Voice Signal Level in H323 Terminal/Gateway
Hi,
I'd like to start a discussion on how to set speech signal level in an H323 environment.
Speech signal level is an important part of QoS. It directly impacts end user's perception of how good the system works. It is also an integral part of interoperability testing. After the interoperability issues at the protocol level have been ironed out, the signal levels from different vendors have to "match" in order to provide true interoperability for end users.
Some vendors make speech signal level a configurable parameter in their equipment and leave it to end user to decide. However, not all end users are capable of making intelligent decisions on this issue. When signal levels are set improperly, end users will complain about low volume, high volume, and/or echo. If the system consists of equipment from multiple vendors or involves interworking with PSTN/PBX, things can become complicated quickly.
So a set of guidelines on how to determine speech level can benefit equipment manufacturers as well as their customers, including end users and service providers.
I envision the guidelines should cover a number of topics:
1. desired end-end signal attenuation and acceptable range; 2. desired place of attenuation (tx or rx?) and how much; 3. desired level for DTMF and tones 4. desired level of echo cancellation; 5. connection scenarios: terminal-terminal, terminal-gateway, and gateway-gateway.
Issues that we have to contend with include: end user acceptance, voice coder and echo cancellor optimal operation ranges, idle noise levels.
For many of these issues, we can draw upon the wealth body of knowledge from the telephony world. Some of them may only take a reference to an existing standard. Others may require some serious thinking and, more importantly, agreement among multiple vendors.
The discussion certainly is not restricted to voice only applications. But I am not sure if video has an interoperability issue as serious as voice. I am open to any suggestions. We may continue the discussion online or we can discuss face to face at SuperOp next week.
Now here is my question #1: Is SourceInfo/DestinationInfo structure in H225 call setup messages a reliable way to distinguish end point type (terminal, gateway, gk)? What if in a GK-routed call? Does the GK modify that field?
Thanks,
Shan Lu
NexTone Communications, Inc 9700 Great Seneca Highway Rockville, MD 20850 +1 240-453-6315
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ For help on this mail list, send "HELP ITU-SG16" in a message to listserv@mailbag.intel.com
At 14:14 -0400 00/06/30, Shan Lu wrote:
I'd like to start a discussion on how to set speech signal level in an H323 environment.
For the H.320 system, a similar situation arose. After the initial version of H.320 had been approved in December 1990, audio level setting problem was raised and specifications were added in the March 1996 revision as section 4.2.1 referring to P-series Recommendations. For your convenience an excerpt is attached at the end of this message.
I hope this is also applicable to the H.323 system or becomes at least a good starting point.
Best regards,
Sakae OKUBO *********************************************************** Waseda Research Center Telecommunications Advancement Organization of Japan (TAO) 5th Floor, Nishi-Waseda Bldg. 1-21-1 Nishi-Waseda, Shinjuku-ku, Tokyo 169-0051 Japan Tel: +81 3 5286 3830 (to be transferred) +81 3 3204 8194 (direct) Fax: +81 3 5287 7287 e-mail: okubo@giti.or.jp ***********************************************************
Excerpt from H.320 (this has been converted to text from Word file)
4 Terminal requirements
4.1 Environments Under study.
4.2 Audio and video arrangements
4.2.1 Audio arrangements
A terminal can have one or more of three different arrangements:
- handset function; - handsfree function for a small group of users (up to three users); - handsfree function for more than three users (conference terminal).
The audio characteristics are defined for each of these functions. Furthermore, the bandwidth of the transmitted speech is taken into consideration.
The principles used are identical with those for telephony terminals. That is, the sensitivity for handset function and handsfree function designed for personal use or for a small group of users is specified in loudness ratings, and the sensitivity for conference terminals is specified as output levels.
4.2.1.1 Test principles
4.2.1.1.1 Handset function
The sensitivity measurement of a terminal when a handset is used shall be based on the principles described in Recommendation P.64. The loudness rating shall be calculated as described in Recommendation P.79.
4.2.1.1.2 Handsfree function for a small group of users
The sensitivity measurement of the handsfree function of a terminal designed for a small group of users shall be based on the principles described in Recommendation P.34. The applied test signal level at the digital input when measuring receive sensitivity shall be -30 dBm0.
The user position for a visual telephone terminal depends on the design of the terminal. The real user position as recommended by the supplier might be different compared with the position used for measurements. A correction factor shall be used. The correction factor is:
F(dB) = 20 log10 {Ds/D0}
where Ds is the distance between the recommended user position and the terminal and D0 is the reference distance of 50 cm.
The loudness rating shall be calculated as described in Recommendation P.79.
4.2.1.1.3 Handsfree function for a conference terminal
The principles described in Recommendation P.30 shall be used.
4.2.1.2 Sensitivity
4.2.1.2.1 General
For handset terminals and handsfree terminals designed for a small group of users, the sensitivity shall be specified as loudness ratings: Send Loudness Rating (SLR) and Receive Loudness Rating (RLR). The definitions of SLR and RLR are found in Recommendation P.10.
For conference terminals, the sensitivity shall be specified in terms of input and output levels.
4.2.1.2.2 Receive volume control
For handsfree and loudspeaking terminals, a volume control shall be provided.
Where a manual receive volume control is provided, the minimum control range shall be to -15 dB from the test position. Where an automatic receive volume control is provided, the RLR value obtained with a line level of -15 dBm0 shall not exceed that RLR value which is obtained with a line level of -30 dBm0 by more than 15 dB.
4.2.1.2.3 Handset function
The requirements of Table 4 shall be met.
Table 4/H.320 - Sensitivity of the handset function 3.1 kHz bandwidth(Note) 7 kHz bandwidth SLR 8 8 RLR 2 7
NOTE - 3.1 kHz bandwidth includes both G.711 and G.728 coding.
The manufacturing tolerances are +/-3 dB.
4.2.1.2.4 Handsfree function
The requirements of Table 5 shall be met.
Table 5/H.320 - Sensitivity of the handsfree function 3.1 kHz bandwidth(Note) 7 kHz bandwidth SLR 13 - F 13 - F RLR -7 - F -5 - F
NOTE - 3.1 kHz bandwidth includes both G.711 and G.728 coding.
The receive RLR requirement shall be met when the receive volume control is in its maximum position. The manufacturing tolerances are +/- 4 dB.
4.2.1.2.5 Conference terminals
The procedures and values specified in Recommendation P.30 shall be used.
4.2.2 Video arrangements
Under study.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ For help on this mail list, send "HELP ITU-SG16" in a message to listserv@mailbag.intel.com
participants (2)
-
Sakae OKUBO -
Shan Lu