Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!doc.ic.ac.uk!warwick!zaphod.crihan.fr!univ-lyon1.fr!swidir.switch.ch!scsing.switch.ch!xlink.net!howland.reston.ans.net!pipex!uunet!noc.near.net!analog.com!analog.com!chris
From: chris.cavigioli@analog.com
Subject: Re: G.728 test pattern?
Message-ID: <1993Nov1.182044.19756@analog.com>
Originator: chris@thor
Sender: usenet@analog.com
Organization: Analog Devices, Inc.
References: <2aiqrp$h7j@news.csie.nctu.edu.tw>
Date: Mon, 1 Nov 1993 18:20:44 GMT
Lines: 54

In article <2aiqrp$h7j@news.csie.nctu.edu.tw> cky902@isdn2 () writes:
>Does anyone know where I can get the standard test pattern of G.728?
>
>Kuo-Yu Chen
>E-mail: cky902@isdn2.ccl.itri.org.tw
>Computer and Communication Lab,
>Industrial Technology Research Institue
>Taiwan, ROC

This is available, but allow me to explain why two answers exist:

ANSWER #1
---------
The G.728 Recommendation has been published in Sept 92 by the CCITT and 
is 63 pages long.  CCITT documents can be purchased from the ITU in 
Geneva, from your country's local standards body, and also in the USA
by a company (PPI) which can be reached at 1-800-OMNICOM (USA calls).
This algorithm was implemented using floating point arithmetic.  Digital 
test sequences exist for this algorithm.

ANSWER #2
---------
Since publication of G.728, there has been further work which standardized
implementation of this algorithm on very fast 16-bit DSPs (such as the new 
ADSP-2171 from Analog Devices, etc).  The "fixed point version" of G.728
appears as an "Annex" to the Sept 92 Rec. G.728.  I don't believe that this
annex is currently included in Rec.G.728 from the publishers, but will be
sometime soon.

Source of Annex text and digital test sequences
-----------------------------------------------
Before official publication, Mr. R. Cox (AT&T) and Mr. Hayashi (NTT) are the 
owners of the annex text (in TROFF format) and the digital test sequences
for both the floating point and fixed point versions.  (Actually, both 
implementations use the exact same input sequences, and then there are two
sets of output sequences ... one for the flt pt, and one for the fx pt.

Possible (?) public ftp site for these
--------------------------------------
I have been discussing with R. Cox about the possibility of putting the 
contents of the 6 floppy diskettes he sent to me (incl. troff text and 
digital test sequences) on our public ftp site here at Analog Devices.
However, likewise to R.Cox's own situation, there is concern about
network security here at Analog Devices in having a public ftp site.
  * any guidance from the peanut gallery would be welcome *
In the meantime, once the diskettes (which are in the USA) get to me here
at my office in Munich, I could send you a copy...

Any other interests, please e-mail postal address and e-mail address to:

    chris.cavigioli@analog.com

and I will be able to see the volume of requests to help decide if we'll do
the public ftp.
