TIA TIA-894

ONLY AVAILABLE IN PDF FORMAT

This document specifies the procedures to test implementationsof SMV-compatible variable-rate speech codecs either by meetingthe bit-exact implementation, or meeting recommended minimumperformance requirements. This speech codec is the ServiceOption 56 described in TIA-893-1. The Service Option 56 speechcodec is used to digitally encode the speech signal fortransmission at variable data rates of 8550, 4000, 2000, or 800bps. Like some other speech coding standards, this standardprovides a bit-exact method of testing the codec for minimumperformance. In this optional procedure, a given set of testvectors are input to the test codec, and the output vectors fromthe test codec must be bit-exact with the output vectors generatedfrom the Software Distribution which is part of this standard. Ifthey are bit-exact, the codec passes the minimum performancerequirement and no further testing is required. The bit-exact modeof testing, however, is only applicable to codecs whose designconforms in all respects to the algorithmic description of the SMVin TIA-893-1, including either of the noise suppression options andeither of the voice activity detection options. These options aredescribed in TIA-893-1. The bit-exact SMV validation includesspecific testing of the OEM SMV to the frame erasure processingcharacteristics of the standard SMV. Should the candidate SMVdiffer in any of these components, the codec shall be tested usingthe objective and subjective tests prescribed by this standard.That is, SMV compliance of a “test codec” can be achieved byeither:

• demonstrating bit-exactness according to the proceduredescribed in Section 3.6, specifically Sections 3.6.5.1 to 3.6.5.4,or

• following the objective and subjective testing procedures setforth in Sections 2.1 and 2.2 of this standard, using processesadditionally defined in Section 3.6, specifically Sections 3.6.5.5to 3.6.5.8.

With the exception of Section 3.6, the remaining text appliesonly to implementations that do not satisfy the requirement forbit-exactness.

Testing the codec is based on two classes of procedures:objective tests and subjective tests. In the event that the testcodec fails any of the objective or subjective tests, the testcodec fails the compliance test. Objective tests are based uponactual measurements from the speech codec function. Subjectivetests are based on listening tests to judge overall speech quality.The minimum subjective requirement for the test codec is based uponthe ability of the test codec to demonstrate performance equivalentto or better than that of the SMV floating-point master codecwithin a fixed allowable statistical error.

The purpose of the testing is not only to ensure adequateperformance between one manufacturer’s encoder and decoder but alsothat this level of performance is maintained with operation betweenany pairing of manufacturers’ encoders and decoders. Thisinteroperability issue is a serious one. Any variation inimplementing the exact standard must be avoided if it cannot beensured that minimum performance levels are met wheninter-operating with all other manufacturers’ equipment meeting thestandard. This standard provides a means for measuring performancelevels while trying to ensure proper interoperation with othermanufacturers’ equipment.

The issue of interoperation can only be definitively answered bytesting all combinations of encoder/decoder pairings. With thenumber of equipment manufacturers expected to supply equipment,this becomes a prohibitive task; therefore, the objective andsubjective tests rely upon the use of a “master codec”. The mastercodec is defined as the floating-point implementation of TIA-893-1written in the C-programming language.

The master codec software, described in 3.4, is used as part ofthe interoperability testing. The master er codec is available inan associated Software Distribution as part of TIA-893-1.

By convention in this document, the Courier font is used toindicate C-language and other software constructs, such as file andvariable names.

Edition:

05

Published:

08/01/2005

Number of Pages:

192

File Size:

1 file , 730 KB