Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
as per https://en.wikipedia.org/wiki/Comfort_noise Comfort noise:
Comfort noise is artificial background noise used in radio and wireless communications to fill the silent time in a transmission resulting from voice activity detection.
Voice activity detection (VAD, or VOX) is a form of squelching where low volume levels are ignored by the transmitting device. In digital transmissions, this saves bandwidth of the communications channel by transmitting nothing when the source volume is under a certain threshold, leaving only louder sounds (presumably the speaker's voice) to be sent.
The result of replacing low volume, mostly ambient sounds with complete silence has a number of unanticipated effects on the listener, such as:
* the listener may believe that the transmission has been lost
* the speech may sound "choppy" and even be hard to understand
* the sudden change in sound level is jarring to the listener
To counteract these effects, a comfort noise generator can be used, usually on the receiving end, to fill in the silent portions of VAD-processed transmissions with artificial noise, with a low volume level appropriate for the average volume level of the received signals.
& as per www.answers.com Comfort noise:
A tone that is generated for packetized voice transmission to simulate the traditional TDM circuit where you hear a slight background noise when nobody is talking. Dead silence is disconcerting for most people.
Sinu_Gowde has covered the use of comfort noise in much detail, however I intend to explain the use of CNG (Comfort Noise Generator) from an audio subsyatem point of view, when DTX/VAD (Discrete transmission/ Voice activity detection) is ON in the sytem then the algorithm has to options
1) To mute the system by applying decaying mute or a slow mute which gradually decrease the output instead of an abtrupt decrease.
2) To call up-on the CNG.
Generally if the first method is employed then the audio syatem has to be ramped-up i.e the gains of various stages and states values have to be re-evaluated the time taken for this is though very small is considerable during call, using CNG saves us from this ramping hence the latency over communicaton is avoided.
No comfort noise is not same as the white noise. Comfort noise is detemined by the nature of signal that was present at receiver before no speech was detected [No Voice Activity Detection] based on the previous input samples the CNG buffer interpolates the Comfort Noise.
Attached is the ETSI spec of VAD this should give you some idea and the mathematical equation of the CN in general.