To the IETF AVT working group and ITU SG16:
--> Interested parties need to respond by end of day PST Monday 2/14 <--
The following has come up during IESG discussion of the draft-ietf-avt-tones-06.txt document:
In section 3.11 there are two events (CRd and MRd) where the signals are different for the initiator and the responder. This implies that the entity doing the translation between audio and events need to be aware of the higher layer roles of the communicating peers. Is this a reasonable assumption?
This appears to be a valid concern that we did not realize in the working group review of this document. One would want the gateway that is doing the translation from an event to a tone to not be required to know the higher layer roles.
It would be possible to give two different codepoints to each of these events. I don't know if there has already been enough implementation of this protocol that this would cause a problem, and the purpose of this message is to find out. Given that the modem and fax tones were added fairly late in the development of the draft, I suspect is would not be a problem to add two codepoints in that section (32-47) without changing the other sections.
The author (Henning Schulzrinne) agreed:
This is at least ambiguous, particularly for a "pass-through" gateway that's pretty clueless as to call state. There are effectively four tone sequences
1375/2002 followed by 1150 1529/2225 followed by 1150 1357/2002 followed by 1900 1529/2225 followed by 1900
My suggestion would be to split the Crd and Mrd into Crdi and Crdr for the initiating and responding side. Same for the other pair.
The draft has been revised to make this split, so the codepoints 41-49 are now different. The revised draft has been submitted but is available in the meantime as:
http://www.cs.columbia.edu/~hgs/rtp/drafts/draft-ietf-avt-tones-07.txt http://www.cs.columbia.edu/~hgs/rtp/drafts/draft-ietf-avt-tones-07.ps
I believe that, subject to this change, the draft has been approved by the IESG. If there are no objections from the working group by the end of the day PST on Monday (2/14), then I will give our OK for the -07 draft be sent to the RFC editor. This is a very short response time, but our goal is to get the RFC number assigned before the end of the SG16 meeting on 2/18. To those at the SG16 meeting, please call this issue to the attention of interested parties that might not have seen this message. -- Steve Casner
Steve and all, Thank you for making the observation of the double nature of CRd and MRd. The same cahnges as you have done, I need to do in the Call Type Discrimination package of H.248 discussed now in SG16.
When looking through avt-tones, I saw that the example showing different signals displayed V.21 bits 10 times longer than they usually are.
This is the table from avt-tones-07
------------------------------Table with errors--------------------------------------- Tone name frequency on period off period ______________________________________________________ CNG 1100 0.5 3.0 V.25 CT 1300 0.5 2.0 CED 2100 3.3 -- ANS 2100 3.3 -- ANSam 2100*15 3.3 -- V.21 "0" bit, ch. 1 1180 0.033 V.21 "1" bit, ch. 1 980 0.033 V.21 "0" bit, ch. 2 1850 0.033 V.21_"1"_bit,_ch._2________1650______0.033____________ ITU dial tone 425 -- -- U.S. dial tone 350+440 -- -- ______________________________________________________ ITU ringing tone 425 0.67--1.5 3--5 U.S._ringing_tone_______440+480________2.0_________4.0 ITU busy tone 425 U.S. busy tone 480+620 0.5 0.5 ______________________________________________________ ITU congestion tone 425 U.S. congestion tone 480+620 0.25 0.25
Table 7: Examples of telephony tones ---------------------------------------------------------------------------- ------------------ If the V.21 bits are aiming at being used for 300 bit/s V.21, that is the most common case, they should be specified to have a length of 0.0033 or rather 0.00333 seconds if you can support that precision.
Agree?
----- Comment on your original topic - the V.8 bis signals. In your message, you have 1375 Hz and 1357 Hz. They should both be the same and be 1375. That is also correct in avt-tones-07, so - no problem.
About timing of these tones, V.8 bis is a bit fuzzy,saying that 400 ms is preferred but for some cases, not to cause interworking problems with existing devices it can be reduced to 280 ms. I do not know what is behind that, and do not understand how the DCE can select between these two cases, so I hope that your single timing of 400 ms is OK. Maybe Bruce Adams, rapporteur of Q4/16 and in charge of V.8 bis maintenance can comment on that. -----
Regards
Gunnar Hellström
-----Original Message----- From: casner@cisco.com [mailto:casner@cisco.com] Sent: Friday, February 11, 2000 10:40 PM To: rem-conf@es.net; ITU-SG16@mailbag.cps.intel.com Cc: Henning Schulzrinne Subject: URGENT question on draft-ietf-avt-tones-07
To the IETF AVT working group and ITU SG16:
--> Interested parties need to respond by end of day PST Monday 2/14 <--
The following has come up during IESG discussion of the draft-ietf-avt-tones-06.txt document:
In section 3.11 there are two events (CRd and MRd) where the signals are different for the initiator and the responder. This implies that the entity doing the translation between
audio and events
need to be aware of the higher layer roles of the communicating peers. Is this a reasonable assumption?
This appears to be a valid concern that we did not realize in the working group review of this document. One would want the gateway that is doing the translation from an event to a tone to not be required to know the higher layer roles.
It would be possible to give two different codepoints to each of these events. I don't know if there has already been enough implementation of this protocol that this would cause a problem, and the purpose of this message is to find out. Given that the modem and fax tones were added fairly late in the development of the draft, I suspect is would not be a problem to add two codepoints in that section (32-47) without changing the other sections.
The author (Henning Schulzrinne) agreed:
This is at least ambiguous, particularly for a "pass-through" gateway that's pretty clueless as to call state. There are effectively four tone sequences
1375/2002 followed by 1150 1529/2225 followed by 1150 1357/2002 followed by 1900 1529/2225 followed by 1900
My suggestion would be to split the Crd and Mrd into Crdi and Crdr for the initiating and responding side. Same for the other pair.
The draft has been revised to make this split, so the codepoints 41-49 are now different. The revised draft has been submitted but is available in the meantime as:
http://www.cs.columbia.edu/~hgs/rtp/drafts/draft-ietf-avt-tones-07.txt http://www.cs.columbia.edu/~hgs/rtp/drafts/draft-ietf-avt-tones-07.ps
I believe that, subject to this change, the draft has been approved by the IESG. If there are no objections from the working group by the end of the day PST on Monday (2/14), then I will give our OK for the -07 draft be sent to the RFC editor. This is a very short response time, but our goal is to get the RFC number assigned before the end of the SG16 meeting on 2/18. To those at the SG16 meeting, please call this issue to the attention of interested parties that might not have seen this message. -- Steve Casner
______________________________________ Gunnar Hellström Omnitor AB Alsnögatan 7, 4tr S-116 41 Stockholm Sweden
Mob +46 708 204 288 Tel +46 8 556 002 03 Fax +46 8 556 002 06 Video +46 8 556 002 05 Txt (All kinds) +46 8 556 002 05 E-mail gunnar.hellstrom@omnitor.se WWW: http://www.omnitor.se
participants (2)
-
Gunnar Hellstrom -
Stephen Casner