The HaH project aims to overcome the limitations of state-of-the-art supportive audio signal processing (SASP) technology and research and develop powerful SASP technologies. These efforts are to parallel visual lip-reading support for home information, home entertainment, home automation, and home care applications running on a STB-like Home Information and Communication (HIC) platform.

Achieving these aims ensures the accessibility of public services for the hearing-impaired benefiting from the STB emerging technology. The HaH project helps to provide equal access, independent living, and participation for all in the Information Society. To reflect the multi-disciplinarity of the field, HaH will bring together SASP developers, researchers on artificial face synthesis for lip-reading, home information, home entertainment, and home automation service providers, and representatives of Europe?s hearing-impaired. The latter guarantees an optimal level of user involvement to draw attention to the human-machine interfaces of the HIC platform with hearing support.

Another objective focuses on the adaptation of an STB platform to overcome the current limitations in an environment for hearing impaired persons. STB platforms are hardware with limited CPU performance. It will be especially challenging or even impossible to implement the SASP functionality for first and second generation STBs. The main focus therefore lies on a STB-like HIC platform that is similar to third and next generation STB systems with more advanced hardware.

The core of the SASP will be an advanced, PC-based audio signal processing framework for hearing support applications, which we shall henceforth refer to as the "Master Hearing Aid" framework, or MHA. Instead of receiving an audio stream from a sophisticated PC-based soundcard, SASP will be required to process multiple audio formats (mono, stereo, 5.1, DTS, ?) with possibly limited or even extended quality compared to standard hearing support technology. Furthermore, the signal processing has to be adapted to fulfil the needs of the visual support algorithms.

The results of the Synface project shall be adapted and ported to compensate for the limitations of the HIC platform, as it will be defined. For instance, a special issue will be a known delay of 200 ms that is inherent to the system under the current implementation of support for a telephone conversation. Similarly, for the processing of combined audio-video signal, the video component must be buffered through a time-shift operation and resynchronized with the audio stream.

In addition, the HIC platform needs to be integrated with or connected to a Residential Gateway (RG) that is the entry point to the home networks (HAN). The RG is a commercial platform supporting the use of an OSGi framework. This will enable the HIC platform to interact with various home automation networks such as Lonworks, X10, or EIB. The OSGi specification is a standard that allows the dynamic provisioning, management, and sharing of services via its component integration platform in a very efficient way. The OSGi standard is currently Java-based, allowing the integration of new services and the removal of no longer needed services at runtime. Within the current standardization process, support of the handicapped has never been a big topic. Focussing on hearing-support, as in HaH, will create the necessary awareness among software developers and facilitate the necessary software techniques and standards for future product development in this direction.

It is the core objective of the HaH project to combine the three technologies MHA, Synface, and home automation on a STB-like HIC platform. Unlike hearing aids, STB and similar technologies are widely accepted. As a result, little or no acceptance problems are expected.

As previously mentioned, the SASP strategy decomposes into front-end and back-end components. The front-end component will include user-independent and thus global SASP (G-SASP) functionality. This component is responsible for audio signal pre-processing, that is, noise-reduction and signal classification for context-aware processing of audio signals. Even the normal hearing will benefit from G-SASP. The back-end, user-dependent component, and thus individual SASP (I-SASP) functionality, requires a user-based fitting tool for parameterization. It is a further objective of the HaH project to enable the end-users to perform parameterization without the assistance of professional acousticians.

This strategy is a novel approach. The strategy eliminates the need for end-users of regular hearing aids to consult an expert at his office. The I-SASP functionality, to be developed within the HaH Project, shall be completely configured in the home environment. Wizards of varying complexity will be available depending on the technical capabilities of the user. For example, it will be possible to recall the last 15 seconds of a broadcast by continuously recording a TV show and playing it back, This time-shift operation, enables the user to review the context that occurred within this time frame but did not satisfy the user′s needs.