This is a technical tutorial on CODECs Digital Signal Processors.

This is an update to an earlier blog item but instigate by the following announcement:
"Verizon Pushing Ahead With Integrated Transcoding in IP Core, Speeding Path to Convergence - Design Eliminates 'Double Transcoding' for Multimedia Content, Regardless of User Gear
Working with our next-generation network infrastructure suppliers, Verizon has designed a transcoding architecture for our packet-based core network that supports Verizon's direction toward open networks by facilitating the introduction of media encoding technology while ensuring interoperability," said Tim Dwight, senior technologist at Verizon, during a panel discussion at Supercomm2009.
"For example, in the case of VoIP, our design resolves the 'tower of Babel' problem by allowing the sender and the receiver to negotiate a common encoding format, which, if successful, eliminates the need for media format conversion, or transcoding, and provides a network-based media conversion capability for use in cases where the end devices support no common codec," he said.
"It does this all in the packet domain, avoiding reliance on the circuit-switched core. And where transcoding is necessary, it is performed directly between the media encoding formats required by each device, eliminating the double transcoding problem that plagues networks that elect to interwork disparate access technologies across a circuit switched core."
In such networks, media is produced by the sending device in one format, transcoded by the network into an intermediate format, and later transcoded from that intermediate format to the one required by the receiving device. This increases cost and degrades service quality, compared with the Verizon solution."

Meanwhile here are more details pertaining to this issue:
A CODEC-COder-DECoder (also known as an encoder-decoder and COmpression-DECompression system when used in video systems) is a computer chip (semiconductor) digital signal processing system. Source codecs are designed specifically for speech, whereas Waveform codecs work well with any type of sound. Depending on the audio or voice application would drive the selection of the Source or Waveform CODEC. While there are many types of CODECs, G.711 & G.729 are the two most-commonly CODECs used in VoIP systems. Shown in the animated tutorial is a G.711 encoded audio stream is 64/56/48 KBPS-Kilo Bits Per Second. Each 13/14 bit sample of the original signal (voice-audio) is encoded into an eight Byte/Octet. Compression algorithms operate by sampling voice and quantizing the analog sound into digital values. G.711 is based on traditional Nyquist-Shannon sampling theorem that the sampling frequency rate must be at least twice as high as the highest input frequency for the result to closely resemble the original signal. A 4,000 Hz-Hertz voice pattern would be sampled at a rate of 8,000 BPS-Bits Per Second. Next, in the animated tutorial here is a G.729 coding at 8,000, not 64,000 samples per second. Transcoding is also related to the concept of Tandem Encoding or Tandem Compression. Tandem Encoding is traditional the concept of the transfer of TDM traffic between different telephone carriers via tandem Class 4/5 switches as they process telephone calls. Tandem Encoding is also the process of interconnecting same or different company packet voice. That is, different CODECs (e.g. G.711/729) may be used at different locations within the same company and more likely between different companies. Most SIP providers support G.711 and G.729. RTA-Real-Time Audio developed by Microsoft is used with OCS-Office Communications Server. MOS-Mean Opinion Scores reduces rapidly with each time a voice conversation is processed by a CODEC. Whether you call it Transcoding or Tandem Encoding, the CODEC in a SIP network should be standardized whenever possible.

Here's the "so what" or "why should I care about this." For example, different CODEC sampling rates may start synchronized but shortly become un-synchronized which can cause encoding problems and voice to jitter. To measure and manage jitter, RTP-Real Time Protocol uses the time-stamp function in the protocol to assess jitter based on the delay between arrival (interarrival) times of each packet. Changing the number of bits sampled and quantized can dramatically impact the voice quality. However, LAN-Local Area Network and WAN-Wide Area Network bandwidth limitations may have an equal or greater impact on VoIP performance. Echo can also occur as a result of Asynchronous Transcoding. Transcoding is the process of conversion between circuit-switched (PSTN-Public Switched Telephone Network) and packet-switched networks such as Frame Relay, IP-Internet Protocol and ATM-Asynchronous Transfer Mode. The point is that Asynchronous Transcoding should be avoided. According to Intel, "The term "asynchronous transcoding" refers to a situation when, for example, one endpoint is talking G.711 to another endpoint talking G.729 or two different encodings." According to Cisco, "Although it can seem logical from a financial standpoint to convert all calls to low-bit rate codecs to save on infrastructure costs, exercise additional care when you design voice networks with low-bit rate compression. There are drawbacks to compressing voice. One of the main drawbacks is signal distortion due to multiple encodings (called tandem encodings). For example, when a G.729 voice signal is tandem encoded three times, the MOS-Mean Opinion Score drops from 3.92 (very good) to 2.68 (unacceptable). Another drawback is codec-induced delay with low bit-rate codecs." Our analysis and recommendation is to always use G.711 based on the assumption that processing voice across other networks will probably use lower-bit rate CODECs reducing MOS quality. In other words, start with quality and hopefully but not likely you will end with quality. In addition, there is no requirement by any customer or vendor to provide end-to-end G.711 or any other CODEC. If you consider multi-party conference calls, the problem gets compounded.