Re: [codec] #5: Mention DTMF in requirements

[stephen botzko <[e-mail]>]

I think quite a few people seemed to support the SHOULD requirement (with
MUST be tested, since all requirements need to be tested). It might be
enough to say there is a rough consensus.

I'd suggest a show of hands (or hum) be used at the next meeting, in order
to get a better sense of the community's view.

I do not favor tabling a requirement for X until we find "a volunteer that
provides an X testing environment."

If there is agreement that the requirement exists (either in MUST or SHOULD
form), then we need to find a way to test it.  The tests in general will not
be exhaustive (for instance, we will certainly not produce MOS scores for
every language under every channel condition), but they need to be broad
enough that we can say we've characterized the performance with some
confidence.

IMHO we will need to specify and create the testing environment for each of
the requirements.  It seems to me unavoidable.

Implementing a test environment for this particular requirement is frankly *
much* easier than most of the testing we will have to do. Raymond Chen's
suggestions seem very reasonable, and would result in an objective automated
test score.

Stephen Botzko



On Sat, Apr 3, 2010 at 1:01 AM, codec issue tracker <[e-mail]>wrote:

> #5: Mention DTMF in requirements
>
> ------------------------------------+---------------------------------------
>  Reporter:  hoene@…                 |       Owner:
>     Type:  defect                  |      Status:  new
>  Priority:  major                   |   Milestone:
> Component:  requirements            |     Version:
>  Severity:  Active WG Document      |    Keywords:
>
> ------------------------------------+---------------------------------------
>
> Comment(by hoene@…):
>
>  No consensus until now.
>  I see do opinions on the mailing list:
>
>  a) I think DTMF should be ignored entirely as far as requirements.  If
>  someone wishes to perform tests, they are welcome to do so and present
>  results, but there is no need to require anything.
>  There are many different machine-to-machine communications systems that
>  have been designed to operate over channels that also carry audio.  We
>  should not attempt to support any of them.  The only purpose of this codec
>  should be to carry sounds to a human ear.
>
>  b) DTMF support is important to ensure interoperability and reliable
>  operation. Thus, the codec SHOULD support the transmission DTMF at most
>  transmission conditions. The DTMF transmission performance MUST be tested
>  and characterized.
>
>  Thus, I suggest to pause the decision until we found "
>
>  PS:
>  It is common consensus that FAX NEED NOT to be support or tested.
>
> --
> Ticket URL: <http://trac.tools.ietf.org/wg/codec/trac/ticket/5#comment:2>
> codec <http://tools.ietf.org/codec/>
>
> _______________________________________________
> codec mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/codec
>