NHK OPEN HOUSE 2012 | Exhibition

9 System for Adjusting Background Sound Levels in TV Programs

As people get older, they tend to find background sounds in TV programs too loud, and find it difficult to hear what is spoken by actors and announcers.We are researching technology that enables people to adjust the balance of speech and background sounds in their own homes to make programs easier to hear. We have built a prototype that can discriminate speech/non-speech segment automatically, and adjust the background sound level with user control.


Automatic discrimination of speech and non-speech segment. We have developed an automatic discrimination method that can identify speech segments and non-speech segments of TV audio with high precision. By applying suitable signal processing to the speech segments and non-speech segments, we can reduce the loss in audio quality that accompanies changes in the background sound level. Independent adjustment of speech segments and non-speech segments It is possible to independently control background sound suppression processing of speech segments and non-speech segments. This makes it possible for individual users to fine-tune the system for their own ease of listening.

We are developing a program audio adjustment system that offers elderly viewers even greater ease of listening by combining the technology presented here with a method for controlling the acoustic characteristics of speech.