Digital Speech Standard
The Digital Speech Standard (DSS) is a proprietary (i.e. not open), manufacturer-independent audio codec for the low-loss storage of speech data in dictation devices .
Emergence
A few years before the market launch of the first digital dictation machines, Grundig developed the first DSS codec in cooperation with the University of Erlangen-Nuremberg in 1994 . Building on this, the DSS standard was presented in 1997 as a joint development by Grundig, Olympus and Philips . At the same time, the manufacturers involved founded an open standardization committee - the International Voice Association (IVA) - for the uniform specification and further development of the DSS standard.
Due to high requirements for device development and license fees, DSS is only used in the professional devices of IVA members.
properties
The DSS audio codec was specially developed to store human speech and, like the much better known MP3, uses a psychoacoustic model . A relatively high compression rate is achieved with only a small loss of quality. Nevertheless, the lossy compression worsens the recognition rate of computer-aided speech recognition . This weakness was improved in the subsequent “DSS Pro” standard.
In addition to optimizing the memory requirements of the recordings, further information can be added to the recordings. For example, in the header files of the recordings. B. an author identifier or similar. get saved.
DSS files can be further processed by special software; the manufacturers of the dictation solutions offer correspondingly powerful writing software and workflow management programs. But there are also free playback programs. However, further processing with other software is not possible. The demonstration program is only available for Microsoft operating systems. Olympus also offers free playback software for macOS (DSS-Player Lite), which can be downloaded from the website there. Unlike DSS Player for MAC, the Lite version of the software has not been maintained since 2003 and no longer runs under macOS Sierra (as of Oct 2012).
After a brief examination of the DSS format, the following findings were obtained:
- DSS is a container that can be used to store different codecs.
- There are currently two known recording modes: LP and SP. LP uses G.723.1 codec.
The successor DSS Pro
The quasi-standard for professional applications DSS was slightly modernized in its second version: On the one hand, the sampling rate was increased to 16 kHz in order to further improve the speech quality and to better meet the requirements of speech recognition software. With a DSS Pro dictation device it is now possible to transfer dictations directly to the Dragon NaturallySpeaking speech recognition system . The first dictation machines with DSS Pro are the Philips DPM 9600 or 9620 and the Digta 420 in two slide switch variants from Grundig Business Systems. With the DS-5000, Olympus has meanwhile also brought a DSS Pro capable dictation device onto the market.
Individual evidence
- ↑ grundig-gbs.com ( Memento of the original from October 11, 2007 in the Internet Archive ) Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice.
- ↑ diktiertechnik.de
- ↑ Olympus DSS Lite Player July 12, 2006
- ↑ bugzilla.libav.org
- ↑ diktiertechnik.de
