Sp-3-4350. 120-rv3 (to become ansi/tia-470. 120-C) Draft 22a ast ballot Comments with track changes on Telecommunications Telephone Terminal Equipment Transmission Requirements for Analog Speakerphones
Download 0.58 Mb.
|
(Informative) - BibliographyThe following is a list of some generally applicable basic standards that are relevant to the requirements of this Standard. ANSI/TIA-464-C-2002 – Requirements for PBX Switching Equipment. ANSI/TIA-464-C-1-2004 – Requirements for PBX Switching Equipment, Addendum. ANSI/TIA /EIA-810-AB-2000 2006 – Telecommunications –Telephone Terminal Equipment - Transmission Requirements for Narrowband Voice over IP and Voice over PCM Digital Wireline Telephones. ANSI/T1.508-2004 – Loss Plan for Evolving Digital Networks. TIA-912-AB-2004 2009 – Telecommunications - IP Telephony Equipment - Voice Gateway Transmission Requirements. TIA-920.110-A-201002 – Telecommunications –Telephone Terminal Equipment - Transmission Requirements for Wideband Digital Wireline Telephones. ITU-T Recommendation G.101 (2003) – The transmission plan. ITU-T Recommendation P.58 (1996) – Head and Torso Simulator for Telephonometry
This annex is a version of an internal memo on the results for preferred listening levels for handsfree terminals (speakerphones), completed in the BNR (now Nortel) Subjective Assessment Lab in August 1986. This study examined user preference for speakerphone receive output levels in quiet and with several levels of room noise. Recommendations for speakerphone nominal level were given based on user response to specific levels across a range of room noise values. Nortel provided this data in support of the development of TIA-470.120-C, These results were originally contributed to IEEE in 1995 in support of similar work, and they are consistent with TIA and ITU recommendations for the electro-acoustic levels of speakerphones.
A loudspeaker was placed on the table 31.5 cm behind the standard handsfree position, and tilted 10 degrees upwards. Acoustically transparent foam hung, from ceiling to table height, in front of the loudspeaker. A disconnected Northern Telecom Companion 4 terminal was placed in the standard experimental position (40 cm in front of subject). The speech samples were bandpass filtered at 500-3500 Hz to simulate the telephony speech band and equalized (per sentence) for signal level. Speech levels were measured using a B&K sound pressure meter that fed into a BT Speech Voltmeter SV6. Measurements were made using a 60 second sample of recorded speech (same samples and talkers used in the study). Levels were measured from the standard handsfree measuring position, 40 cm in front, and 30 cm above the handsfree terminal. The speaker amplifier was calibrated to produce an unattenuated speech level of 89.5 dB SPL. A programmable attenuator was used to control the speech level heard on each trial.
Subjects sat facing the loudspeaker, with the chair placed at a natural distance from the table. A speech sample consisting of one sentence was played through the loudspeaker. Although speech samples were presented through the loudspeaker, subjects reported that speech appeared to come from the handsfree terminal. Subjects were asked to rate whether speech samples were "Too Soft' , "OK", or "Too Loud", and indicate their response using the button box. Twelve trials were randomly presented at each of the ten speech levels shown in the table below. Three blocks of trials were given to each subject, one at each of the three room noise levels: 35, 45, and 55 dBA. Blocks occurred in random order. The 360 trials took approximately 20 minutes to complete. Results below are based on responses from 20 subjects.
As room noise increases, preferred listening level increases. At 35 dBA the preferred listening level is 70 dB SPL, at 45 dBA it is 73 dB SPL, and at 55 dBA preferred listening level is 76 dB SPL. The rate of increase of preferred listening level with increasing room noise is approximately .3 dB/dB, which is in line with Gardner's (1969) and Ericsson's (1977) results. There appears to be a threshold between 79 and 82 dBSPL where speech is considered too loud, regardless of the room noise level. Compression of the acceptable range of speech levels as room noise increases, as reported by Gardner and others, is not readily apparent in the present results. In all room noise conditions, there are at least three speech levels where %OK is above 80%. Substantial compression may not occur until higher room noise levels, or maybe the compression information is lost when data is averaged.
A good compromise for nominal level appears to be 73 dB SPL, since it gives over 90% acceptance (94,97,95) at all three noise levels. An acceptable range for volume would be 61 to 82 dB SPL; 64% found 61 dB to be too low even in the quiet room, and 55% found 82 dB to be too loud in the noisiest room. The step size should be smaller than 3 dB, since a single 3 dB step can change opinion by over 40%. It should be remembered that the present results were obtained in the large sound room, which is not a very reverberant room. Since sound is absorbed readily in this room, subjects may prefer higher listening levels in this room than in a more reverberant room. This should be taken into account when generalizing these results to other environments.
The following are the acceptance of speech levels for three levels of room noise Table FG.1 – Room noise 35 dBA User Acceptance Levels
Table FG.2 – Room noise 45 dBA User Acceptance Levels
Table FG.3 – Room noise 55 dBA User Acceptance Levels
The following is a list of items for consideration in a future revision of this document.
Speakerphone voice switching performance can be thought of part art and part science with a number of inter-related variables. When the number of variables exceeds seven the design becomes an art. Speakerphone deals with many more than seven and many of them are not independent variables involving background noise, switching speeds, integration times, frequency content, volume controls, voice levels, voice cadence etc. There are a variety of good design solutions that work but juggle the parameters differently. A speakerphone must perform over a wide number of conversation dynamics among a variety of interlocutors. The speaking style, cadence, pitch, loudness, interruption initiation, even politeness of each person comes into play as well as background noise, etc. Applying a few, or even considerably more than a few, test signals to the device doesn’t guarantee that a bad switching design fails, or that a good one passes. The methods described in IEEE Std 269 & IEEE Std 1329 are complicated to implement but probably not complicated enough to do a reliable job. They may be great development or screening tools once the parameters are set for a specific design implementation. At a minimum a correlation test with subjective tests using a variety of subjects should be conducted to verify voice switching and this is outside the scope of this document.
Consideration should be given to replacing the Send Loudness Rating (SLR) and Send Loudness Rating Directionality (SLRD) requirements in 4.4.1.2 and 4.4.1.3 with Send Level and Send Level Directionality (SLD) requirements in a manner analogous to the Receive Level and Receive Level Directionality (RLD) requirements in 4.4.2.1 and 4.4.2.2. The following text has been proposed.
The CPE is expected to encounter talkers with a wide dynamic range of input speech levels. A microphone volume control (manual or automatic or both) may be present providing a means to compensate for this variable input level. Different volume control settings may be used to meet the requirements at the different test loops.
To ensure adequate gain, the CPE shall meet the network level requirements in Table H-1. Table G.1H-1 – Speakerphone Send Level Limits
NOTES: The test signal input levels were derived as follows:
The SLD values should be within the limits given in G .2Table H-2. Table G.2H-2 – Send Level Directionality Limits.
NOTES:
Editorial Note: The title of Figure 8 will need to be changed from “Speakerphone Send LR Directionality Measurement Setup” to “Speakerphone Send Level Directionality Measurement Setup.” 
NOTES:
Figure G.1 – Speakerphone Send Level Directionality Measurement Setup † The 1979 version of IEEE Std 661 used a free field microphone aligned coaxially with the artificial mouth for its calibration. The revised 1992 version of IEEE Std 269 adopts the ITU method of using a pressure microphone aligned perpendicular to the mouth axis. The latter method very nearly approximates a true free field measurement, whereas the former method indicates a sound pressure that is about 0.5 dB too high. This results in lowering the drive level to the artificial mouth by a corresponding amount. †† ITU-T Supplement No. 19 to the P-Series Recommendations lists slightly different values for these relationships. In particular, it uses +56 for the send relationship, -50 for receive, and +8 for sidetone. Directory: standards standards -> Request for proposal contract form for the transit industry standards -> Indiana Academic Standards Resource Guide Grade 8 United States History – Growth and Development (to 1877) Updated April 2016 standards -> Us history Dictionary Based on the sc state Standards standards -> Integrating Standards Education into the Business School Curriculum standards -> Use of Sprint U301 3G/4g mobile Broadband usb device Installation standards -> International organisation for standardisation organisation internationale de normalisation standards -> Standards of Excellence standards -> Common Core State Standards for Mathematics (ccssm) Download 0.58 Mb. Share with your friends:
|