Question 10/XII - Speech transmission characteristics for digital handset telephones (continuation of Question 10/XII studied in 1985-88; new wording) Considering (a) that we now have Recommendation P.31 concerning SLR, RLR, STMR, LSTR, to send and receive frequency responses for digital sets, the following Question should be studied: 1. What are the desirable speech transmission characteristics for digital telephone sets? Contributions should address the headings listed hereafter but the headings under 1.1, 1.2, 2.1, 2.2 and 3.1 should only be addressed if a revision of Recommendation P.31 seems necessary. 1. Sending 1.1 Sending frequency characteristics 1.2 Sending loudness rating 1.3 Distortion 1.4 Noise 1.5 Linearity 2. Receiving 2.1 Receiving frequency characteristics 2.2 Receiving loudness rating 2.3 Distortion 2.4 Noise 2.5 Linearity 3. Sidetone 3.1 Sidetone masking rating 3.2 Distortion 4. Telephone acoustic loss 4.1 Echo 4.2 Stability 5. Delay 6. Other 6.1 Discrimination against out-of-band input signals 6.2 Spurious out-of-band signals 2. How can Recommendation P.35 Part B about handset dimensions be improved. Note - Test methods are dealt with under Question 38/XII. Wideband telephony are dealt with under Question 17/XII. ANNEX (to Question 10/XII) Report on digital telephones (study period 1985-1988) 1.Send and Receive LR:s for Digital Sets 1.1Provisional Recommendation from the previous study period In the annex to Question 10/XII 1.3 a provisional recommendation on LR:s for digital sets was given. 1.2 Contributions -COM XII-18 (Special Rapporteur for Question 19/XII) - "On the optimum CRE and LR" The document is summarized and discussed in COM XII-R 4. -COM XII-72 (BT) "Preferred loudness ratings for digital telephones" -COM XII-108 (NTT) "Desirable SLR and RLR values for digital telephone sets" -COM XII-110 (NTT) "Allowable LR for digital telephone sets considering echo effect under the mixed analogue digital environment" These contributions are summarized and discussed in COM XII-R 11. -COM XII-169 (BT) "Loudness ratings for digital telephones" The contribution is summarized and discussed in COM XII-R 16. -COM XII-179 (Norway) "Transmission aspects for digital telephony" -COM XII-R 20 Annex 1 to reply to Q.10 (Norway) "Comments on transmission aspects for digital telephony" (i.e. COM XII-179) -COM XII-R 20 Annex 2 to reply to Question 10 (BT) "Comments on COM XII- 179" The contributions are reported in COM XII-R 20. Working Party XII/2 - Delayed D.52 (ATT) - "Send loudness ratings for digital telephones" The main finding in Working Party XII/2-D.52 was that linear microphone equipped telephones produced an average 2.8 dB higher active speech level on the telephone line as measured with an SV-6 speech voltmeter conforming to Recommendation P.56. Although the study was done for sets having equal objectively measured EARS-based Transmit Objective Loudness Ratings (TOLRs), there is every reason to believe that similar results would hold for carbon and linear sets having equal Send Loudness Ratings (SLRs) based on Recommendation P.79. The implication is that linear microphone equipped telephones should have about 3 dB different (less loud) SLRs in order to produce the same speech levels on the telephone line, and hence the same subjective loudness for the far end listener, as carbon transmitter equipped telephones. This subject is still under study by the EIA, and no official U.S. position has been taken on the matter. But any recommendation concerning the sending level for digital telephones which is derived from grade-of-service models and experiences based on carbon transmitter equipped telephones measured using swept sine waves may want to allow for such differences being taken into account. The findings above may influence the implementation of recommended LR values. More information is needed before more precise conclusions can be drawn. 1.3 Discussion From the previous study period the following ranges were provisionally recommended: nominal SLR within the range 3 to 5 dB nominal RLR within the range 0 to 2 dB These values were mainly based on an investigation by BT showing an optimum for OLR around 5 dB. During this study period several administrations have expressed their concern about transmission degradation from echo. This degradation will decrease with increased OLR. As the optimum around 5 dB is fairly flat the provisionally recommended ranges were changed to the following short term objective: nominal SLR within the range 5 to 11 dB nominal RLR within the range -1 to 5 dB The range has also been increased to allow for compatibility with the existing analogue network because of differences in the transmission plans in different countries. However, as a long term objective the following values have been agreed upon: nominal SRL = +8 dB nominal RLR = +2 dB COM XII-179 is in a form of a standard (NET 5) as an outcome of work within Europe which contains tolerance limits for SRL and RLR. However, it has been agreed upon not to include such tolerance limits in the proposed recommendation. The draft Recommendation (P.31) as agreed upon in the October meeting 1987 (COM XII-R 20) is found in section 5. At the meeting in June 1985 (COM XII-R 4) some problems with LR:s for digital telephones connected to PABX:es were reported. To achieve compatibility with analogue telephones in a mixed analogue and digital network as low LR values as SLR = +5 dB and RLR = -4 dB sometimes had been chosen. The largest difference relative to the present proposed recommendation seems to be at receiving. One way out of the problem, so that the customers can achieve receiving levels as they are used to, may be to allow for a volume control for instance as allowed for in the proposed NET 5 standard (COM XII-179 amended by Annexes 1 and 2 of the report of Q.10 in COM XII-R 20) see 4.2.2.2. 2. Sidetone characteristics for digital sets 2.1 Suggestion from the previous study period In the annex to Question 10/XII 3, a suggestion for STRE was given. 2.2 Contributions -COM XII-115 (BNR) "Sidetone characteristics for digital sets" COM XII-116 (NTT) "Desirable STMR for digital telephone sets" The contributions are summarized and discussed in COM XII-R 11. -COM XII-169 (BT) "Loudness ratings for digital telephones" The contribution is summarized and discussed in COM XII-R 16. -COM XII-179 (Norway) "Transmission aspects for digital telephony" -COM XII-R 20 Annex 1 to reply to Q.10 (Norway) "Comments on transmission aspects for digital telephony" (i.e. COM XII-179) -COM XII-R 20 Annex 2 to reply to Q.10 (BT) "Comments on COM XII-179" -Delayed Contribution D.46 (October meeting 1987) (Special Rapporteur on Question 9/XII) "Sidetone for digital telephony" The contribution is discussed in COM XII-R 20. -COM XII-226 (Sweden) "The effect of delayed sidetone on the overall quality of a telephone connection" (New contribution) The contribution presents the results from a conversation test where the following parameters were varied Overall Echo Loudness Rating (OELR) = 0, 10, 20, 25 and 30 dB STMR = 15 dB and the electrical path disconnected Delay times = 0, 4, 16 and 32 dB The investigation shows: The delayed sidetone is marked by the direct sidetone at STMR = 15 dB when the delayed sidetone is weak (OELR = 30 dB). If the electrical path is disconnected the weak delayed sidetone (OELR = 30 dB) has an obvious influence on the quality at the longest delay, i.e. 32 ms. At OELR = 20 dB the delayed speech is disturbing and increasing with the delay and the disturbing efect is higher with the direct sidetone switched off. At OELR = 10 dB the delayed sidetone is dominating over the direct sidetone and the degradation is considerable already at a 4 ms delay. 2.3 Discussion The question of sidetone is discussed more in detail in connection with Question 9/XII. From our previous study period a suggestion for a STRE value was given. However, it has been agreed that STMR and LSTR are more adequate measures for sidetone effects at talking and listening respectively. A compromise value STMR around 15 dB with respect to both quiet and noisy room conditions was suggested in COM XII-R 11. With respect to the study connected with Question 9/XII and what people are used to with present analogue telephones a nominal range of 10-12 dB was agreed upon at the Working Party XII/2 meeting in October 1987 (COM XII-R 20). (Manufacturing tolerances should not be given.) However, at the Working Party XII/2 meeting in Geneva, January 1988 it was agreed to include also the value +15 dB in the recommended range which means 10-15 dB which takes into account the results of our agreement to include LSTR as a factor. It should also be noted that a certain sidetone level will also mask the talker echo as reported in 3.2.2 above. As suggested in the work with Q.9 it was also agreed at the Working Party XII/2 meeting in October 1987 (COM XII-R 20) that the concept of listener sidetone should also be adapted and that a LSTR > 15 dB should be recommended. The proposed recommendation for STMR and LSTR are included in Recommendation P.31 (see section 5 below). 3. Send and Receive frequency responses 3.1 Suggestion from the previous study period In the annex to Question 10/XII 2 a suggestion for frequency responses was given as reference for further study. 3.2 Contributions -Working Party XII/2-Delayed D.1 (LM Ericsson) (Working Party XII/2 meeting June 1985) "On effect of mains frequency noise and room noise with digital telephones" This contribution was summarized and discussed in COM XII-R 4. -COM XII-54 (NTT) "Preliminary study on the desirable sending sensitivity frequency characteristics for digital telephone sets" This contribution is summarized and discussed in COM XII-R 9. -Working Party XII/2-Temporary Document 19 (Special Rapporteur on Question 10/XII) (Working Party XII/2 meeting September 1986) "Corrections on the frequency response limits in COM XII-1 page 61" ( 2 in the annex to Question 10/XII). This contribution is considered in COM XII-R 11. -COM XII-164 (NTT) "Sending and receiving frequency responses for digital telephone sets" -COM XII-168 (BT) "Desirable sending frequency response characteristics for digital telephones" This contributions are summarized and discussed in COM XII-R 16. -COM XII-179 (Norway) "Transmission aspects for digital telephony" -COM XII-R 20 Annex 1 to reply to Q.10 (Norway) "Comments on transmission aspects for digital telephony" (i.e. COM XII-179) -COM XII-R 20 Annex 2 to reply to Q.10 (BT) "Comments on COM XII-179" -COM XII-229 (Sweden) "Desirable sending frequency response of telephone sets" These contributions are reported in COM XII-R 20. 3.3 Discussion From the previous study period we had a suggestion for frequency characteristic limits as shown in the annex to Q.10. At the meeting in April 1987 (COM XII-R 16) the earlier proposed limits were considered to be too narrow and should be opened up. However, wide limits will give very little guidance of the preferred frequency characteristics. At the October meeting 1987 (COM XII-R 20) it was therefore agreed to recommend frequency responses instead of giving upper and lower limits. Manufacturing tolerances should not be included. COM XII-229 makes a review and comparison of available information. It makes the following conclusion assuming a substantially flat response for the receiving side, for the sending side a recommended range of response curves might cover linear slopes between 0 and +3 dB/octave within the transmission band. The slope +3 dB corresponds to a level difference of about 12 dB between the upper and lower band limit. The same difference is obtained by a response curve which is flat below 1 kHz and raises linearly towards 12 dB at 4 kHz. This type of response curve is expected to give at least as good quality as a constant slope within the whole band. Based on available contributions and the agreement at the last Working Party XII/2 meeting a draft Recommendation for sending and receiving frequency characteristics is given in section 6. 4. Telephone acoustic loss (TAL) and to TAL related echo 4.1 Contributions -COM XII-97 (BT) "Considerations of echo caused by the acoustic loss path of telephone sets" -COM XII-109 (BT) "Measurement of acoustic loss" These contributions are summarized and discussed in COM XII-R 11. -COM XII-165 (Norway) "Assessment of talker echo caused by the acoustic echo-path in a 4-wire telephone set" -COM XII-166 (BT) "Measurements of the acoustic loss of handset telephone sets on human and artificial ears" -COM XII-Delayed D.25 (BT) (Working Party XII/2, April 1987) "Measurement of HAL using a simple simulation" These contributions are summarized and discussed in COM XII-R 16. -COM XII-179 (Norway) "Transmission aspects for digital telephony" -COM XII-Delayed D.38 (BT) (Working Party XII/2 October 1987) "Suggestions for amendments to draft Recommendation on digital telephone testing" -COM XII-Delayed D.39 (BT) (Working Party XII/2 October 1987) "A human head simulator for handset acoustic loss determination" -COM XII-218 (BNR) "Measurement of acoustic echo path loss in handsets" These documents are summarized and discussed in COM XII-R 20. -COM XII-Delayed D.48 (BNR) "Measurements of acoustic echo path loss in handsets" is a revised version of COM XII-218. The revision is that the echo path is also quoted in LR figures. 4.2 Discussion In a 4-wire telephone in a fully digital network the coupling between the earphone and the microphone will be the main and perhaps only path that will contribute with a loss that surpresses echo. Therefore it is important to reach a recommendation on the loss that can be achieved and which should be recommended. This will also determine if extra echo surpression will be needed. In the study contributions have addressed the different sources for coupling: -acoustic coupling -electric coupling -sesmic coupling (in the handset) All of these need consideration when designing a handset. It has been shown that in carefully designed handsets the acoustic coupling will dominate. It has been agreed that the coupling should be measured simulating real use. For this purpose a measuring method has to be defined. A real use simulator will also be needed to facilitate the measurements. Contributions on coupling loss measured on human heads has been reported as well as a suggestion of a simulator. At present some more information is needed to arrive at recommendations but the sugestion in COM XII-Delayed D.39 of -a human head simulator and of -a telephone acoustic loss > 46 dB should be considered in the forthcoming study. It is also suggested that the question of a human head simulator should be studied in connection with Question 12/XII. 5. Others 5.1 Contributions -COM XII-167 (BT) "Suitable characteristics to be specified for digital sets" This contribution is discussed in COM XII-R 16. -COM XII-179 (Norway) "Transmission aspects for digital telephony" -COM XII-R 20 Annex 1 to reply to Q.10 "Comments to transmission aspects for digital telephony" (i.e. COM XII-179) -COM XII-R 20 Annex 2 to reply to Q.10 (BT) "Comments to COM XII- 179" The contributions are reported in COM XII-R 20. 5.2 Discussion Relevant parts of COM XII-167 have been incorporated in the proposed new question in section 7 below. The other three contributions form the present proposal for the NET 5 standard and contain parts related to all the headings given in the proposed new question in section 7 and should therefore be remembered in the forthcoming study of the new question.