The motivation to create AmpliCell

About 5% of all mobile phone users suffer from moderate to severe hearing loss.
Only a quarter of all people reporting hearing impairment own and use a hearing aid, but most hearing impaired are nowadays users of mobile phones. The number of mobile phones sold is several hundred times larger than the number of hearing aid sold. Sources:
Hearing impaired mobile phone users are additionally challenged when using a mobile phone: GSM signal artifacts and environmental noise add to intelligibility difficulties during phone conversations. Often, hearing impaired mobile phone users follow a simple but not necessarily satisfying strategy: they increase the volume of their mobile phones to the maximum. This results in a frequency independent linear amplification which does not reflect the user's individual amplification needs.
How Traditional Hearing Aids address Individual Amplification Needs
The characteristics of hearing losses are highly individual and hearing thresholds vary substantially from person to person. The most common type of hearing loss is the sensorineural hearing loss which is caused by an impaired inner ear. This type of hearing loss is characterized by the loudness recruitment phenomenon. To compensate loudness recruitment more amplification is needed for soft sounds than for loud sounds. The core functionality of the software hearing aid is thus (a) compensating for the frequency dependent sensitivity loss of the impaired human ear and (b) compensating for loudness recruitment by means of a non-linear automatic gain control (AGC).
AmpliCell addresses this issue and helps out by implementing a software hearing aid on the phone. The implementation of the software hearing aid is based on the core functionalities of traditional hearing aids, but it benefits from the technological possibilities of modern mobile phones. People with a mild to moderate hearing loss that will benefit most from a better sound quality AmpliCell provides.
Technological Possibilities for Amplification in Mobile Phones
  1. In comparison to traditional (physical) hearing aids, the frequency range is much wider, ranging in many cases from 20 Hz to 20.000 Hz. While the smaller frequency range of traditional hearing aids is sufficient for speech, this wider range allows excellent sound quality when listening to music.
  2. The input to the signal processing in the software hearing aid is the audio signal of the phone conversation. Obviously, the signal from the phone's microphone itself is not amplified. This is not the case in hearing aids because hearing aids can not easily distinguish between the users own voice and the voice of the incoming call. The fact is beneficial for the development of the software hearing aid in a mobile phone. Firstly, feedback as known from physical hearing aids cannot occur. Secondly, a signal delay in the order of magnitude of one hundred milliseconds will go unnoticed by the user. Thus, a look-ahead delay can be introduced in the AGC algorithm in order to improve sound quality. Thirdly, there is no risk for comb filter effects due to the superposition of direct (un-amplified) and delayed path, no unpleasant pre-echo effects can occur and no stuttering is provoked by any delayed and amplified own voice signal.
  3. In addition, the software hearing aid in a mobile phone can make a qualified guess whether the signal to be amplified is speech or music: a signal that stems from the phone's GSM unit consist most likely of speech while the content of media, such as MP3 files, consists most likely of music. This assumption allows tailoring the signal processing to either speech intelligibility or sound quality.

Source Code

We have decided to publish our Android source code. At this stage, the audio signal from phone conversation can not yet be captured; see issue 2117.

iPhone development is stalled until background processes are allowed on the iPhone.

About AmpliCell

AmpliCell is made by
Joachim Neumann
Dr. Joachim Neumann is a German physicist living in Barcelona and presently employed at Telefonica Research. He earned his Ph.D. degree at one of the leading audiological laboratories in Europe: the department for medical physics at the University of Oldenburg, Germany. Since 1997, Dr. Neumann is working in the hearing aid industry. He obtained his profound knowledge in digital signal processing at the research center 'Eriksholm' which does "blue sky" research for the Danish hearing aid manufacturer Oticon, one of the largest hearing aid manufacturers. His developments found their way into millions of hearing aids and he owns several patents in the field.
joachim@joachimneumann.com
Nicolas Wack
Nicolas Wack is a telecommunication engineer who received in 2003 his degree with majors in signal processing and computer science from the ENST in Paris. Since then, he has been leading the development of the music analysis and the search engine developed at the Music Technology Group of the Universitat Pompeu Fabra, which is also now commercialized by BMAT.
wackou@gmail.com