| Professor criticises audio quality on DAB15th February 2007 Sverre Holm, a professor of signal processing at the Centre for Imaging, Department of Informatics, University of Oslo, has written a report criticising the audio quality provided on DAB stations in Norway. Professor Holm's conclusions are stark: the audio quality on DAB simply is not good enough because the audio quality is worse than on FM. To provide some background to the situation in Norway: Norway has the unenviable claim of being the last country to choose to adopt DABv1. And with DAB+ now being ready, all countries other than the UK, Denmark and Norway are expected to adopt DAB+, so it is likely that Norway will keep this unenviable claim forever. The driving force behind DABv1 in Norway is the national public service radio broadcaster, NRK, which has, like other broadcasters under fire for using the outdated DAB system, been, how shall I put it, somewhat economical with the truth. And in the English translation of the Summary, Professor Holm explains that: "This analysis of the audio quality of DAB has been made independently of the broadcasting companies and aims at balancing their information", because NRK has been incorrectly claiming that the audio quality on DABv1 is better than FM. The most bizarre thing about Norway adopting DAB is that sales had not taken off, yet they decided to stick with the old version of DAB even though they will have been aware that DAB was about to be upgraded. And in 2006 there were only 50,000 DAB sales in Norway, which included 15,000 DAB radios given away as gifts by a bank. Professor Holm's analysis clearly shows that NRK's claims about the audio quality on DAB are incorrect, in particular because 12 out of the 15 stereo stations on DAB are using 128 kbps, which he says provides a lower audio quality than on FM -- based on what I can hear on DAB and FM, 160 kbps and 192 kbps MP2 also sound worse than FM. Professor Holm's analysis clearly shows that NRK's claims are incorrect, in particular because 12 out of the 15 stereo stations on DAB are using 128 kbps, which he says provides a lower audio quality than on FM -- based on what I can hear on DAB and FM, 160 kbps and 192 kbps MP2 also sound worse than FM. The following figure shows the English translation of the Summary:
Just to explain a couple of things that people might not understand: When Professor Holm describes the treble cut-off being as low as 14 kHz, this is a result of NRK's use of 128 kbps, because the MP2 audio codec used on DAB is supposed to be used at bit rates between 192 kbps and 256 kbps for stereo stations. A common strategy used in low bit rate audio coding is to reduce the bandwidth as the bit rate is reduced. For example, on most MP3 encoders, if you reduce the bit rates the audio bandwidth is automaticaly reduced by the encoder. Another related audio encoding strategy is that once the available bits have run out to simply not transmit the highest frequencies, and with a bit rate of 128 kbps with the inefficient MP2 codec, the bits will run out for the vast majority of the time. And Professor Holm's comment about younger people finding the low treble cut-off more annoying refers to the fact that the bandwidth of human hearing reduces the older you get, so younger people will be able to hear higher frequencies that older people cannot. On the subject of joint stereo coding: DABv1 uses the MP2 audio codec, which is limited to using the "intensity stereo" method of joint stereo coding. Intensity stereo coding adds the amplitudes of the left (L) and right (R) channel signals together and calculates the ratio of the energy of the L and R signals at different frequency bands. Unfortunately, this destroys the signal 'envelopes' of the L and R signals, and because the ear uses the relative time at which sound waves reach each ear to determine which direction a sound is coming from, intensity stereo produces a very poor and unstable stereo image that frequently collapses altogether. All the other audio codecs, such as MP3, AAC/AAC+, WMA etc all use "mid/side" joint stereo coding, which is lossless in that information is not destroyed in the joint stereo coding/decoding process. Mid/side works by adding the L and R channel signals and subtracting them as follows and transmitting the mid (sum) and side (difference): At the encoder: M = L + R S = L - R And the decoder recovers the L and R signals as follows: Left = 0.5 x (M + S) = 0.5 x (L + L + R - R) = 0.5 x 2L = L Right = 0.5 x (M - S) = 0.5 x (L - L + R - (-R)) = 0.5 x 2R = R So the original L and R signals are recovered without errors due to the joint stereo coding process. Mid/side joint stereo coding is also the method that FM uses, where the mono signal is the sum of the L and R channels, and the difference signal is transmitted separately. Most of the codecs other than MP2 do have the option of using intensity stereo, but it's just that encoders choose not to use it because it doesn't work very well. And here's a quote that sums up the situation on other encoders, but MP3 in particular: "The main point to appreciate about Intensity Stereo is that it was only ever intended to be used in situations where - for whatever reason - the bitrate needs to be restricted to 96k or less, but where some form of stereo signal is desired. In these circumstances, it is generally preferable to accept some loss of stereo definition than to attempt to encode in full stereo with insufficient bits. Most of the music which circulates around the Internet these days tends to be encoded at bitrates of 128k or higher. Above 96k, all the current generation of encoders will default to Mid/Side Joint Stereo, and hence, most music fans will rarely encounter Intensity Stereo."
| |||||||||||||||
Digg
del.icio.us
Reddit
Facebook
Google
Stumbleupon
Slashdot