at very low audio quality as part of the BBC's strategy of wanting to deter people from listening to radio via the Internet. After all, Simon Nelson, who was the BBC Controller in charge of digital radio until 2007, said on Radio 4 Feedback that "of course the BBC would prefer it if everybody listened to digital radio via DAB". That hardly leaves a great deal of doubt about where the BBC's preferences lie. The BBC used a bit rate of 32 kbps for the Internet streams for Radios 1, 2 and 4 until 2007Here's a recording of Radio 1's Internet radio stream from 2006 (FLAC lossless version). It is no exaggeration to describe the audio quality of that recording as being absolutely diabolical. But that recording is perfectly representative of the level of audio quality the BBC delivered its Internet radio streams at over the period from 2003 up to mid 2007, because the BBC was using 32 kbps for the Internet streams for Radios 1, 2 and 4 (the other stations were using 44 kbps) over that period. In 2007 the BBC finally deigned to increase the bit rate levels to 64 kbps, but even after the bit rates had been increased the audio quality was very poor. To put these incredibly low bit rate levels into context, the BBC launched the iPlayer in December 2007, and the iPlayer TV streams started off using a bit rate level of 500 kbps, which is 15 times higher than the 32 kbps that the BBC was using for the Internet radio streams just a few months beforehand. The following year saw the bit rate levels of the iPlayer TV streams increase to 800 kbps (25 times higher than 32 kbps), and in April this year the BBC launched high-definition iPlayer TV streams at a bit rate of 3,200 kbps (100 times higher than 32 kbps), along with simultaneously increasing the bit rates of the existing 'normal' and 'high' quality standard-definition iPlayer TV streams from 500 to 800 kbps (25 times higher than 32 kbps) and from 800 to 1,500 kbps (47 times higher than 32 kbps) respectively. The following figure shows the bandwidth that the iPlayer TV streams have been consuming since their inception in December 2007, along with the combined bandwidth consumed by the BBC's live and on-demand radio streams.
Anthony Rose, the BBC Controller in charge of the iPlayer, has also made the following prediction about future bandwidth requirements for the iPlayer TV streams in an interview for the EBU Technical Review in August 2008:
The following graph shows the effect this would have on the amount of bandwidth required for the iPlayer TV streams assuming that the 10-fold increase takes two years (from August 2008) to happen (bandwidth obviously wouldn't actually increase in a straight line).
In case you missed it, the graph above also shows the bandwidth consumed by the Internet radio streams up to the present day -- it's the blue line near zero on the graph (if you're having difficulty seeing it there's a Magnifier available in Windows under Programs > Accessories > Accessibility). If what Anthony Rose predicted does indeed materialise, the iPlayer TV streams would be consuming 30,000 TB/month by August 2010. In comparison, the total bandwidth consumed by the BBC's Internet radio streams was just 317 TB/month in August 2007 (which was itself after the BBC had increased the bit rates of the Internet radio streams from 32 to 64 kbps). So the iPlayer TV streams would be consuming 95 times as much bandwidth as the radio streams were consuming just 3 years previously. Even allowing for the price of Internet bandwidth falling, the above figures categorically show that the BBC could easily have afforded to use far higher bit rate levels for the Internet radio streams than it did do over the last few years. In other words, the BBC chose to deliver its Internet radio streams at very low audio quality for several years even though it could have delivered them at far higher quality. If that isn't due to bias, I would be very interested to hear a plausible explanation for why the BBC considered it necessary to use 32 kbps until mid 2007, why the BBC only increased the bit rate levels to 64 kbps in 2007, and why the BBC is still using 64 kbps today. The BBC also butchered the audio for the Internet radio streams for six years just for good measureIf the disgraceful "engineering" decisions mentioned above weren't bad enough, the BBC also 'transcoded' the audio for the Internet radio streams for six years until last autumn. The BBC knew full well that transcoding is diabolical audio engineering practice because of the effect it has on audio quality, and it is perfectly reasonable to describe the practice as "audio butchery". Transcoding's effect on audio quality also becomes more severe the lower the distribution bit rate level is, and the BBC was distributing its streams at a bit rate of 32 kbps! The BBC transcoded the audio for the Internet streams in order to save the £5,000 - £10,000 per annum that was needed for a dedicated leased-line to be installed to transport the audio to the BBC's Internet servers in Maidenhead. In comparison, the BBC has been spending between £10 to £14 MILLION per annum on transmitting DAB over the same period, which is a figure that is likely to increase to between £40 to £70 million per annum once the BBC's national DAB multiplex is extended to cover the entire UK population. This was deliberate, gross mis-prioritisation of licence-fee payers' money, and the same thing can be said about most of the other issue I've raised in this complaint. The BBC delayed deploying new encoders for the Internet radio streams for 4 yearsThe following is a quote made by Alan Ogilvie, who works in the BBC Audio & Music's Distribution Technologies team, about the new "Coyopa" encoders, which were deployed last autumn to encode the audio for the new Internet radio streams:
So the BBC drew up the requirements for the new encoders, then they postponed deploying these new encoders for four years even though the BBC knew perfectly well that it was delivering the Internet radio streams at diabolically low audio quality. I do hope that the BBC put to good use the few hundred pounds it saved in interest payments by delaying the deployment of the new encoders for four years. Presumably the few seconds' worth of Eastenders it helped to pay for were all the better for it. I somehow don't think the people who spend 17 million hours per month listening to the live Internet radio streams -- which could equate to as much as 500 million hours over four years even taking into consideration the growth in Internet radio listening (and that doesn't even include listening to the on-demand streams!) -- would agree that the BBC saving literally a few hundred pounds per year was a worthwhile cost saving when the BBC spends £14 million per annum transmitting DAB. It is examples like this that completely rule out the possibility that the decisions were motivated by anything other than bias.
Conclusions
The decisions taken by the BBC with regards to the Internet radio streams over the past five or six years have shown a systematic and extreme level of bias towards DAB and against Internet radio. The numbers speak for themselves: we're literally talking about the BBC refusing to spend sums of money on Internet radio that are between a thousand to several thousand times less than it spends on transmitting DAB each year -- even though the BBC knew perfectly well that the audio quality of the Internet radio streams was absolutely diabolical when it took these decisions. A BBC executive has chosen to allocate financial resources to the different digital platforms in the proportions just mentioned, and I'm afraid that there is simply no possibility whatsoever that that person has not displayed an extreme level of bias towards DAB and against Internet radio. I also consider it to be self-evident from the facts given that the BBC has deliberately delayed launching the higher quality live Internet radio streams for the last year.
I look forward to hearing your response.
Quotes from the BBC Internet blog and BBC Radio Labs blog referred to aboveThe following are quotes listed in chronological order that are relevant to the discussion above written by James Cridland on the BBC Internet blog and the BBC Radio Labs blog last year. The first quote, from the 'Streaming Radio Online' blog, shows that the BBC wanted to deliver the live Internet radio streams at lower quality than the on-demand Internet radio streams:
As will become clear below, James Cridland used the word "differentiating" to mean that the BBC was looking to deliver the live Internet streams at lower quality than the on-demand streams. He then repeated this intention in the 'Under the iPlayer hood for radio' blog:
and he said the same thing in the 'Latest on Coyopa...' blog:
He also said the same thing in an email replying to myself:
| |||||||||||||||||||||||
Digg
del.icio.us
Reddit
Facebook
Google
Stumbleupon
Slashdot