Speech coding research at Bell Laboratories
Abstract
The field of speech coding is now over 70 years old. It started from the desire to transmit voice signals over telegraph cables. The availability of digital computers in the mid 1960s made it possible to test complex speech coding algorithms rapidly. The introduction of linear predictive coding (LPC) started a new era in speech coding. The fundamental philosophy of speech coding went through a major shift, resulting in a new generation of low bit rate speech coders, such as multi-pulse and code-excited LPC. The semiconductor revolution produced faster and faster DSP chips and made linear predictive coding practical. Code-excited LPC has become the method of choice for low bit rate speech coding applications and is used in most voice transmission standards for cell phones. Digital speech communication is rapidly evolving from circuit-switched to packet-switched networks to provide integrated transmission of voice, data, and video signals. The new communication environment is also moving the focus of speech coding research from compression to low cost, reliable, and secure transmission of voice signals on digital networks, and provides the motivation for creating a new class of speech coders suitable for future applications.
Related Papers
- → CELP Coding for high-quality speech at 8 kbit/s(2005)10 cited
- → Celp With Priority To Critical Segments(1998)
- → Pitch-controlled variable bitrate CELP speech coding(1998)
- → A 4 kbit/s renewal code excited linear prediction speech coder(2002)
- Variable-rate enhanced double-band excitation LPC low bit rate speech coding algorithm(2003)