Speex
lastmod 2014-09-19created 2014-09-192 edits
Speex is a patent-free audio compression format developed for speech, and a freeware speech codec used in VoIP software and podcasts. It is based on the CELP speech coding algorithm. Speex aims to be free of all patent restrictions and is licensed under the revised (3-clause) BSD license. It is used with the Ogg container format or transmitted over UDP/RTP.
The Speex developers see the project as complementary to the Vorbis general-purpose audio compression project. Speex is a lossy compression format that permanently degrades quality to reduce file size.
The Speex project started on February 13, 2002. The first development versions of Speex were released under the LGPL license, but like the 1.0 beta, Speex was released under the Xiph Foundation's revised BSD license. Speex 1.0 was announced on March 24, 2003. The final version of the Speex encoder and decoder is 1.1.12. The Xiph.Org Foundation considers Speex obsolete and regards Opus as its successor.
Description
Speex was developed for use in VoIP and file-based compression. The development goal was a codec optimized for high-quality speech and low bitrates. To achieve this goal, Speex uses multiple bitrates and supports ultra-wideband (32 kHz sampling rate), wideband (16 kHz sampling rate), and narrowband (telephone quality, 8 kHz sampling rate). Because Speex was made for VoIP rather than mobile phones, it has to be robust to lost packets. For this reason, CELP was chosen as the encoding technique for Speex. One of the biggest reasons is that CELP has long been proven to perform well at both low and high bitrates and to scale well. Its key features are as follows.
- Freeware/open source, patent-free, royalty-free
- Integration of narrowband and wideband in the same bitstream
- Wide range of available bitrates (2 kbit/s to 44kbit/s)
- Dynamic bitrate switching and variable bitrate (VBR)
- Voice Activity Detection (integrated with VBR) (not supported in version 1.2)
- Variable complexity
- Ultra-wideband mode at 32 kHz (up to 48 kHz)
- Intensity stereo encoding option