[R-C] Strawman: Requirements

[Varun Singh]

Hi,

On 20.10.2011, at 8.09, abheek saha <abheek.saha at gmail.com> wrote:

Hi,


Does "better" mean better media quality or more bandwidth, or does "better"

mean "more closely models real-world open-internet performance"?



We got smoother goodput and better PSNR, the average media bit rate

was higher in the case of droptail but the variation in media bit rate

was also much more.


How did you measure psnr in ns2? We have been struggling with this for

some time now in our own simulations?



The encoder and decoder are external to ns2. They are connected by an
interface to the ns2 application. Instead of relying on the system clock the
codecs rely on ns2's clock. Basically ns2 polls the codecs at small
intervals of 1ms to check if they have data to send.

I've mentioned this before: a 300sec simulation can take anything from
30mins to 2hrs. Depending on the complexity of the simulation.

If you are interested this environment has been used here:

Evaluation of Error Resilience Mechanisms for 3G Conversational Video;
http://www.netlab.tkk.fi/~jo/papers/2008-er-3g-conf-ext.pdf
Rate adaptation for conversational 3G Video;
http://ieeexplore.ieee.org/xpls/abs_all.jsp?tp=&arnumber=5072183&tag=1

There is also Evalvid-RA. http://www.item.ntnu.no/~arnelie/Evalvid-RA.htm

If we want to also compare PSNR for the above simulations then I'd prefer to
eventually rub the simulations in the real world using
dummynet/netem/planetlab.


-best regards,

Abheek

_______________________________________________

Rtp-congestion mailing list

Rtp-congestion at alvestrand.no

http://www.alvestrand.no/mailman/listinfo/rtp-congestion
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.alvestrand.no/pipermail/rtp-congestion/attachments/20111020/95948e04/attachment.html>