VoIP Voice and Fax Signal Processing

6.10 ECHO CANCELLER TESTING

As per G.168 requirements [ITU-T-G.168 (2004)(2004)], the tests recommended are given in Table 6.3. These tests are identified with a test number. The numbers remained the same across several revisions of echo canceller recommendations, and the results are usually referenced with the test number. In the table test number, the test name and a brief description are given. For more details on these tests, it is suggested to refer to the G.168 recommendation [ITU-T-G.168 (2004)]. These tests ensure time-dependent echo attenuation [ITU-T-P.340 (2000)] for different power levels, NLP tracking and time responses, double talk conditions, echo attenuation over frequency, and characteristics of fax/modem signals and tones.

Table 6.3. Echo Canceller Tests as per G.168 Recommendation

6.10.1 Simulated Tests

Echo canceller test signal generation, testing procedure, and timing sequence are given in G.168. The test signal consists of synthetic speech, pseudo-random noise, and silence zones. Developers/implementers also generate additional test vectors to create various corner cases and actual conversation speech files. Such an extended suit is used for internal testing. Most of these tests are useful for actual software development and a simulation environment. These ...