Why use “Error Correction Code”?
Sending digital data via wireless broadcasting and mobile reception without any transmission error is quite tricky. The car radio’s reception conditions are too unstable when driving. The received signal strength on the moving car antenna varies enormously. Even with a perfect error correction, one will never succeed in achieving 100% error-free mobile digital reception.
Error correction in all digital communication and, more specifically, in digital mobile communications is an absolute must.
It is precisely this phenomenon that will define the reception area for the DAB+ transmitter. As one moves towards the edge of the transmission area, bit errors will start to increase more and more. Once we reach the point where the error correction cannot adequately correct the bit errors, the signal becomes unusable, and reception ends. The range of a DAB+ transmitter is therefore not only determined by the power of the transmitter but also by the applied audio coding and the applied error correction.
Why does audio coding also play a role?
The applied error correction code in DAB(+) depends on the involved audio coding. For DAB with MP2 coding, “Unequal Error Protection (UEP)” is used.
For DAB+, “Equal Error Protection (EEP)” is used. Tests also show that the HE-AAC v2 encoding with EEP used in DAB+ performs better than the MP2 encoding with UEP used in DAB. We must conclude that the choice of the audio encoder and the error correction code used with it will play a role in planning the transmit power for a given transmission range. So it becomes a bit more complex than the FM issues we know.
DAB+ (HE-AAC audio coding with EEP error correction)
has better coverage than
DAB (MP2 audio coding and UEP error correction).
UEP versus EEP
UEP (used in DAB) is for data where certain bits are more important than others, which corresponds well to how MPEG-2 audio is encoded. One will use the “LDPC error correction” (Low-Density Parity-Check).
EEP (used in DAB+) is for data where all bits are equally important. It corresponds well to the way HE-AAC v2 is encoded. EEP uses “RS error Correction Code” (RS Code), named after the people who developed this code, namely Irving Stoy Reed and Gustave Solomon.
We find RS error correction also on a CD. That code ensures that a scratch (error) on the CD record sometimes does not affect the sound quality. The error correction happens while you listen to the music. So the whole process of error correction has to be very fast. It is also why they do not use software algorithms but much faster hardware (gates and shift registers) to achieve high-speed encoding and decoding for audio and video streams.
So DAB+ is indeed an evolution of DAB. Apart from using a different audio codec, a different and better error correction for the DAB+ system.
Both ways of error correction are purely mathematical matters. If you want to go deeper into this, we recommend an Internet search (BCH -Bose-Chaudhuri-Hocquenghem- codes, finite fields, Langlois fields, and syndrome calculation).
Important in error-correction codes is that audio bits and error code bits are interlaced.
Delineated blocks of consecutive bits of a certain length are processed. By placing them in a table. A table with columns and rows. Certain positions in the table are to be kept free for error-correction bits. One can fix more errors in the table according to the number of table positions for error-correction bits. Compare it to a crossword puzzle where you sometimes find the word by using the letters of other words. By testing the parity of the columns and rows, you can detect and fix errors in the table.
Spreading the data in time can improve the error correction. DAB(+) interleaves the data in time and frequency.
By spreading the bits over 4 consecutive broadcast frames, we lower the risk of unrecoverable data. We call this Time interleaving or concatenated coding. A DAB broadcast frame lasts 96ms and contains 4 audio frames of 24m (including their error correction). The audio will be delayed (4 x 96ms = 384 ms) at the receiver end, which is an acceptable delay for radio broadcasting.
After time-interleaving, one goes on to distribute the bits once more over the 1.536 carriers according to a particular pattern. It is called frequency interleaving. This method of time-interleaving and frequency-interleaving, together with the RS-error correction, makes the signal very robust against interference.
The greater the duration of time-interleaving, the more robust the signal becomes against bit errors in reception. Of course, one cannot use too long delays for radio broadcasting. However, the method was adopted by NASA, albeit with much more significant delays, for “deep space” data communication (relaying photos). It shows how technologically advanced the DAB+ system is.
The error code and time interleaving is added in the DAB(+) transmitter before the OFDM modulator is reached.
The error correction type is specified in the Ensemble MUX. It is communicated to the transmitter by data registered in the 24ms ETI frame.