[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [AVT] RTCP Inter-Arrival Jitter
RTP packet delay variation (RPDV) is a matter of
a) traffic source (RTP endsystem) and
b) network.
"CBR" and "VBR" are primarily characteristics of the traffic source (a).
The RFC 3550 RPDV calculation algorithm is (and may) not differentiating
between jitter "introduced" by (a) or (b). It's just observing the
traffic arrival process.
The distinction between "CBR" and "VBR" is rather a concept of a
*traffic descriptor*, particular for each (RTP) application.
E.g., a "burst size" parameter is e.g. a property to describe (a), i.e.
differences between "CBR" and "VBR". Or a "delay variation tolerance"
parameter could be a measure for expected jitter introduced by (b).
Consequently, the measured and reported jitter by the RTP layer may be
interpreted correctly by the RTP served user instance (when having some
knowledge about traffic descriptor parameters).
-Albrecht
> -----Original Message-----
> From: avt-bounces at ietf.org [mailto:avt-bounces at ietf.org] On Behalf Of
> Randell Jesup
> Sent: Dienstag, 12. Februar 2008 21:53
> To: Robert Gendron
> Cc: avt at ietf.org
> Subject: Re: [AVT] RTCP Inter-Arrival Jitter
>
> "Robert Gendron" <robert.gendron at exfo.com> writes:
> > I was looking at the RFC 3550 and in the RTCP
RR
> >messages there is a requirement to report the Inter-Arrival Jitter.
It
> >is not clear to me that the suggested calculations herein are
applicable
> >to Constant Bit Rate (CBR) or Variable Bit Rate (VBR) streams. In the
> >case of IPTV, depending on the deployment the two types of streams
can
> >be transported. Since the calculations in RFC 3550 resembles the ones
> >presented in RFC 3393 (IP Packet Delay Variation Metric for IP
> >Performance Metrics (IPPM)) which are prescribed for CBR (Poisson
> >distribution) is there any arm is using the RFC 3550 calculation
methods
> >to address these two cases? Could you shed some light on what I might
be
> >missing?
>
> From 6.4.1:
>
> The jitter calculation MUST conform to the formula specified
here
> in order to allow profile-independent monitors to make valid
> interpretations of reports coming from different
implementations.
>
> The calulations aren't specifically targeted at one or the other,
though
> you could say that the calculation has an inherent (and hard to avoid)
> assumption that packet sizes are the same (or close to); i.e.
packet-level
> CBR. (Note that "CBR" often means in practice "constant over time
period
> X", not "constant bits per packet", especially in video.)
>
> If you want *any* other calculation for jitter, see extended RTCP
blocks
> (RTCP-XR, RTCP-HR and the like).
>
> --
> Randell Jesup
> rjesup at wgate.com
> "The fetters imposed on liberty at home have ever been forged out of
the
> weapons
> provided for defence against real, pretended, or imaginary dangers
from
> abroad."
> - James Madison, 4th US president (1751-1836)
>
> _______________________________________________
> Audio/Video Transport Working Group
> avt at ietf.org
> http://www.ietf.org/mailman/listinfo/avt
_______________________________________________
Audio/Video Transport Working Group
avt at ietf.org
http://www.ietf.org/mailman/listinfo/avt