Re: [rmcat] Roman Danyliw's Discuss on draft-ietf-rmcat-eval-criteria-12: (with DISCUSS and COMMENT)

[Joerg Ott <[e-mail]>]

Dear Roman,

thanks much for your comments.  I just posted an updated draft (-13) to
address those.  See below.

> ----------------------------------------------------------------------
> DISCUSS:
> ----------------------------------------------------------------------
> 
> Section 5.2.  Per “Sample video test sequences are available at: [xiph-seq] and
> [HEVC-seq].  The following two video streams are the recommended minimum for
> testing: Foreman and FourPeople.”, these test sequences seems underspecified.
> 
> ** Is the “recommended” here intended to be normative?  There is no RFC2119
> boiler plate in this document to guide the parsing of the text.

Recommended is recommended as per English language and it not used as a
verb but as an adjective, indicating a suggestion.

> ** From the text, there wasn’t much precision in where to find these
> recommended videos (Foreman and FourPeople).  At the url pointed to by
> [HEVC-seq], I found the filenames “FourPeople_1280x720_60.yuv” and
> “foreman15_4000.yuv”, is that them?

This is fixed in the rev now, also the link that disappeared is removed.

> ** Is it expected for
> http://www.netlab.tkk.fi/~varun/test_sequences/foreman15_4000.yuv to be 0
> bytes?  I tried on 03/03/2020 at ~0950 EST

The broken link is gone now.  Both videos can be found in a single place
now.  It is clearer specified where to find them.

> ** Give that that one of the recommended urls doesn’t work even before this
> draft is published, I have great reservation with keeping a normative
> “recommended” to such external repositories.  However, providing pointers to
> repositories of “sample video test sequences” makes sense to me and is helpful.

Should be addressed by now.

> ----------------------------------------------------------------------
> COMMENT:
> ----------------------------------------------------------------------
> 
> ** Section 3.1.  The purpose of this log isn’t clear.  What is the relationship
> of it to the metrics described in the previous section?  Where does it fit into
> the measurement workflow?  Is this constructed on a per packet capture file
> basis?

Added a clarifying sentence:

These data should suffice to compute the media-encoding independent
metrics described above.  Use of a common log will allow simplified
post-processing and analysis across different implementations.

> ** References:
> -- Section 3.  Consider adding a citation for tcpdump and wireshark

done

> -- Section 4.4.  Consider adding a citation for the “Bilbert-Elliot” model

We got this comment also from Mirja, to which I replied:

I could include a reference here for Gilbert-Elliot such as:
http://www.ohohlfeld.com/paper/hasslinger_hohlfeld-mmb_2008.pdf

However, this not mandatory text and the second part of the sentence
is even less specific and does (deliberately) not give a specific
example.  We don't want to create any prejudice here beyond random
losses.  The Internet changes, after all.

In order to avoid this being a showstopper here (we don't have enough
hour tonight for another rev round, I am adding this in now.  Please
provide feedback if this is really needed.  I could still remove it
in two weeks when I-D submission reopens.

> ** Editorial:
> -- Section 3.  Recommend explicitly spelling out PCAP as packet capture.

This is spelled out in brackets; since this also refers to a commonly
used file format, I would keep it as is.

> -- Section 4.5. s/is is/is/

Fixed.

Thanks much,
Jörg