3 Comparison of the P.57 type 3.4 simulator to human ears

Transcription

1 INTERNATIONAL TELECOMMUNICATION UNION TELECOMMUNICATION STANDARDIZATION SECTOR STUDY PERIOD COM E Febr Original: English Questions: 6/12, 14/12, SQEG STUDY GROUP 12 - CONTRIBUTION SOURCE*: TITLE: Federal Republic of Germany Proposal for an annex to P.64 concerning handset positions to be used with HATS according to P.58 and pinna simulator according to P.57 type Summary In this contribution a new annex D to P.64 ist proposed describing the position of handsets measured with HATS and P.57 type 3.4 pinna simulator. Background information is gien on the type 3.4 pinna simulator, especially to comparison measurements with human ears and the influence of the angles defining the orientation of the handset. A short summary of the round robin tests on reproducibility of measurements performed with the type 3.4 pinna simulator is gien also. In conclusion, handset positions are defined. Measurement results achieed in the proposed positions are added. At the end of this contribution a draft of the proposed annex D is added. 2 Introduction A lot of contributions concerning the P.57 type 3.4 simulator hae been submitted since 1993 [1], [3] - [9], [11], presenting the pinna simulator itself [1], [7], numerous measurement results [3], useful handset positions [6], results of reproducibility tests [4], [5], compatibility tests to P.57 type 1 [5], [8] and type 3.2 pinna simulator [5] and first experiences in possible applications, especially measurements of cordless and mobile telephones [9]. In conclusion from all that data in this contribution an Annex D to P.64 is proposed defining the measurement setup, especially the handset position and orientation for HATS measurements using the P.57 type 3.4 artificial ear analog to the definition gien in Annex C for testhead measurements [10]. 3 Comparison of the P.57 type 3.4 simulator to human ears Comparison measurements were presented already in 1993 [1]. In the present contribution the results are summarized. Fig. 1 shows receiing sensitiity transfer functions of four different telephones applied to the artificial ear in comparison to aeraged transfer functions of the same telephones applied to the ears of six male test persons. Contact person: H.W. Gierlich, HEAD acoustics Tel.: / DOC Fax: /57799

2 For measurements in human ears a small electret microphone was inserted into the ear channel at a depth of 4mm. The influence of the microphone position was corrected as described in [1].The force applied to the handset was adjusted to 13, 4 and 1N. Since the human ear does not behae like a spring, forces below 8N do not result in well-defined positions. Thus, for low forces the position was defined by the geometrical distance to the position corresponding to a force of 8N. Fig. 1: Upper left german standard handset no.''7'', upper right ''Kiel'', lower left ''Nizza'', lower right ''Piccolo''. Cure parameter: force resp. distance, referred to the ''8N-position'', from top to bottom: 13N, 4 and 8 mm(- -) resp. 13N, 4N (Piccolo: 8N) und 1N (Piccolo: 4N) (- --). Fig. 1 shows a good correspondence between the frequency responses measured by the artificial ear and the aeraged frequency responses measured at human ears, see also [1]. The resonances at about 3 khz and 5 khz show a large interindiidual ariance, so they are more attenuated at the human ears due to the aeraging process, while the artificial ear shows a typical resonance. 2

3 4 Results of the round robin test conducted in 1994 / 1995 A detailed description of the test procedure and the results can be found in [5]. Main goal of the procedure was to test the reproducibility of the type 3.4 simulator and to compare the results to type 3.2 and type 1 simulators. The tests which were performed in 1994/1995 in 10 laboratories proed the good reproducibility of measurements performed with the type 3.4 pinna simulator. The standard deiations of the receie loudness ratings (RLR) in db achieed in trials are listed in the following table 1 (from [5]). Table 1: Standard deiations of the receie loudness ratings (RLR) in db P.57 Type 3.4 P.57 P.57 Tel. Name 2N 4N 6N 8N 13N Type 1 Type SIEM SIEM SIEM SIEM DBP-T BT BT BT BT Ag For an pressure force of 13 N the RLR alues are comparable to those measured by the type 3.2 simulator (low leakage option). 5 Influence of angles that determine the handset orientation Fig. 2 shows a typical positioning system indicating all angles and ectors which may be useful for the definition of the position and orientation of the handset in relation to the HATS. A similar system was used in the round robin test [4], [5]. The receiing sensitiity frequency response is mainly determined by the pressure force and the angle α (typical 10,2 ). The angles β (typical 24 ) and γ (typical 0,5 ) hae a minor impact on the receiing direction, as shown for γ in the following section. For the sending direction, howeer, β and α are the most important parameters. 3

4 Fig. 2: Schematic diagram of a typical positioning system for the definition of angles and ectors. 5.1 Inestigations on influence of angle γ The influence of the angle γ has been inestigated for a group of 8 subjects by the following setup [1]: Using a microphone positioned within the ear canal [2] the receiing sensitiity frequency responses were measured for two different positioning instructions: (1) The subject is told to press the handset against the pinna to get it optimum sealed. In that position the frequency response H sealed (f) is measured. Afterwards the acoustical leakage between handset and ear is increased by increasing γ, keeping the upper edge of the handset close to the ear, until a difference of 12 db as compared to the sealed position is obtained for a frequency of 1 khz. The frequency response H tilt (f) is measured in that position. (2) The subject is told to press the handset against the pinna to get it optimum sealed. In that position the frequency response H sealed (f) is measured. Afterwards the acoustical leakage between handset and ear is increased by parallel moement, keeping γ, until a difference of 12 db to the sealed position is obtained for a frequency of 1 khz. The frequency response H parallel (f) is measured in that position. Fig. 3 shows the aerage difference H tilt (f) - H parallel (f) for 5 different handsets (aerage of 8 subjects). Up to 5 khz no significant differences occur. 3 of 8 subjects show a resonance shift at approx. 2 khz (maximum difference see fig. 4 right), all other subjects are comparable to the example shown in fig. 4 left). Listening tests with untrained subjects show no perceptual differences, if speech signals filtered according to the frequency response shown in fig. 4 (right) are presented ia headphones 1. Trained subjects are able to notice slight differences in sharpness and presence in plosie phonems, but only after repeated presentation. In conclusion can be stated that no significant influence of γ exists. Only the leakage dimensions itself determine the receiing sensitiity frequency response. 1 Noise stimuli filtered according to the frequency response shown in fig. xy (right) show audible differences. 4

5 db a (f) - a (f) HO HO A K 7er Nizza Dallas Kiel Piccolo f / khz Fig.3 : Difference H tilt (f)-h parallel (f) for fie handsets (aerage of 8 subjects) according to [1]. Fig. 4 : Examples of H tilt (f) (K) and H parallel (f) (A) according to [1]. 5

6 6 Definition of handset positions During the Round Robin Tests a handset position was used which was adapted as much as possible to LRGP. The problem with that position is that the handset relatie to the mouth is tilted to much downwards. The angle α which is 24 for the HATS is set to 39 in LRGP-position. The reason for this difference is that HATS is always oriented in the "Frankfurter horizontal plane" whereas for LRGP a tilting of the head of about 15 is assumed. Using the same angle for HATS would result in a much lower sending sensitiity for the handsets due to the distance between mouth and handset microphone. So a different position was defined in order to get the typical leakage effect which is shown in the results of the Round Robin experiment [5] a sending sensitiity comparable to the LRGP-position. [10] This new position should be the standard one and is defined as follows: 6.1 Standard position The handset position has been defined according to the procedure described in Annex C to P.64 [10]. The orientation of the handset is defined by ectors normal to the plane of the ear cap and the plane of symmetry of the handset. 1. Unit ector normal to the plane of the ear cap: n EC = ± ( , , ) 2. Unit ector normal to plane of the symmetry of the handset: n HS = ± ( , , ) The relatie position between EEP and plain of lips are defined in ITU-T P.58. The center of the earcap is shifted from the EEP by +11,5 mm in x-direction and +8,0 mm in z-direction. 6.2 Alternatie position For ery flat handsets, especially for mobile phones this position may be impossible, since the handset gets in touch with the "check" of HATS (as well as with the check of subjects) outside the pinna area (see also [9]). This can be aoided by decreasing α by 5. The ectors for that alternatie position are: 1. Unit ector normal to the plane of the ear cap: n EC = ± ( , , ) 2. Unit ector normal to plane of the symmetry of the handset: n HS = ± ( , , ) 6

7 Fig. 5: Definition of unit ectors n EC and n HS relatie to handset (left) and in connection to HATS (right). In addition the ear canal entrance point (EEP), the center of the earcap (CE) and the mouth reference point (MRP) are indicated. 7 Measurement results for the standard position 7.1 Loudness ratings For 5 handsets used in the Round Robin test the sending sensitiity frequency responses were measured in the proposed standard position. The sending loudness ratings were calculated according to ITU-T P. 79 (The other sets used in the Round Robin test were either not equipped with a microphone or no appropriate feeding for the microphone was aailable). All measurements were carried out with a pressure force of 13 N for the handset. The measured sensitiity in receiing direction for this position is comparable to the measurements with the type 3.2 ear simulator, low leak ersion [5]. For this new position the loudness ratings listed in table 2 were measured. In comparison the results measured in the LRGP-position using the artificial mouth according to ITU-T P.51 are shown. Table 2: Comparison of SLR Handset SLR P.79, SLR P.79, new HATS- LRGP-Position Position SIEM db 4.1 db -0.2 db DBP-T 4.7 db 3.8 db -0.9 db BT db -0.1 db -0.8 db BT db 2.4 db 0.5 db BT db 2.1 db -0.2 db 7

8 7.2 Frequency responses in sending direction The sending sensitiity frequency response for the new HATS position compared to LRGP-position is shown in the Figs In general it can be seen that the frequency responses for the different handsets are in principle quite similar for the HATS position in comparison to the LRGP position. Howeer, there are some remarkable differences due to reflections between head and handset which are not present when using only the artificial mouth according to P.51. Especially for the handset BT2 (see Fig. 9) it can be seen that in a frequency range of 3 khz the measured frequency response is no longer flat. Obiously a reflection between the ery flat handset shape and the HATS leads to this narrowband attenuation. 7.3 Frequency responses in receiing direction The receiing sensitiity frequency responses for the new proposed HATS position in comparison to the results of the Round Robin Test [5] and the Type 3.2 ear are shown in Figs The sensitiity frequency responses are shown for 2 N and 13 N pressure force. In general it can be seen that the results between the Round Robin Test measurements and the new proposed HATS position are ery similar. Howeer, there is a tendency to a closer sealing of the handset in comparison to the Round Robin Test. Neertheless it can be seen that the results, achieed by the Round Robin Test again are ery similar to the Type 3.2 ear for 13 N pressure force (except handset BT2, which can not be pressed to the HATS in such a way that the handset seriously sealed. The handset touches the "check" of the HATS already with 4 N pressure force.). 7.4 Summary of measurements results The new proposed HATS position allows a ery realistic measurement of handsets in dependence of pressure forces. In general the results for the sealed condition are comparable to the results achieed with the Type 3.2 ear (low leak ersion) and the artificial mouth according to P.51 in LRGP position. Remarkable differences can be found for the sending direction where the shape of the handset interferes with the shape of the HATS. Those results are certainly more realistic than the measurements using only the simplified mouth according to P.51. For the receiing direction a ery realistic pressure force dependent frequency response and loudness rating measurement can be achieed. 8 Measurement results for alternatie position Handset 7 (BT2) used within the round robin procedure has an extremely flat shape. Thus, the handset gets in touch to the HATS outside the pinna area already for low pressure forces. For handsets showing this effect the alternatie position should be applied. Fig. 20 shows measurements in standard position, fig. 21 in the alternatie position. 8

9 References [1] A New Artificial Ear for Telephone Measurements ITU-T SG 12 Q12, December 1993, Genea, COM [2] W. Krebber, Speech Quality of Telephone Handsets. PhD thesis RWTH Aachen 1995 (In German, summary in English). VDI Verlag, Reihe 10, Nr [3] Possible Errors When Measuring With a Simplified Pinna Simulator ITU-T SG12, Q12, December 1994, Genea, Delayed Contr. D.48 [4] Status of the Round Robin Test for the new Proposed Simplified Pinna Simulator ITU-T SG12 Q12, December 1994, Genea, Delayed Contr. D.55 [5] Report on the Round Robin Experiment for Ealuating the Repeatibility of Measurements on the New Simplified Pinna Simulator. ITU-T SG12 Q12, June 1995, COM E [6] Proposal for a Handset Position to be Used With HATS and the Proposed Simplified Pinna-Simulator Type 3.4 ITU-T SG 12, , Genea, Delayed Contr. D.67 [7] Proposal for a new Text Added to P.57 ITU-T SG12, , Genea, Delayed Contr. D.66 [8] DRP-ERP Correction for P.57 Type 3.4 Artificial Ear ITU-T SG 12 Meeting, , Genea, COM E [9] Application of the new P.57 Type 3.4 Ear Simulator With HATS and the new Proposed Handset Position for HATS: Proposal for the Measurement of Mobile/Cordless Telephones ITU-T SG 12 Meeting, , Genea, COM E [10] Determination of sensitiity/frequency characteristics of local telephone systems. ITU-T Recommendation P.64 March 1993 [11] Impedance Measurements of Handsets Used for Ealuating the DRP-ERP Correction Procedure ITU-T SG 12 Meeting, , Genea, Delayed Contr. D.022 9

10 #### Diagramme 6-19 entsprechen 1-14 aus [6] #### Neumessung Bilder 20, 21 Figure 6: Measurement results (sending) for HATS position as proposed in this contribution (standard position) in comparison with results measured in LRGP position. Handset SIEM1 Figure 7: Measurement results (sending) for HATS position as proposed in this contribution (standard position) in comparison with results measured in LRGP position. Handset DBP-T Figure 8: Measurement results (sending) for HATS position as proposed in this contribution (standard position) in comparison with results measured in LRGP position. Handset BT Figure 9: Measurement results (sending) for HATS position as proposed in this contribution (standard position) in comparison with results measured in LRGP position. Handset BT2 Figure 10: Measurement results (sending) for HATS position as proposed in this contribution (standard position) in comparison with results measured in LRGP position. Handset BT3 Figure 11: Handset SIEM1. Figure 12: Handset SIEM2. Figure 13: Handset SIEM3. Figure 14: Handset SIEM4. Figure 15: 10

11 Handset DBP-T. Figure 16: Handset BT1. Figure 17: Handset BT2. Figure 18: Handset BT3. Figure 19: Handset BT4. Figure 20: Comparison of measurement results (receiing) for HATS position as described in this contribution. Standard position ( ) and alternatie position ( ), using pressure forces 2N and 13 N. Figure 21: Comparison of measurement results (sending) for HATS position as described in this contribution. Standard position ( ) and alternatie position ( ) 11

12 9 Proposed Annex D to P. 64 The title should be the following: Definition of handset positions for measuring loudness ratings and frequency responses using the P.57 type 3.4 artificial ear on HATS D.1 This annex describes two handset positions which should be used to measure the sensitiities of commercial telephone sets in sending and receiing direction. The first position described in this annex is the standard position compatible to the LRGP position in terms of sending sensitiity. The second one is an alternatie position which can be applied for ery flat (e.g. mobile) handsets which may touch the head outside the pinna area in case the standard position is used. The handset positions are defined according to the procedure described in Annex C. The orientation of the handset is defined by a ector normal to the plane of the ear cap ( n EC ) and a ector normal to the plane of symmetry of the handset ( n HS ), as shown in figure D.1./P.64. Fig. D.1/P.64: Definition of unit ectors n EC and n HS relatie to handset (left) and in connection to HATS (right). In addition the ear canal entrance point (EEP), the center of the earcap (CE) and the mouth reference point (MRP) are indicated. D.2 The standard position is defined by the following ectors: 1. Unit ector normal to the plane of the ear cap: n EC = ± ( , , ) 2. Unit ector normal to plane of the symmetry of the handset: n HS = ± ( , , ) 12

13 The relatie position between EEP and plain of lips are defined in ITU-T P.58. The center of the earcap is shifted from the EEP by +11,5 mm in x-direction and +8,0 mm in z-direction. D.3 For some recent hadsets, especially for handies this position may be impossible, since α may be too large. For that case an alternatie position can be used, with α decreased by 5. The ectors for that second position is 1. Unit ector normal to the plane of the ear cap: n EC = ± ( , , ) 2. Unit ector normal to plane of the symmetry of the handset: n HS = ± ( , , ) The relatie position between EEP and plain of lips are defined in ITU-T P.58. The center of the earcap is shifted from the EEP by +11,5 mm in x-direction and +8,0 mm in z-direction, thus the alternatie position show s the same displacement as the standard position described in D.2 13

14 14