Audio Engineering Society Standards Committee

Comments on AES3 coaxial revision

[Last printing 3 June 2001] Comments to date on Draft Revised AES-3id-xxxx DRAFT REVISED AES information document for digital audio engineering -- Transmission of AES3 formatted data by unbalanced coaxial cable published 2001-01-08 for comment.

Caine 2001-01-12

From:   SMTP%"" 10-JAN-2001 08:07:14.06
Subj: New fig 3 for AES-3id draft

Julian Dunn has pointed out a minor error in the eye diagram, fig 3, in AES-3id-xxxx. I've made a replacement from a real plot of an AES-3id signal and annotated it as required.


Robin Caine

Dunn reply 2001-01-15

From:   SMTP%"" 15-JAN-2001 12:02:51.38

Dear Mr Caine,

I agree that your figure corrects an error in the figure given in the CFC. The figure in the CFC does not show the eye diagram correctly and would confuse the reader.

The original measurements from which the value for Vmin was derived need to be checked in order to determine if the error affects the quoted value of Vmin, mentioned in note 2 to clause 7.3.

Best regards

Julian Dunn
AES SC-02-02 vice chair.

Tsuchiya 2001-03-06


Subj: Comment on Draft Revised AES-3id


I am Yoshikazu Tsuchiya, Sony Corp. I have a comment on an equation given in the chapter [sub-clause] A.2 of ANNEX.A (page 12).

Though an editorial error of the equation ,T=1/6,144x1/fB==0,16276usec, be corrected , fB is not helpful parameter to understand other equations and figures on the ANNEX. It is better to change the equation to clarify the time unit as below after deleting 1/fB.


Best Regards

Y. Tsuchiya

Dunn reply to Tsuchiya 2001-03-08

From:   SMTP%""  8-MAR-2001 10:29:04.81
To: AES Standards Secretariat

The comment of Tsuchiya-san is editorial. Here are my suggestions for editorial changes on that page:

The parameter, T, that Tsuchiya-san mentions, is equivalent to the unit interval, which is now defined in AES3. The equations in the draft also show confusion of units so I recommend the following changes:

Delete all the equations and the first paragraph on page 12 and replace with:

                                   (1+cos(pi.f.T))     pi.f.T
E(f) = a.(root)f + 20.log(---------------.-----------)
                                   2    sin(pi.f.T)

Where E(f) is in dB, f is in Hz, T is in seconds, and a is in dB per root Hz.

Because the equalizer compensates cable characteristics, it is reasonable to suppose one practical cable type to be used. The parameter T (in seconds) is the duration of the unit interval corresponding to the rate of transmission being used. For a 48kHz frame rate the unit interval is 162.8ns. The parameter a (in dB Hz[-1/2]) is given by the type and length of coaxial cable. For 1000m of 5C2V, a is 0.0085 dB Hz[-1/2] (equivalent to 8.5 dB MHz[-1/2] as provided in table A.1).

Typesetting note: The [] indicates a font change to indicate "raised to the power". Hz[-1/2] can also be "per root Hz" and MHz[-1/2] can be "per root MHz", It is a matter of style.

The second paragraph and the figure on page 12 should remain.


Best regards
Julian Dunn
Vice-chair SC-02-02

For more information about standards activity:


(C) 1999, Audio Engineering Society, Inc.