Conference hips::uk_audioo

You've heard of a lot.  Tried listening to anything?

Well, I've listened to people who've tried some of these things and 
reported no audible (to them) effect.  I freely confess to not having 
tried such things myself, but then I'm not unhappy with the sound of my 
current system.  I am intrigued though by such apparently wacky ideas; I 
mean, what is the rationale for even *trying out* the effect of freezing 
a CD ?  Did anyone ever try this with vinyl ?

Aural perception is so subjective.  I don't doubt that people spending 
500 pounds on speaker cables get more pleasure out of their system than 
if they'd spend 50 pounds, or even 5 pounds - after all who in their 
right mind would would spend such sums for no perceived effect ?  But I'm 
equally in no doubt that other people couldn't tell the difference and 
suffer no impairment to their listening pleasure.

The placebo effect is a well known phenomenon and a convenient 
explanation for sceptics - though the perceived effect for the converted 
is no less real as a result.  But if this isn't the answer then there 
must be a 'real' reason and, for the example in question, I'm intrigued 
as to what hypotheses are offered for the proposal that freezing then 
thawing a CD can actually make a difference to the resulting sound.

Mike

I needed no prior rationale, nor justification, to try this...
rushed out to the freezer with my U2 Joshua *vinyl*, and gave it
15 minutes at 5 dedrees C.  Result:  it still lacked a bit of clarity to my
ears.  Maybe I'm cooling the wrong thing ?  I'll try the arm/cartridge next.

I did notice some ice crystals, then some minor condensation, and a bit of
warpage.  The sound may have changed as the temperature rose.
(Wonder what that condensation would do to the CD error-correction ? )

I'll post my cryogenic CD results R.S.N.  ;^)

The article I read (Sunday Times, 10th Oct, front page no less) said the 
CDs in question were sealed in plastic bags to prevent condensation and 
then thawed prior to being played.  

Perhaps this could happen 'naturally' if the delivery van is parked 
overnight somewhere in the depths of winter.  How would one know ?

Mike

    I also heard you need to get them down to Kelvin(?) which is around
    -273, so no domestic appliance will do this.

Who has a fridge that can go close to Absolute Zero ?

Who in the Hi Fi World has tried it temperatures as low as this?

Mark.

The Sunday Times article only mentioned domestic freezers.  I guess this 
means somewhere around -10 degrees centigrade/celsius.  I don't know 
where the near-absolute zero suggestion came from.

At that temperature I'd be more concerned about shattering the CD; I 
can't see how that would help the sound quality :-)

I can think of a number of styles of 'music' that would, to my ears, be 
much improved by shattering the CD!

(only � :-) )

    Sometimes it's funny reading old magazines.... From the HIFI News &
    Record Review April 1991, the Headroom page from Ken Kessler:
    
    "Last year (!), Ed Meitner of Museatex - probably the most radical
    designer in all of hi-fi - discussed the effects of cryogenic freezing
    on materials"
    "Stereophile asked Ed Meitner to treat a number of copies of the
    magazine's first test CD, for comparison purposes. So wild were the
    results that Stereophile offered 'frozen' copies for sale at CES, where
    I bought my copy. Back in the UK, I did a side-by-side comparison with
    the standard copy of the test CD. The results were - how shall I put
    this? - worying.
    So I enlisted the help of a known Golden Ears from one of the larger
    hi-fi companies, one with a reputation for addressing audiophile
    concerns. This individual shall remain nameless, because he left as
    mystified as I was.
    So drastic were the differences between the normal and 'frozen'
    versions that we wondered if Stereophile had supplied Ed Meitner with
    different takes of certain tracks. In particular, Track 13, with Lesley
    Olsher singing Gershwin's Summertime sounded like a different
    performance."
    His conclusion:
    "Listen: I am the least spiritual, least metaphysical slob you'll ever
    meet. But there's something straight out of E Nesbit going on with this
    cryogenic thing."
    
    Would be interesting to listen to a 'frozen' CD and a 'normal' copy....
    
    Willem

    I can offer one consideration about comparing tracks that should be
    kept in mind.
    
    Some years ago when I worked for Solid State Logic (a mixing desk
    manufacturer) one of the design engineers walked into one of the test
    studios and asked several of us to decide what the difference was
    between two theoretically identical CD's when listened to.
    
    All sorts of words were used to describe the very subtle difference
    between the two CD's, and none of us could decide what was responsible
    for the variation. Finally, one of the engineers twigged what was
    causing the change in audio quality; the second of the two disks was
    playing back about .25 to .5 of a dB louder, and the manifestation of
    this was a very subtle change in overall character....
    
    All is often not what it seems....
    
    Russell.

.10 seems a more plausible explanation (IMHO of course) ... after all, 
what's the difference between a music CD and a CD-ROM (or even CD-V) at 
the physical/bit level ?  Yet I've not heard any suggestions that 
freezing CD-ROMs 'improves' data (or even CD-V images).  Intriguing.

Mike

    re .11
    There is a big difference between audio CD and CD-ROM, Photo-CD etc.,
    not in the technology but in how they are used. Audio CD is the only
    one that, during replay, the read data moves continuously and in real 
    time from the digital domain to the analogue domain.
    
    How small changes in this process can affect sound quality isn't fully
    understood yet.
    
    Dave

re .12
What about digitally-derived images, eg CD-V ?  Isn't that moving 
continuously from the digital to the analogue domain ? (although I'd 
accept that the eyes are more tolerant of image degradation than the ears 
are to audio degradation).

I'm still trying to understand any mechanisms by which freezing a CD 
could affect the final sound.  The CD contains only digital data so any 
effects of freezing must be limited to the digital domain - in which 
case, wouldn't the same effects be noticed with CD-ROMs ?


Mike

Digital data is a stream of 0's and 1's

Any change you make to the data can only be flipping the state of those bits
and having done that you have corrupted the data and trigger the data integrity
checks.

Apart from the fact that a cold disk put in a warm CD player might in some
way improve the analogue part of the device, I think its hogwash! 

-Mike-

    re .14.....
    
    It should be remembered that the electrical waveform derived from
    reading the data on a CD is an analogue representation, not a nice
    clean 1 or 0 state; that only comes later.
    
    Apart from the data, what is also derived from this waveform is a clock
    that is used to synchronise the player's internal clock with the clock
    used for the mastering. Although degradation of the shape of the
    electrical waveform should not affect the data itself, it can affect the
    timing of when the waveform transits from 1 to 0 or vice versa, and
    thus the stability of the clock. Again this shouldn't affect the data
    itself.
    
    But data is sequenced into the DAC using a derivative of this clock, and 
    variations in timing much less than current industry standards can, all 
    other things being above a certain standard, show up as changes in sound 
    quality. Various things, such as acoustic feedback or mechanical
    feedback from the CD mechanics, to power supply load changes during
    seeking or error correction, can cause clock sidebands.
    
    It may just be that freezing the CD subtly changed the shape of the
    pits on the CD, and therefore the waveform shape, and this is where
    changes in sound come from. Most of the research has been subjective.
    so who knows?
    
    Dave

I realise the digital bit stream is initially read as an analogue waveform, but
ultimately there is a string of 1's and 0's which is the real data and is most
unlikely to be any different whether the CD is hot or cold. Do we agree on that?

So now we come to the timing. I confess to not knowing exactly how this works
from a CD standpoint, but at the D to A end, samples must be converted at the
same rate that they were generated in order to faithfully reproduce the original
analogue signal.

So now it depends on the details....

Presumably the D to A conversion clock would be phase locked to an imbedded clock
on the disk and assuming a conversion clock which is fixed, then the sensible
circuit design would have the PLL feedback loop fairly sluggish so that the clock
wouldn't jitter all over the place. But regardless of this, any change in timing
brought about by changing the waveshape of an imbedded clock signal would surely
give rise to a slight phase shift in the conversion instance rather than change
the frequency of conversions... 

Lets try and show an example...

Heres an example waveform and its converted clock
   ___         ___         ___         ___         ___         ___
  /   \       /   \       /   \       /   \       /   \       /   \       
 /     \     /     \     /     \     /     \     /     \     /     \     
/       \___/       \___/       \___/       \___/       \___/       \__

  _____       _____       _____       _____       _____       _____     
 |     |     |     |     |     |     |     |     |     |     |     |    
 |     |     |     |     |     |     |     |     |     |     |     | 
_|     |_____|     |_____|     |_____|     |_____|     |_____|     |____

Heres a distorted waveform and its derived clock...
    __          __          __          __          __          __
   /  \        /  \        /  \        /  \        /  \        /  \       
 --    --    --    --    --    --    --    --    --    --    --    --
/        \__/        \__/        \__/        \__/        \__/        \__

    ___         ___         ___         ___         ___         ___     
   |   |       |   |       |   |       |   |       |   |       |   |     
   |   |       |   |       |   |       |   |       |   |       |   |    
___|   |_______|   |_______|   |_______|   |_______|   |_______|   |_____

Can you explain to me how those two clocks, when used for D-to-A sampling
are going to make the sound any different?

Actually, I just thought of one possibility and thats a relative phase shift
between left and right channels but surely the design must be bullet proof
in allowing this to happen....

-Mike-

    re .16....
    Well, from your example there would be no difference. And I agree with
    your comment about changes in waveshape not affecting the data.
    
    But your second clock example assumes:-
    1. That successive clock pulses are all exactly the same shape and
    cross the 1/0 threshold at the same point in time relative to the
    previous pulse;
    
    2. That the word clock that is divided down from the bit clock you
    illustrate will be unconditionally stable, which it is only likely to
    be if your first assumption is correct, all other things being equal.
    
    I'm not arguing whether all this is horsefeathers or not, and I have
    neither the time nor the inclination to experiment myself, let alone
    the equipment to carry out exhaustive, objective scientific tests. But
    the medium is new and, like psychoacoustics, not fully understood.
    That's why I don't dismiss it, even though it's not relevant to me.
    
    I still prefer LPs anyway.
    
    Dave

I think there are another couple of points too...

1) If the cooling brings about a waveshape change, is it not likely to
   consistantly modify successive pulses in a similar way?

2) There must be something in the design of a CD system that damps any clock
   jitter on the final D to A conversion, whether this jitter comes from
   zero crossing detection of the analogue signal, speed variations in the disk
   or whatever. Why? Because the original sampling would have been jitter free..

Thanks for the discussion

-Mike- (CD convert of many years...) 

    Mike,
    
    First point 1. Yes, that is reasonable, but I don't think anybody
    knows. The freezing process was to near absolute zero anyway; maybe it
    does something to inconsistencies in the molecular structure of the
    disc that could cause waveshape variations, but that's just a
    particularly uninformed supposition on my part!
    
    Point 2. Total jitter removal is an ideal that is never achieved in the
    real world of real hardware. The less the jitter the better the sound
    is pretty well accepted now, but it is difficult to characterise the
    effect that jitter can have on sound, and even small amounts can make a
    difference to what is heard.
    
    Jitter isn't the only thing affecting sound anyway, it's just the
    currently-fashionable thing to reduce, like distortion in 1970's
    transistor amps. In a few years it'll be something else; Pioneer are
    already getting rave reviews for their domestic DAT recorder that
    samples at 96kHz. If the industry moves to a new sampling standard that
    is more than double the current rate, we'll get a whole new set of hifi
    folk devils to worry us!
    
    That's why I'm sticking to relatively technology-free analogue for my
    main source ;-) It dampens the hobby instinct but, oh, the music....
    
    Dave_who_isn't_really_cynical (honest...)

It seems to me that all of this "jitter" stuff can be fixed by double buffering
the data in RAM (ala MINIdisc).

You can totally decouple the timing variations caused by the CD reading process
from the timing of data to the DAC simply by having a large enough buffer - it
doesn't matter that the data is read a couple of seconds before it is sent to the
DAC (given that the long term data rate is correct, you won't overflow or
underflow you buffer this way).

So - now the jitter etc in the CD reading process can have no effect on the
timing of the data to the DAC, we need only worry about the stability and jitter
or the clock that clocks the data out of RAM into the DAC - which can be very
good indeed - tho it could still affect the sound of course.

The point is, any CD transport should be as good as any other if linked to a
double buffering DAC! (assuming it can read the data OK in the first place) - if
you have a 600Mbtye odd buffer, you could even read all of the data in 
before you acutally start to send it to the DAC.

The pit distortion/sampling clock waveshape ideas are certainly 
interesting and seem like a possible basis for such effects.  However, 
until .19 mentioned it, I hadn't connected the double buffering mechanism 
used in some portable and in-car CD players with this 'phenomenon' - or 
rather _dis-connected_ it!

As I understand it, these portable players have a 10 sec (or so) data 
buffer.  The CD data is written into the buffer asynchronously to the DAC 
circuitry reading it out, which gives the player up to 10 secs (or so) to 
recover from a physical jolt.  Now I don't know the exact details of how 
the DAC clock is generated in such players but it seems reasonable to 
assume it is completely isolated from the actual CD transport and, 
therefore, and cryogenically induced distortion of the CD itself.

But would that simply mean that such players would be unable to deliver 
any 'benefits' from a frozen CD ?  It would certainly be an interesting 
basis for some blind listening trials.

Thanks for an interesting discussion.
Mike