8.Amendments to Recommendation G.711 (Fascicle III.3 of the Red Book)

8.1    Delete Note 2 to Tables 3 and 4 and the associated small table.

8.2    Add a fourth paragraph to Note 3*:

       "The A-æ-A transparency for bits 1 to 7 was achieved by  modifying  the
table slightly from the optimum conversion in that æ-80 is converted  to  A-81
instead of A-80, and A-80 is converted to æ-79 instead of æ-80.  This  has  an
insignificant effect on quantizing distortion".

8.3    Add a new section 3.6 as follows:

"3.6   Conversion to and from uniform PCM

       Every "decision value" and "quanitized value" of the A  (resp.  æ)  law
should be associated  with  a  "uniform  PCM  value".  (For  a  definition  of
"decision value" and "quantized value", see Recommendation G.701 and in particular 
Figure 2/G.701.) This requires the application of a 13 (14) bit uniform PCM code. 
The mapping from A-law PCM, and æ-law PCM, respectively, to the uniform code is 
given by Tables 1/G.711 and 2/G.711. The conversion to A-law or  æ-law  values
from uniform PCM values corresponding to the decision values, is left  to  the
individual   equipment   specification.   One   option   is    described    in
Recommendation G.721,  4.2.7 subblock COMPRESS.".

8.4    Add the following note to clarify the content  of  Tables  1/G.711  and
2/G.711:

"Note 5 - In Tables 1/G.711 and 2/G.711 the values of the uniform code  are  given
in columns 3, 5 and 7."

       Furthermore, at the heads of Tables 1 and 2, the entry in column  7  should
be changed as follows:

                                 "Quantized value
                                (value at decoder
                                   output), Yn"

8.5    Insert a new section 4 and to renumber present section 4 as section 5:

"4     Transmission of character signals

       When character signals are transmitted serially, i.e. consecutively on  one
physical medium, bit No. 1 (polarity bit) is transmitted first and No. 8 (the least 
significant bit) last."

8.6    The following note should be added to new section 5:

Note - The use  of  another  digital  periodic  sequence  representing  a  nominal
reference frequency of 1 020 Hz at a nominal level of -10 dBm0 (preferred value, see 
Recommendation O.XX) or 0 dBm0 is acceptable, provided that the theoretical accuracy of 
that sequence does not differ more than + 0.03 dB from a level of   -10 dBm0 or  0
dBm0 respectively. In accordance with Recommendation O.XX, the specified frequency 
tolerance should be 1 020 Hz + 2 Hz, - 7 Hz.

       If a sequence representing -10 dBm0 is used, the nominal level at the voice 
frequency outputs should be -10 dBm0.