Development and analysis of an International Speech Test Signal (ISTS)

Abstract

Abstract

For analysing the processing of speech by a hearing instrument, a standard test signal is necessary which allows for reproducible measurement conditions, and which features as many of the most relevant properties of natural speech as possible, e.g. the average speech spectrum, the modulation spectrum, the variation of the fundamental frequency together with its appropriate harmonics, and the comodulation in different frequency bands. Existing artificial signals do not adequately fulfill these requirements. Moreover, recordings from natural speakers represent only one language and are therefore not internationally acceptable. For this reason, an International Speech Test Signal (ISTS) was developed. It is based on natural recordings but is largely non-intelligible because of segmentation and remixing. When using the signal for hearing aid measurements, the gain of a device can be described at different percentiles of the speech level distribution. The primary intention is to include this test signal with a new measurement method for a new hearing aid standard (IEC 60118-15).

Sumario

Para analizar el procesamiento del lenguaje por medio de un instrumento auditivo, se requiere de una señal estándar de prueba que permita obtener condiciones reproducibles de medición, y que contenga la mayor cantidad de las propiedades más relevantes del lenguaje natural como sea posible, p.e., el espectro promedio del lenguaje, el espectro de modulación, las variaciones en la frecuencia fundamental junto con sus armónicos apropiados, y la co-modulación en diferentes bandas de frecuencia. Las señales artificiales existentes no llenan adecuadamente estos requisitos. Más aún, los registros de un hablante natural representan solo un idioma y por lo tanto, no son aceptables internacionalmente. Por esta razón, se desarrolló una señal internacional de prueba para lenguaje (ISTS). Se basa en registros naturales pero es esencialmente ininteligible debido a la segmentación y el re-mezclado. Cuando se usa la señal para mediciones de un auxiliar auditivo, la ganancia del dispositivo puede ser descrita a diferentes percentiles de la distribución de nivel de intensidad del lenguaje. La intención primaria es incluir esta señal de prueba con un nuevo método de medición para un nuevo estándar para auxiliares auditivos. (IEC 60118-15).

Key Words:

• Test signal
• Hearing aid measurement
• International standard
• Speech

Acknowledgements

Thanks to:

• The ISMADHA working group for the very fruitful and intense cooperation and expert monitoring: Nikolai Bisgaard and Brian Dam Pedersen at GN Resound; Volker Kuehnel at Phonak; Frank Rosenberger at Siemens; Ivo Merks at Starkey Laboratories, Carsten Paludan Müller at Widex, Johnny Andersen at Oticon, and Todd Fortune at Interton.

• EHIMA for financial support.

• The international expert advisors Robyn Cox, Harvey Dillon, Gitte Keidser, and Åke Olofsson for their very valuable input.

• Volker Hohmann, Jörg Bitzer, Kirsten Wagener, and Kathrin Kliem for fruitful discussions and support.

• Richard Schultz-Amling, Jörn Anemüller, Jörg Bitzer, Monika Kappelmann, Björn Ohl, Jan Schaffmeister, and Marco Wilmes for their technical support.

• Three anonymous reviewers for helping to improve the paper.

• Last but not least, thanks to the speakers for reading out the story several times.

• This study was co-funded by AGIP, the Lower Saxony Department of Science and Culture, Hanover, and the European regional funding EFRE.

Parts of this article were presented at the Erlanger Kolloquium audiologisch tätiger Physiker und Ingenieure, February 15–16, 2007, Erlangen, Germany; 33. Deutsche Jahrestagung fÜr Akustik – DAGA 2007, March 19–22, Stuttgart, Germany; 10th Congress of the German Society of Audiology, June 6–9, 2007, Heidelberg, Germany; and the International Hearing Aid Research Conference (IHCON), August 13–17, 2008, Lake Tahoe, California, USA.

Declaration of Interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.