Transmission of speech in its raw format requires a lot of bandwidth. Software components
that compress the raw speech data into smaller bitstreams are called speech encoders.
Software components that uncompress that compressed bitstream back into the raw speech
suitable for play out using a loudspeaker are called speech decoders. Typically, an
implementation will consist of both an encoder and a decoder. For simplicity, such an
implementation is then usually called a "speech codec."
Agora Labs has seven speech codecs to choose from. Two of our speech codecs support the
ITU-T International Recommendations G.711 and G.722. Each of these speech codecs were
developed entirely at Agora Labs. In addition, Agora Labs has five proprietary speech codecs:
the SX1200P, SX2000P, SX7300P, SX8300P, and SX9600P which support 1200 bits-per-second (bps),
2000 bps, 7300 bps, 8300 bps, and 9600 bps bandwidth respectively. Each of these speech codecs
were originally developed by Lucent elemedia™.