5196106213

I

Conducted by Neville Williams

# •

• • * • f    •

error correction:

Is quality at risk?

On page 19 of the April issue, I posed a somewhat tongue-in-cheek question which could well be re-phrased thus: Is it conceivable that CD error correction circuitry, while ostensibly concealing certain data faults might, in fact, be translating them into incongruous components which can be perceived as such by keen-eared listeners?

In the particular article, I had madę the point that supporters of the tradi-tional analog system were tending nowadays, to disregard the fact that most modem black discs are sourced from digital master tapcs. Yct many of them maintain their opposition to compact discs on the grounds that the digital technology, on which thcy are bascd, rcnders them discemably “un-musical".

Whatever the term means, it invites some clarification as to why digital tcch-nology should be okay for studio mas-tering but unacceptable for consumer-level products.

With no such explanation forthcom-ing, I offered to invcnt one, bascd on the fact that most modern digital master recorders (e.g. the DASH format) allo-cate two or morę tracks to each chan-nel, with instantaneous automatic track switching in the evcnt of a data fault being encountered.

With tongue planted firmly in cheek, I pointed out that, by contrast, the Cl) system rclies on a single data strcam to scrve both channels It is backed up, howcver, by powerful error correction circuitry, capable of correcting or concealing discontinuities in the data stream that might othcrwise be heard as clicks or plops.

“Could it be", I asked, “that CD technology rclies too heavily on correction circuitry, particularly in the case of discs and/or players with morę than their fair share of data strcam errors?"

Is it possible that, in the process of concealing an excessive number of er-rors, it also conceals (or otherwise pro-cesses) an excessivc quota of high fre-quency signal informatiori — thereby imparting that allegedly incongruous, unmusical quality to the sounds?" (My dictionary tellą me that, amongst other things, incongruous means “not har-monious in character").

Not long tó wait

While strcssing that the suggestion lacked “any known basis in fact", I nevertheless wondered privatcly how long it woiild be before someone else adoptcd the idea or‘, alternativcly, came up with it independently. I didn^łt havc long to wait

In the “Views" page of the May 1987 issue of “Hifi News & Record Review", a correspondcnt (R.Webb, London) had this to say on the matter:

. *7/ seems that if your CD player has a to uch of the “jitters” or any other vibra-ńonal problem of focussing on the daki Information, then the CIRC or error correct ions fabricates • a mean a\ erage of what’s missing.

“If this is happening regularly and every so often in each codę word (as is likely with a wobbly disc or chassis yibration) then Vm not surprised if the sound is “less sweet” with “a lessening of the sound stage” when a disc is used without a Mod Squad Dam per Kit ... sińce it’s the definition of the fas ter fre-quencies that will be liardest hit.

“ . . . manufacturers should look into giving us an indication of whal percent-age of the musie we are hearing is “correct” and what is “madę up” ... How about one of those simulated wice chips,

which waits until youve removed the disc from the player and then States . . . 50% of what you’ve been listening to was guessed by the Computer inside this player!”

R.Webb certainly docsn’t theorise by halves, beginning, as he docs, with pure speculation and ending up with an im-plication — even if consciously exaggcr-ated — that 50% of what we hear from a CD player could be Computer guess-work. I do conccde, however, that ex-aggerated or not, the suggestion has about it a certain credibility and that, somewhere down the track, it will need to be invcstigated and quantificd in relation to typical present-day discs and players.

Published ctaims

While I have no doubt that relevant figures alrcady exist in laboratory notes and limited circulation research papers, up-to-date information is very patchy in-deed as far as consumer lcvcl authors and publications are concerned.

Some suggest that this is no accidcnt: that manufacturers sce little point in un-settling their one reasonably prosperous market area with gratuitous tcchnical debate, it is sufficicnt to maintain that the CD system employs powerful error detection and correction technology, able to ensure a smooth flow of sound, free from distortion and unmarred by dropouts and clicks.

For the morę technically mindcd, it has long sińce been identified as the Cross-Intcrleave Recd-Solomon Codę (CIRC) which, as per the original speci-fications, makes possible “complete correction" of errors spread over 4000 bits — equivalent to a 2.5mm section of pit track. This can take care of:

•    Random errors — involving relatively isolatcd bits, caused in disc production by inaccurate photo resist coating, or by other minutę blemishes, and

•    Burst errors — resulting from visibic

26

ELECTRONICS Australia, September 1987