Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 (Tek) 9/26/83; site tekig1.UUCP
Path: utzoo!linus!decvax!ucbvax!ucbcad!tektronix!tekig1!gregr
From: gregr@tekig1.UUCP
Newsgroups: net.audio
Subject: Re: CD Error Correction Questions - (nf)
Message-ID: <1545@tekig1.UUCP>
Date: Sun, 19-Feb-84 14:53:20 EST
Article-I.D.: tekig1.1545
Posted: Sun Feb 19 14:53:20 1984
Date-Received: Sat, 25-Feb-84 03:42:23 EST
References: <237@ihlpf.UUCP>
Organization: Tektronix, Beaverton OR
Lines: 43

Eric, I think I can answer two of your three questions.

>>  1.	Do ALL CD players correct to the fullest extent made possible
	by the coding redundancy or do some only correct errors up to,
	say, 2000 bits? 

Unfortunately this is the one I can't answer.  However the bulk of the
error correction circuit is done in a single IC within the Phillips
system so I should think that most companies with custom IC capability
should be able to duplicate the full correction at a fairly competitive
price.  I would guess that Phillips sells its IC to others that license 
its approach??? 

>>  2.	What is the nature of the interpolation process and what
	degradation (from perfection) of the signal results?

In the Phillips system the process is linear interpolation between the
preceeding value and the following value.  The effect is a very momentary
increase in distortion.  Remember also that this is a relatively rare
occurance (one sample interpolation every 10 hours at a BER (bit error rate)
of 1*E-4).  To really get a feel for the effectiveness and complexity of this
error correction and concealment system you should get a copy of the Phillips
Technical Review, vol 40, #6.  I think it's too technical to continue to
discuss in the newsgroup.

>>  3.	The bit rate of the PCM is obviously 2 * 44.1 * 16 kbps, but what
	is the bit rate as the data comes off of the CD (including
	redundancy)?

The actual bit rate coming off the CD is 4.32 Mbits/sec.  The bit density 
increases in the following manner.  The information is divided into frames
of six 16 bit samples per channel (192 = 2 x 16 x 6 audio bits/frame).  After 
CIRC encoding additional C&D (control and display) bits and parity bits are
added which yields 264 data bits/frame.  The data is then modulated by the
EFM (Eight to Fourteen Modulation) I discussed before, 3 merging bits are added
for each resulting 14 bit pattern,  and 27 sync bits are added for each frame.
This brings the total to 588 bits/frame ( [264 * {14+3} / 8] + 27 ).  Hence the
original audio rate of 1.41 Mbit ( 2 * 44.1 * 16 kbps ) is increased by a factor
of 588/192 = 3.06.  This yields 4.32 Mbits/sec.

					Greg Rogers