Lindsay Cornell, Principal Systems Architect, BBC and Chairman, WorldDAB Technical Committee
The WorldDAB Technical Committee has completed work on revising the main character set used for station names and text accompanying radio programmes. The revision has been made to ensure that the spread of DAB into Eastern European countries is not hampered by poor rendering of text on radios. Two tasks made up the work:
Regarding the first task, the WorldDAB Country Update was consulted to see which countries were using or about to use DAB. This provided a priority list of languages that had to be included. Next, a second tier of languages were added, which comprised all the Latin-based languages used in Europe. It was found that additional characters were needed to provide the complete alphabets used to write in Czech, French, Hungarian, Latvian, Lithuanian, Maltese, Polish, Romanian, and Slovak.
The original character set was the result of work by the EBU in the late 1970s to support the introduction of teletext. The character set was then adopted for RDS and DAB as they were specified in the 1980s and 1990s. It is an 8-bit character set, which means that there are 256 (two to the power of eight) code points available to represent characters. Some code points had no associated character and so were available to be defined, but there were not enough free code points to accommodate all the desired new characters. An analysis was made of characters that would be very unlikely to be used in station names or programme text and that could therefore be replaced with new characters without causing difficulties for existing broadcasters.
The original EBU character set contained 222 characters. The new character set retains 191 of these characters, defines characters for 30 code points that were previously unused, and redefines 31 code points to provide more useful characters. The new character set therefore contains 252 displayable characters along with four reserved code points for control and compatibility reasons.
Having agreed the new character set, the TC task force began designing the presentation of each character on a 14-segment starburst display. These displays are often used for products at the cheaper end of the market. For the majority of characters, there was an obvious representation, but for a few, the decision was not easy. We debated whether to allow some variations to take account of the aspect ratio and segment shape of different displays, but in the end we agreed that a single presentation for each character would be best.
There are some compatibility issues related to this change. However, these are expected to be minimal. There are four possible combinations of old and new character set usage for broadcasters and receivers, as presented in the table below.
|Broadcast character set||Receiver character set||Compatibility issue||Mitigation|
|Original||New||The broadcaster uses a character that exists only in the original character set. The receiver displays the new character.||Broadcasters who have not yet updated their equipment should not use the characters that have been replaced. Since these characters are unlikely to be used, the impact should be minimal.|
|New||Original||The broadcaster uses a character that exists only in the new character set. If the character was defined, the receiver will display the original character.||This combination is more likely in established markets where large numbers of receivers have already been sold. However, in these markets the new characters are less likely to be used as the original character set supported the required languages.|
The character set is standardised in Annex C of ETSI TS 101 756v1.8.1. All broadcast equipment should be updated to use the new character set, and all new equipment designs, both for broadcast and reception, must be designed for the new character set.
Please note that this article was revised 03.12.15.