Re: [rtcweb] RTP Usage: Is RTP Retransmission REQUIRED or RECOMMENDED

["Roy, Radhika R CIV (US)" <[e-mail]>]

Hi, Fabio:

I could not retrieve the first one, but other two links are OK.

I understand these things, but still they are NOT sufficient enough to generalize these results because of the following:

1. Experiments seem to be local without any precise measurements under all possible circumstances noting the network traffic, audio quality, video quality, and precise applications and number of subjects under test using different applications (e.g. point-to-point audio & video conf, multipoint audio & video conf, streaming audio/video applications). That is, it is not global, that means, there is NOT a wide area network (e.g. public Internet) connected with two different access networks where end-to-end one-way propagation delay is around 120-150 milliseconds, types of applications, traffic loads, subjects who were under test, and others.

2. If we do experiments over the high-speed LANs connecting users in the same LANs, we may find many things are possible including retransmissions of audios/videos. We can then extend these networking considerations unless and until we get a situation where retransmissions are no longer applicable. RFCs with retransmissions of audios/videos may be acceptable considering the trade-offs of end-to-end-delay and jitter-buffer-sizes under those limited circumstances. Each of those precise conditions can be noted generated. Again, these are very specific and case-by-case basis and cannot be generalized. 

3. Video quality measurements need to be taken in terms of real-time point-to-point and multipoint conferencing along with audio and video. Then these experimental results need to be measured separately for both audio and video separately. For example, at what point lip-synchronization is no longer possible.

Unless one can note all of those special cases for retransmissions of audio/video, it will always be technically confusing to use the retransmissions of audio and video as a generalized one.

I also know one case such as wiretapping applications. For silent recording in a distance location, audio & video retransmissions are needed because the exact things that are produced legally in the court need to be near 100% correct.

Hope this helps.

Radhika 


-----Original Message-----
From: Fabio Pietrosanti [mailto:[e-mail]] On Behalf Of Fabio Pietrosanti (naif)
Sent: Wednesday, June 27, 2012 2:23 PM
To: Roy, Radhika R CIV (US)
Cc: Cameron Byrne; [e-mail]
Subject: Re: [rtcweb] RTP Usage: Is RTP Retransmission REQUIRED or RECOMMENDED

On 6/27/12 8:14 PM, Roy, Radhika R CIV (US) wrote:
> Hi, all:
> 
> Real-time protocols (Audio/Video payloads for two-way conversations) over UDP/IP were being developed based on the single thing: Retransmission delays are unacceptable (more over congestion problems especially for the high-bandwidth videos).
> 
> If this fundamental performance characteristic for the Real-time protocols (Audio/Video payloads for two-way conversations) need to be changed, the experimental results with lots of actual measurements must be produced to the technical communities.
> 
> Otherwise, someone may make some subjective statements here and there without any actual measurements will only provide evidences what those folks do not know what they are actually talking about creating technical confusions.

It's not a subjective topic that over a mobile network there is packet loss and that there are known condition where this lead to unacceptable voip quality.

The Internet Measurement of VoIP on different Transport Layer Protocols http://netarchlab.tsinghua.edu.cn/PersonalMainPage/wzl/public_html/pubs/2009_ICOIN_transport_measure.pdf

SipDroid: An opensource Mobile client using RTP Retransmission to improve quality over 3G network with packet loss:
http://code.google.com/p/sipdroid/wiki/NewImprovedAudio

RTP Retransmission RFC (explaining how to handle dynamically retransmission from RTCP feedback):
http://tools.ietf.org/html/rfc4588

Fabio