We have been undertaking research into audio signals for the benefit of the elderly. The results of our research were translated into practice in February 2004 when NHK inaugurated in a number of programs that it broadcast an audio service carried on a digital sub-channel and provided for elderly people. This service takes advantage of the characteristics of digital transmissions. Normal sound is sent over the main channel while sound processed to make voice signals easier to hear is sent over a sub-channel, in effect making the broadcast stereophonic. In homes equipped to receive digital broadcasts, it is unnecessary to install special receivers, as viewers can avail themselves of this service, which is the first of its kind in the world, simply by using their remote controller to select the sub-channel.


1. Is it difficult to recognize the diminished sensitivity to sound that comes with old age?

As we get older our eyes become weaker, and without our realizing it our sense of hearing weakens, too. Hearing difficulties caused by aging - "the aging of the ears"- make it it harder to catch everything that is being said in a broadcast. There are, of course, individual differences, but from the age of 65, the condition becomes obvious. We are now on the threshold of digital broadcasting, by which it will be possible to transmit up to twelve sound channels for a single program. Would it not be possible, by making use of this, to establish a special service for the elderly? Our research program is examining this.


2. A detailed look at the service and research activities

To begin with, the "Silver Acoustic Research Committee" started by considering the form the service like this should take. The following three possibilities were put forward.

1) A differential level method: This would broadcast audio signals, with BGM and SFX at different (i.e. lower) levels than ordinary mixing levels.

2) A discreet level method: This would broadcast dialogue tracks and M-E (music and SFX) tracks as discrete channels (2-stereo format). The viewer could then change the balance between the channels to obtain the best balance for listening.

3) A dialogue frequency data option method: This would broadcast dialogue frequency energy data as an optional data service. A built-in data decoder in the viewer’s TV set would then control the dialogue to give the best level.
We reached the conclusion that of these three, the one that would constitute the smallest burden on the listener, and at the same time could be produced and transmitted most economically, and efficiently was the first one.


3. Evaluation testing for establishing the optimum parameters, and the results obtained

We carried out experiments for each type of program in various genres to ascertain the extent to which, in comparison with normal balance, the levels of BGM and background sound effects should be lowered to obtain a condition in which the program would be clearly heard and easy to follow. For each sample, the value for voice sounds in normal balance was given the base value 0 and compared with BGM and sound effects paired in a series of values: +3db, -3db, -6db, and -10db.

For the evaluations, subjects listened to the voice components and rated them according to three criteria: ‘easy to hear’, ‘difficult to hear’, or ‘average’.

At the same time, we carried out a survey under actual home viewing conditions. Viewers were given questionnaires, enabling us to gather comments and impressions of the sound components of TV programs, and by means of audiometers we obtained measurements of hearing characteristics (Fig.1). The same methods were used in all cases.

250 viewers were sampled (230 aged 63 or more; 20 under the age of 30). Analysis of the data obtained showed that when music and sound effects were lowered by 6db from normal balance levels, the intensity of background sounds was not lost and voice sounds were easy to follow. Lowering levels further caused the loss of effectiveness of background sounds and reached the borderline for obtaining data.

From the data on hearing characteristics, it was seen that as frequencies went from 3KHz up to a high level, between 30 and 60db, elderly listeners all found a rapid increase in the difficulty they had hearing sounds. From questionnaire replies and free comments also we were able to gain useful hints.


4. Looking to the Future

Japanese society is continuing to age. In the year 2010, 22 percent of the population will be 65 or older. We consider that being able to hear important information clearly, and without finding the effort tiring, will be appropriate to the needs of society then. But a service that makes sounds easy to hear is not the concern of the transmission side alone, the sound reproduction functions of TV receivers also play a part in it. The idea of being "listener friendly" does not simply apply to extending services to the elderly, it can also be applied to listening to foreign broadcasts, or to the intelligibility of sounds in public spaces, classrooms, railway stations and such environments.

The HBF provided the above research project with a grant.

For further information about this research, please make contact with Mr. Mick M. Sawaguchi, Director of NHK Program production/ops center
(mail to : sawaguchi.m-em@nhk.or.jp).