[ProAud] RE: ProAudio Digest, Vol 7, Issue 23

Tue Apr 19 00:36:45 PDT 2005

Well considering that invalid data is unlikely (easily verified if you 
have the original), jitter is the only thing that's left.

>I would assume that CD players first read the data, then
>correct / validate it, then buffer it, then stream it out
>synched to a stable clock. 

Bingo, you said it yourself. The key-words are "assume" and "a stable 
clock".

Have you ever measured clock jitter on a CD player and how it changes with 
different media? The simplest test is to play a calculated 11.025kHz 
signal and do an FFT using a known-good A/D converter (e.g. use an AP 
set). What you should get is a horizontal noise floor and a single peak at 
11.025kHz. Anything else is jitter. On most CD players the differences 
between and among pressed discs and CDR's are astonishing.

The most common mechanism is also embarrassingly simple. The clock 
oscillator is usually on the very same chip as is the decoder and the 
servo driver. If not, it is still usually running off the same power rail 
as are the decoder/servo. If the disc is unstable, or data is written at 
wobbly intervals, the servos have to speed up / slow down more heavily in 
order to keep the fifo from overflowing or underflowing. This imposes a 
noisy load on the power rail, modulating it. The clock frequency is 
modulated as a consequence.

Just for the heck of it, listen to or measure the power supply that powers 
the clock oscillator circuit, both in play and in stop mode, and with 
different discs.

In other words, your assumption reflects how things should be done (heck 
we all agree on that), but not how they are. The sonic differences between 
CD's and CDR's can be made zero by using a stand-alone, separately powered 
clock circuit on the analogue side and using some other measures to 
prevent crosstalk from the drive to the DAC other than the wanted data. 
There are two reasons why this is not done on the vast majority of 
players. One is ignorance (I'll come to that later). The other is because 
optical drives are sold as standalone building blocks that should be 
usable with the least effort. If you've ever dealt with consumer equipment 
manufs in the World's Most Populous Country, you'll realise that talk of 
them making their own clocks etc is way,yyy,yyy above their heads.

And it's not only cheap players. I've got a Marantz SA-1 (their first SACD 
player), priced at that time at $4k or so. It must've been Very Audiophile 
because half of the box consisted of regulators. One for the digital half 
of the dac chips, two for the analog half, two for the reconstruction 
filter and one for the drive. You've guessed it. The clock oscillator 
lived on the same 5V as the drive. It did have a neat shiny cover over it 
though.
I'll spare you the details of the EMC carnage I found inside. What I mean 
is that even buying high-end players won't guarantee you that it'll play 
copied CD's at the same quality as the original. Only if you know the 
player inside out you can make any statement as to the soundness of the 
engineering that went into it.

__________________________________________________________________________
Bruno Putzeys, Philips Applied Technologies
Versterkerbouwerij - Convertisserie
Interleuvenlaan 80, B-3001 Leuven, Belgium,
Tel +32 16 39 07 45
bruno.putzeys at philips.com

"David Lloyd" <David.Lloyd at sas.com>
Sent by: 
proaudio-bounces at pgm.com
2005-04-18 09:27 PM
Please respond to proaudio

        To:     <proaudio at pgm.com>
        cc:     (bcc: Bruno Putzeys/LEU/PDSL/PHILIPS)
        Subject:        [ProAud] RE: ProAudio Digest, Vol 7, Issue 23
        Classification: 

"My money is on your CD player being sensitive to jitter. Most are."
__________________________________________________________________________
Bruno Putzeys, Philips Applied Technologies Versterkerbouwerij - 
Convertisserie 

============
I don't disagree with the existence of this phenomenon...I've heard 
audible differences between various burns/pressings of the same content 
first-hand.  That said, it makes no sense that this should be possible, 
given the error correction and concealment scheme required by the Red Book 
spec [as described in another post]. 

How, exactly, is a CD player "sensitive to jitter?"  I would assume that 
CD players first read the data, then correct / validate it, then buffer 
it, then stream it out synched to a stable clock.  How can media jitter 
imposed on the read phase affect the inevitable stability of the signal 
streamed out of the buffer in synch with a clean clock?

Does a CD player's jitter sensitivity to low gain/imperfect burns suggest 
the *absence* of internal buffering in the player?  Or...Does the player's 
sample clock somehow derive from each CD placed in the tray?

Any detail on how this works is appreciated,

David

_______________________________________________
ProAudio mailing list
ProAudio at pgm.com
http://www.pgm.com/mailman/listinfo/proaudio