NetNews Usenet Archive 1992 #18/NN_1992_18.iso/spool/rec/audio/11713

Internet Message Format | 1992-08-20 | 10.0 KB

Path: sparky!uunet!mcsun!sun4nl!tuegate.tue.nl!viper!marc
From: [e-mail] (Marc Heijligers)
Newsgroups: rec.audio
Subject: Re: Sony Mini-Disc Questions/Debate
Message-ID: <[e-mail]>
Date: 21 Aug 92 07:16:02 GMT
References: <[e-mail]> <[e-mail]>
    <[e-mail]>
Sender: [e-mail]
Organization: Eindhoven University of Technology, Eindhoven, The Netherlands.
Lines: 164
In-reply-to: [e-mail]'s message of 13 Aug 92 09:18:10 GMT


In article <[e-mail]> [e-mail] (S S Sturrock) writes:

   The machines which are being released at this very moment in the UK do have
   SCMS.  This basically means that although you could stack a whole pile of
   DCC machines and copy a CD onto all of them digitally you cannot make
   digital copies of your recordings whether they were of copyrighted material
   or recordings you made of the baby gurgling.  Philips response is that
   because DCC is digital it is perfect (ha ha ha!) SCMS was needed.  What
   worries me about the whole business of PASC or whatever is the ability for
   the algorithm to encode and decode then recode as 4bit to 16bit to 4bit
   etc.  The Philips bod said that he didn't see this as being a problem but
   then I think he has too much faith in the perfection of digital and no idea
   about precisely what PASC is doing to achieve the compression.

I don't think you can argue just because you 'think' he has to much faith in
his product. He has based his opinions on wide-scale tests.

   A review of the Philips DCC 900 states that the analogue sound is about on
   average for a 100 Pounds (UK) cassette deck and as such I would be loathed
   to abandon the deck I already have,

I don't know who did the review, but when I heard DCC myself yesterday, I can
only state that the recorded sound sounded exactly the same as the original
sound did. I think the reviewer is of a type equal to the one who claims that
CD doesn't sounds right, because the sound is cut into pieces.


   not to mention the fact that the sound
   quality of analogue tapes in my car is very good and I don't really care
   to tape things for posterity anyway.  As the review states, it is a pity
   the DCC deck can't record analogue sound on analogue cassettes.  Also,
   since the DCC tapes are nothing special, just ordinary chrome tapes in a
   fancy box, I wonder how long it will be before someone circumvents the 5 to
   7 Pound pricetag of blank DCC tapes by using ordinary analogue chrome tapes
   if you are willing to accept the lack of the fancy sliding cover just as
   has been done with SVHS and Hi8 video?

It is not ordinary chrome tape. The information on the tape is 'digital' (in
fact it is analogue ofcourse), which means that a degradation of the tape can
hardly be noticed (a degraded '1' has to degrade very hard to become a '0').
CD's also degrade. The best would be 'solid-state audio', but this is future
work.

To give you a more complete view on DCC, here's a letter of my friend (working
at Philips) which he wrote because of a discussion on nlnet (= local net for
the Netherlands, on which the DCC is a hot item too).

Have fun,
Marc

-----------------------------------------------------------------------------

Hi there,

thanx to a friend of mine I'm able to forward this message to you, out there,
participating in the discussion about various DCC matters. I'm NOT talking
as an official Philips spokesman; however, I'm involved in digital audio and
audio data reduction, and currently I have a DCC player (the upcoming DCC900)
at home to try it out. I would like to address several points in the, rather
speculative, discussion that's going between you all.

1. audio inputs: DCC players (at least for some time) will have both an analog
and a digital input. The digital input format complies with the IEC958 standard,
commonly called 'AES/EBU interface', since it was approved by these two
organisations (Audio Engineering Society and European Broadcasting Union).
The digital interface that many of you will find on your CD player is also in
AES/EBU format, or better a subset of it, called SPDIF = 'Sony / Philips
Digital Interface'.

2. audio outputs: DCC players (at least for some years) will have both analog
and digital outputs in the AES/EBU format described in 1.

3. sampling rates: when recording from an analog source, the DCC defaults to a
sampling rate of 44.1 kHz (=CD rate). When recording from a digital source,
the DCC switches over to the sampling rate of the digital audio source. All
three AES/EBU sampling rates (32, 44.1, 48 kHz) are supported. When playing
back, the DCC runs at the sampling rate at which the tape was recorded.
The reason for the existence of 3 different sampling rates in 'digital audio
land' is the independent choice of the (default) sampling rate for different
widespread digital audio applications:
  48   kHz: DAT; professional recording equipment; future DAB
  44.1 kHz: CD, CD-I, DCC
  32   kHz: DSR (Digital Satellite Radio); D2MAC television sound
The 32 kHz rate is used since it requires (with the same number of bits per
sample) 33% less data transmission bandwidth than 48 kHz, whereas only 4 to 5
kHz audio bandwidth limitation is the consequence, which musically speaking is
only a minor third at the highest frequency of what very good (i.e. young)
listeners can perceive (plm. 20 kHz). (How many of you, more 'mature', listeners
can still hear the 15.625 kHz tone normally generated by PAL television sets?)
Since bandwidth is expensive, for the applications mentioned above the small
pay-off in sound quality (only noticed by a limited group of listeners) was
considered subject to the significant reduction in transmission bandwidth.

4. Copy protection: DCC uses SCMS (Serial Copy Management System) as
introduced also on DAT. As one of you already correctly described, SCMS lets
you make only one copy of a *digital* source that's copy protected (such as
CD), and as many generations of a *digital* audio source that's not protected.
When playing back a DCC recording of an *analog* source, the digital AES/EBU
output data has no protection, so it can be copied on and on without inhibits
(in the description above the DCC can then be considered to be a digital audio
source without copy protection). You must realise, however, that every time
you make a digital copy, data reduction is applied another time. Since the
reduction is non-bit-true (the bits coming out of the DCC digital output at
playback are not the same as the bits put in the digital input digitally at
recording) a slight degradation is possible. Tests, however, have shown that
only for the first copies this is slighlty noticeable, and after, say, 5
generations, no more degradation is perceived. Copies were made up to more
than 80 (...) generations, but the audio quality remained very high.

5. Sound quality: the PASC (Precision Adaptive Subband Coding) scheme used for
DCC has shown to be very high quality. Extensive tests have been done by large
groups of expert listeners (audio engineers, recording engineers, musicians,
HiFi magazine journalists) with selected material that's known to be
critical for audio data reduction. The results show only tendencies, if any.
Our experience is that people who claim to hear 'obvious differences' when
they know what they hear, fail to tell what is 'CD original' and what is
'PASC encoded/decoded' in a blind listening test. It's *not* because of my
background that I'm saying this; I can assure you that nobody at Philips
has any interest in claiming a high sound quality, meanwhile risking to make
the Philips label ridiculous when the contrary would turn out to be true!
As a final remark, from a technical point of view (confirmed by tests), with
prerecorded tapes, DCC in some cases can have a *higher* perceived sound
quality than CD, since its dynamic range is far greater than that of CD. Of
course, this holds only if you make a DCC recording departing from a better-
than-CD source (e.g. the commonly used 18 or 20 bit PCM studio master
recording). Have you ever noticed the noise during the fade-out of a track on
CD? That's because the CD quantisation noise has (roughly speaking) always the
same power, independent of the signal level. At low levels, CD has a poor signal
to noise ratio, with the situation made even worse (for the insiders) since at
low levels the noise becomes correlated with the signal, and becomes
therefore more noticeable: it sounds like a 'ringing' distortion. These
artifacts can become much less on DCC. Please realise that the 16 bit PCM
format on CD is also a form of coding, and that one coding scheme does not have
to be always superior or always inferior to another in all cases!

6. Your auditory system: the point above describes some results of listening
tests, because that what we at Philips are interested in. However, take a listen
to DCC with your own ears as soon as you can: never let anybody tell you what
*you* hear.

7. Audio data compression in the future: in addition to DCC, in many
applications that will arrive on the market in the near future, a form of
audio data compression comparable to PASC is used or supported. To name a few:
  * DAB (Digital Audio Broadcasting); will arrive in 1995, and gives high
     sound quality and enhanced other facilities; one of its main advantages
     is the uncomparably better reception in a driving car than with FM; on
     several places in Europe cars are driving around for test purposes;
  * CD-Interactive;
  * HDTV (both in the European and the American version);
  * ISDN (the telephone network's digital successor).

I hope to have answered some of your questions that were still open. I'm NOT
speaking on behalf of Philips, only 'au titre personnel', and only able to
speak as far as my knowledge reaches. If you have any comments, please contact
me via email.

-------------------------------------------------------------------------------
Paul Dillen
Philips Consumer Electronics
e-mail  [e-mail]
-------------------------------------------------------------------------------