March 28, 2008 -
Vocoders Quality Analyzer v2.1 enables speech codec developers to automatically check voice quality that is returned after processing in voice coder. Speech model and algorithms are based on patents RU 61924 and RU 2312405 and software provides precise estimations automatically. Original and decoded signals are compared using internal audio signal comparison function. Window with signal similarity coefficient is displayed, representing identical signals as 100%.
HELSINKI, Finland - March 25, 2008 - Sevana Oy today announces the immediate availability of Vocoders Quality Analyzer 2.1, a new tool that enables speech codec developers to automatically check voice quality that is returned after processing in a voice coder. Featuring a unique speech model and algorithms, based on patents RU 61924 and RU 2312405, the product is one of the easiest and fastest ways to check a vocoder's quality.
Typically, quality of voice coders is measured according to the Mean Opinion Score (MOS), which provides a numerical indication of the perceived quality of received media after compression. The MOS is expressed as a single number in the range 1 to 5, where 1 is lowest perceived quality, and 5 is the highest perceived quality. Obtaining MOS estimations is time-consuming and expensive as it requires hiring experts to make estimations. When a voice coding system is under development, or the developer has to test and compare a couple of audio systems, it's very important to have a possibility for a quick check. Vocoders Quality Analyzer provides precise estimations automatically without the need for overpaid experts. According to the internal tests, the product provides results quite close to what CISCO Systems provide as MOS values for different codecs, but much faster.
The mechanism of checking a vocoder's quality works like this. Vocoders Quality Analyzer requests the codec's frame size and generates a test set of audio data according to the proper frame size. Then a correspondent vocoder is loaded from the DLL and the program passes the audio signal array together with all parameters into it. Vocoder starts encoding. Original and decoded signals are compared using an internal audio signal comparison function, and a window with signal similarity coefficient is displayed. 100% corresponds to identical signals.
The demo version of the product lets developers test their own vocoders but comparison of two audio files is available only in the full version.
On request, Sevana engineers can provide custom editions of the product tailored specifically to the needs and software projects of the customer.
Vocoders Quality Analyzer 2.1 is available today.
About Sevana Oy
The company was founded in 2003 and specializes in the production of affordable software with competitive features, as well as providing outsourcing services. So far the company has developed solutions for data mining, speech processing and business processes automation for SMEs. For more information, please visit http://www.sevana.fi.
Postal address:
Sevana Oy
c/o Ecotax AB
Agricolankatu 11
00530 Helsinki
Finland
Phone: +358 9 23164165
Mobile: +372 53485178
Contacts:
General Information:
Endre Domiczi CEO/Co-founder
INTL
Phone: 358-9-23164165 E-mail this person Company Information: Name: Sevana Oy Address: c/o Hiihtäjäntie 4 C 14 City: Helsinki ZIP: 00810 Country: Finland Phone: http://www.sevana.fi
News provided by ThomasNet Industrial Newsroom® (TIN). TIN is a comprehensive source
of new and timely product information in the industrial marketplace. TIN supplies new product
information to the web sites, e-marketplaces and print publications that serve the
industrial marketplace. For press release submissions please go to
http://www.productnews.com/SubmitPRPage.html.
BY ACCESSING, BROWSING AND/OR USING THIS WEB SITE AND/OR ANY WEB SITES PROVIDED BY
ProductNews.com, YOU AGREE TO BE BOUND BY THE TERMS OF USE
AGREEMENT.
![[Get Copyright Permissions]](http://license.icopyright.net/images/icopy-w.gif){kind=small}
