ANSI TIA EIA-136-250-2001 TDMA Third Generation Wireless - Minimum Performance Standards for ACELP Voice Activity Detection (VAD).pdf

1、 ANSI/TIA/EIA-136-250-2001 Approved: April 23, 2001 TIA/EIA STANDARD TIA/EIA-136-250 TDMA Third Generation Wireless - Minimum Performance Standards for ACELP Voice Activity Detection (VAD) TIA/EIA-136-250 APRIL 2001 TELECOMMUNICATIONS INDUSTRY ASSOCIATION The Telecommunications Industry Association

2、represents the communications sector of NOTICE TIA/EIA Engineering Standards and Publications are designed to serve the public interest through eliminating misunderstandings between manufacturers and purchasers, facilitating interchangeability and improvement of products, and assisting the purchaser

3、 in selecting and obtaining with minimum delay the proper product for his particular need. Existence of such Standards and Publications shall not in any respect preclude any member or nonmember of TIA/EIA from manufacturing or selling products not conforming to such Standards and Publications, nor s

4、hall the existence of such Standards and Publications preclude their voluntary use by those other than TIA/EIA members, whether the standard is to be used either domestically or internationally. Standards and Publications are adopted by TIA/EIA in accordance with the American National Standards Inst

5、itute (ANSI) patent policy. By such action, TIA/EIA does not assume any liability to any patent owner, nor does it assume any obligation whatever to parties adopting the Standard or Publication. Note: The users attention is called to the possibility that compliance with this document may require use

6、 of one or more inventions covered by patent rights. By publication of this standard, Interim Standard, or TSB, no position is taken with respect to the validity of those claims or of any patent rights in connection therewith. The patent holder(s) so far identified have, however, filed statements of

7、 willingness to grant licenses under those rights on reasonable and nondiscriminatory terms and conditions to applicants desiring to obtain such licenses. Details may be obtained from the standards developer. This Standard does not purport to address all safety problems associated with its use or al

8、l applicable regulatory requirements. It is the responsibility of the user of this Standard to establish appropriate safety and health practices and to determine the applicability of regulatory limitations before its use. (From Standards Proposal No. 3-4027-250, formulated under the cognizance of th

9、e TIA TR-45.3 Subcommittee on Time Division Digital Technology.) Published by g211TELECOMMUNICATIONS INDUSTRY ASSOCIATION 2001 Standards and Technology Department 2500 Wilson Boulevard Arlington, VA 22201 PRICE: Please refer to current Catalog of EIA ELECTRONIC INDUSTRIES ALLIANCE STANDARDS and ENGI

10、NEERING PUBLICATIONS or call Global Engineering Documents, USA and Canada (1-800-854-7179) International (303-397-7956) All rights reserved Printed in U.S.A. PLEASE! DONT VIOLATE THE LAW! This document is copyrighted by the TIA and may not be reproduced without permission. Organizations may obtain p

11、ermission to reproduce a limited number of copies through entering into a license agreement. For information, contact: Global Engineering Documents 15 Inverness Way East Englewood, CO 80112-5704 or call U.S.A. and Canada 1-800-854-7179, International (303) 397-7956 NOTICE OF DISCLAIMER AND LIMITATIO

12、N OF LIABILITY The document to which this Notice is affixed has been prepared by one or more Engineering Committees of the Telecommunications Industry Association (“TIA”). TIA is not the author of the document contents, but publishes and claims copyright to the document pursuant to licenses and perm

13、ission granted by the authors of the contents. TIA Engineering Committees are expected to conduct their affairs in accordance with the TIA Engineering Manual (“Manual”), the current and predecessor versions of which are available at http:/www.tiaonline.org/standards/sfg/engineering_manual.cfm. TIAs

14、function is to administer the process, but not the content, of document preparation in accordance with the Manual and, when appropriate, the policies and procedures of the American National Standards Institute (“ANSI”). THE USE OR PRACTICE OF CONTENTS OF THIS DOCUMENT MAY INVOLVE THE USE OF INTELLEC

15、TUAL PROPERTY RIGHTS (“IPR”), INCLUDING PENDING OR ISSUED PATENTS, OR COPYRIGHTS, OWNED BY ONE OR MORE PARTIES. TIA MAKES NO SEARCH OR INVESTIGATION FOR IPR. WHEN IPR CONSISTING OF PATENTS AND PUBLISHED PATENT APPLICATIONS ARE CLAIMED AND CALLED TO TIAS ATTENTION, A STATEMENT FROM THE HOLDER THEREOF

16、 IS REQUESTED, ALL IN ACCORDANCE WITH THE MANUAL. TIA TAKES NO POSITION WITH REFERENCE TO, AND DISCLAIMS ANY OBLIGATION TO INVESTIGATE OR INQUIRE INTO, THE SCOPE OR VALIDITY OF ANY CLAIMS OF IPR. ALL WARRANTIES, EXPRESS OR IMPLIED, ARE DISCLAIMED, INCLUDING WITHOUT LIMITATION, ANY AND ALL WARRANTIES

17、 CONCERNING THE ACCURACY OF THE CONTENTS, ITS FITNESS OR APPROPRIATENESS FOR A PARTICULAR PURPOSE OR USE, ITS MERCHANTABILITY AND ITS NON-INFRINGEMENT OF ANY THIRD PARTYS INTELLECTUAL PROPERTY RIGHTS. TIA EXPRESSLY DISCLAIMS ANY AND ALL RESPONSIBILITIES FOR THE ACCURACY OF THE CONTENTS AND MAKES NO

18、REPRESENTATIONS OR WARRANTIES REGARDING THE CONTENTS COMPLIANCE WITH ANY APPLICABLE STATUTE, RULE OR REGULATION. TIA SHALL NOT BE LIABLE FOR ANY AND ALL DAMAGES, DIRECT OR INDIRECT, ARISING FROM OR RELATING TO ANY USE OF THE CONTENTS CONTAINED HEREIN, INCLUDING WITHOUT LIMITATION ANY AND ALL INDIREC

19、T, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES (INCLUDING DAMAGES FOR LOSS OF BUSINESS, LOSS OF PROFITS, LITIGATION, OR THE LIKE), WHETHER BASED UPON BREACH OF CONTRACT, BREACH OF WARRANTY, TORT (INCLUDING NEGLIGENCE), PRODUCT LIABILITY OR OTHERWISE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAG

20、ES. THE FOREGOING NEGATION OF DAMAGES IS A FUNDAMENTAL ELEMENT OF THE USE OF THE CONTENTS HEREOF, AND THESE CONTENTS WOULD NOT BE PUBLISHED BY TIA WITHOUT SUCH LIMITATIONS. TIA/EIA-136-250iContentsList of Tables .ii1. Introduction11.1 Scope11.2 Definitions .12. VAD Minimum Standards .22.1 Delay Requ

21、irements.22.2 Performance Requirements 22.2.1 Test Cases 32.2.1.1 Source Speech Material.52.2.1.2 Source Noise Material.62.2.2 Test Procedure62.2.2.1 Generation of Test Cases72.2.2.2 Processing of Test Cases 72.2.2.3 Generation of Performance Metrics72.2.2.4 Evaluation of Performance Metrics 8Annex

22、A CD-ROM Contents (normative) 10Annex B Sample Performance Metrics (informative) 12Annex C Sample Evaluation Metrics (informative) .14Annex D References (informative).16TIA/EIA-136-250iiList of TablesTable 2.2.1-1 VAD Test File Cases .3Table 2.2.1.1-1 Speech Data Files.5Table 2.2.1.1-2 Truth Mark Fi

23、les .5Table 2.2.1.2-1 Noise Data Files.6Table 2.2.2.2-1 VAD Mark File Code Format7Table 2.2.2.4-1 VAD Test Evaluation Cases 8Table A-1 CD-ROM Contents11TIA/EIA-136-25011. Introduction1This document supercedes TIA/EIA-IS-727. This standard describes definitions, methods of2measurement, minimum delay

24、and performance requirements for all voice-activity detectors (VADs)3to be used with the enhanced full-rate vocoder (TIA/EIA-136-410) with comfort-noise generation4(CNG). If both TIA/EIA-136-410 (ACELP) and TIA/EIA-136-440 (AMR) are implemented in one5mobile station then GSM VAD1 or GSM VAD2 (GSM 06

25、.94) bit exact implementation is required6and the minimum performance part of this document does not apply. This standard applies to7mobile stations operating in the discontinuous transmission (DTX) mode.8This standard consists of this document and a software distribution on the accompanying CD-ROM9

26、containing ten (10) speech data files, ten (10) truth mark files corresponding to each speech file, four10(4) background noise files, the example files in Annexes B and C, and one (1) C source file for a11software tool to generate test cases and to produce and to evaluate the performance metrics. A1

27、2listing of the CD-ROM contents appears in Annex A.131.1 Scope14This document specifies the procedures to be employed to verify that implementations of VAD15processing in conjunction with the TIA/EIA-136-410 DTX/CNG feature meet strict minimum16performance requirements.17This document describes the

28、test database, a method for generating individual test data files, the test18requirements on candidate VADs, a method to collect performance metrics, and a means to test those19metrics so as to validate conformance to the specification.20The purpose of this standard is to define minimum performance

29、levels for the VAD; however, the21manufacturer should attempt to provide the highest possible level of performance.221.2 Definitions23dBov. Sound level in decibels with respect to 16-bit overload (the maximum value in a 16-bit word,2432768).25Frame. A 160-sample data block corresponding to the lengt

30、h of the TIA/EIA-136-410 processing26block.27Mark. Refers to the indicator which designates a frame in a particular data file as either speech or28non-speech.29Modified-IRS. Modified Intermediate Reference System 1 refers to the characteristic spectral30shaping of speech signals by the telephone net

31、work.31Normalize. The scaling of speech, noise or speech + noise data with respect to 16-bit overload.32Offset. Refers to the few frames of speech at the end of an utterance. For the purpose of this33standard, this shall be defined as three frames.34Onset. Refers to the few frames of speech at the b

32、eginning of an utterance. For the purpose of this35standard, this shall be defined as three frames.36TIA/EIA-136-2502Pclip. The probability that speech frames are incorrectly designated as non-speech by the VAD.1For the purpose of this standard, probabilities shall be listed as fractions such that 0

33、 1 Pclip .2Pdetect. The probability that speech frames are correctly designated as speech by the VAD. For3the purpose of this standard, probabilities shall be listed as fractions such that 0 1 Pdetect .4PCM. Pulse-Coded Modulation refers to speech, noise, or speech + noise data whose sample5amplitud

34、es are stored linearly in 16-bit, twos complement words.6SNR. Active speech level with respect to noise level expressed in decibels.7Steady-State, or Steady. Refers to the speech frames between onsets and offsets.8VAD. Voice-Activity Detector, or voice-activity detection refers to the process or alg

35、orithm that9labels data frames as speech or non-speech.10VAF. Voice-Activity Factor is the fraction of time that a VAD designates frames as speech.11DVAF. Delta Voice-Activity Factor is the difference between the candidate VAF and the true VAF12divided by the true VAF, where the true VAF is determin

36、ed from a perfect VAD which correctly13calls all speech frames as speech and all non-speech frames as non-speech.14Vocoder. A device for the coding of a signal into a digital representation152. VAD Minimum Standards16The candidate VAD shall meet objective delay and performance requirements. There ar

37、e no17subjective performance requirements.182.1 Delay Requirements19Whether the VAD is a stand-alone pre-processor or is integrated with the TIA/EIA-136-41020vocoder, the algorithmic delay, which includes the buffering and look-ahead delays, for the VAD-21plus-vocoder combination shall not exceed th

38、at of the TIA/EIA-136-410 vocoder by itself.222.2 Performance Requirements23When operational, the VAD shall be used in conjunction with the TIA/EIA-136-410 enhanced full-24rate vocoder and comfort-noise generation to allow for variable-rate operation. Variable-rate25operation of the vocoder will deg

39、rade the quality of the transmitted speech as the number of speech26frames that are misclassified as non-speech increases. Likewise, if a VAD classifies more non-speech27frames as speech, the advantage of the DTX/CNG feature in reducing the mobile station transmitter28duty cycle is not fully realize

40、d. To ensure a maximum level of voice-quality during VAD operation,29strict limits on the number and type of speech frames that are allowed to be misclassified as non-30speech shall be met. To ensure that the DTX/CNG feature also realizes sufficient power savings,31strict limits are placed on the de

41、viation from the true voice-activity factor.32The objective performance requirements are based on statistics from processing test cases consisting33of conversational test files which have been determined to contain sufficient distributions of speech34utterance types to adequately exercise the VAD al

42、gorithm.35TIA/EIA-136-25032.2.1 Test Cases1Test cases have been chosen to exercise the range of VAD processing. Speech data files are2presented to the VAD at low, nominal and high normalization levels, with different types and SNR3levels of background noise having both stationary and non-stationary

43、characteristics.4Ten speech data files and four noise data files are combined in various ratios according to Table 2.2.1-51 to yield 80 test cases. Normalization values are in units of dBov and SNR values are in units of dB6relative to the speech ( dB refers to the no-noise case). For example, Case

44、1. consists of speech file7m1left1.nom normalized to -36 dBov combined with noise file train.nom added for a SNR8of 6 dB.9Table 2.2.1-1 VAD Test File Cases10Case Speech file Normalization Noise file SNR1. m1left1.nom -36 Train.nom 62. m1left1.nom -26 Garage.nom 123. m1left1.nom -16 Street.nom 184. m

45、1left1.nom -26 - 5. m1left1.nom -16 - 6. m1left1.nom -36 Car.nom 187. m1left1.nom -26 Train.nom 128. m1left1.nom -16 Garage.nom 69. m2right1.nom -26 Garage.nom 610. m2right1.nom -26 Street.nom 1211. m2right1.nom -16 Car.nom 1812. m2right1.nom -36 - 13. m2right1.nom -26 - 14. m2right1.nom -36 Train.n

46、om 1815. m2right1.nom -36 Garage.nom 1216. m2right1.nom -16 Street.nom 617. m3left2.nom -36 Street.nom 618. m3left2.nom -26 Car.nom 1219. m3left2.nom -16 Train.nom 1820. m3left2.nom -16 - 21. m3left2.nom -36 - 22. m3left2.nom -26 Garage.nom 1823. m3left2.nom -36 Street.nom 1224. m3left2.nom -16 Car.

47、nom 625. m4right2.nom -26 Car.nom 626. m4right2.nom -36 Train.nom 1227. m4right2.nom -36 Garage.nom 1828. m4right2.nom -26 - 29. m4right2.nom -16 - 30. m4right2.nom -26 Street.nom 18TIA/EIA-136-250431. m4right2.nom -16 Car.nom 1232. m4right2.nom -16 Train.nom 633. m5left3.nom -26 Train.nom 634. m5le

48、ft3.nom -16 Garage.nom 1235. m5left3.nom -36 Street.nom 1836. m5left3.nom -36 - 37. m5left3.nom -26 - 38. m5left3.nom -26 Car.nom 1839. m5left3.nom -16 Train.nom 1240. m5left3.nom -36 Garage.nom 641. f1right3.nom -36 Street.nom 642. f1right3.nom -26 Car.nom 1243. f1right3.nom -16 Train.nom 1844. f1r

49、ight3.nom -26 - 45. f1right3.nom -16 - 46. f1right3.nom -36 Garage.nom 1847. f1right3.nom -26 Street.nom 1248. f1right3.nom -16 Car.nom 649. f2left4.nom -26 Car.nom 650. f2left4.nom -26 Train.nom 1251. f2left4.nom -16 Garage.nom 1852. f2left4.nom -36 - 53. f2left4.nom -26 - 54. f2left4.nom -36 Street.nom 1855. f2left4.nom -36 Car.nom 1256. f2left4.nom -16 Train.nom 657. f3right4.nom -36 Train.nom 658. f3right4.nom -26 Garage.nom 1259. f3right4.nom -16 Street.nom 1860. f3right4.nom -16 - 61. f3right4.nom -