Jump to content
Sony Insider Forums

MiniDisc Lifespan

Rate this topic


Recommended Posts

Philippe, are you sure about this?

Looking at the schematic for the W1, it seems that the connections between the deck are normal S/PDIF connections, switchable between direct connection and optical in. Would you agree with me that the ATRAC converter lies after this level of signal?

If that's the case then really the MDS-W1 is nothing special, just two decks (specifically MDS-JE520's) hooked back to back.****

Now comes the potentially interesting part. If you are correct in your claim about no loss (and I don't doubt you on this point), then the ATRAC transform itself is essentially "Magic", and mathematically ensures that ATRAC->S/PDIF->ATRAC is the identity matrix.

Stephen,

What makes the W1 unique is the disc to disc MOVE function. That function bypasses the ATRAC decode/reencode steps, and directly copies ATRAC data from disc to disc. The copy occurs at 4x real-time speed, because only compressed ATRAC data bits are transmitted between discs. I insist: this is a bit to bit copy (even if the disc layout can differ when defrag is involved). For the record, the trick: after MOVE completion, one have to super UNDO the source disc to actually achieve the perfect 1rst gen copy.

The W1 also permits the usual deck to deck copy, ie ATRAC -> decode -> transmit -> reencode -> ATRAC, that honours the SCMS protections: decoded bits are transmitted in the digital domain when allowed, otherwise they are transmitted via the built-in analog connection. I never use that regular lossy copy function (why would I when I can do lossless copies with MOVE ?), except when I want to copy from STEREO to MONO, or change the recording level.

Link to post
Share on other sites
  • Replies 61
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Posted Images

Have you copied a disk, uploaded each with RH1, converted to WAV and done a file compare, then? If what you say is true then this should work, I think.

Can you show me the signal path on the circuit diagram? Again I don't doubt you - I want to believe.

Link to post
Share on other sites

Have you copied a disk, uploaded each with RH1, converted to WAV and done a file compare, then? If what you say is true then this should work, I think.

Here you are:

$ ls -l *.wav

-r-------- 1 felix users 27417196 Oct 4 18:33 001-Jacques Grello.wav

-r-------- 1 felix users 27417196 Oct 4 18:33 036-Jacques Grello.wav

$ cmp *.wav

001-Jacques Grello.wav 036-Jacques Grello.wav differ: char 27370545, line 264673

$ cmp -l *.wav | head

27370545 301 300

27370655 111 112

27370659 331 332

27370661 175 176

27370665 24 25

27370679 170 171

27370703 221 222

27370707 307 310

27370717 303 304

27370729 317 320

First 27.370.544 bytes are strictly identical. Last 46.652 bytes differ slightly. That is is probably a firmware bug, that can easily be circumvented: pad the track with a sound quantum; copy; remove the padding. Practically, the difference is unnoticeable, except in rare cases where sound does not fade at the end of the track.

Link to post
Share on other sites

Those are data, right? Too much for a digital copy imo.

But as I say, my theory that ATRAC->S/PDIF->ATRAC is uni-whatsit would still produce something like this.

Did you take a look at the service manual schematic? I would expect to be seeing a physical interconnect between some pre-decoded data on one drive and the DSP chip on the other. And I don't.

Either way it's amazing. Now the question is, can it be reproduced with an SCMS-killed regular copy? Or one made without digital amplification by making a WAV file and transferring it back to an SP disk?

Stephen

Link to post
Share on other sites

Those are data, right? Too much for a digital copy imo.

But as I say, my theory that ATRAC->S/PDIF->ATRAC is uni-whatsit would still produce something like this.

Did you take a look at the service manual schematic? I would expect to be seeing a physical interconnect between some pre-decoded data on one drive and the DSP chip on the other. And I don't.

Either way it's amazing. Now the question is, can it be reproduced with an SCMS-killed regular copy? Or one made without digital amplification by making a WAV file and transferring it back to an SP disk?

Stephen

I am unable to read the schematic. But I know for sure that the W1 does ATRAC domain copies, that avoid ATRAC decoding/encoding stages. I can ear the difference between a 1st and a 2nd gen track, and I can assure you that the W1 does 1st gen copies. I never pay attention to which copy is the original, because they all are.

I usually edit my current MD hundreds, if not thousands, of times. Each time fragmentation becomes excessive (too many seeks required, to much space lost), I do a full copy on a blank disc, and restart editing from there. The defragmentation that accompanies the copy typically recovers 1 or 2 minutes of recording time from a full disc.

I often repeat that procedure many times while composing a single disc. When I am done, the final master MD will have been copied half a dozen times. And I will make one more copy for my wife, or for friends. And all those copies will be as many more master copies. I have never lost an "original" recording, with the blessed W1.

SCMS bits are ignored, because they control how tracks are copied, not moved :-) SCMS bits are copied along, regardless of their values. My guess is that the generous Sony engineers that built the W1 were perfectly aware of what they were doing: they deliberately suppressed all copy restrictions, but could not advertise it :-(

As I said in an earlier post, I would not have invested in MD, if the W1 had not existed. IMHO, the capacity to make perfect copies is the essence of digital recording. I was looking for a durable, convenient, versatile, physical medium, and I found it. Again, I do recommend the MDS-W1 for long term original sound recording archival on MD.

Link to post
Share on other sites

I am unable to read the schematic. But I know for sure that the W1 does ATRAC domain copies, that avoid ATRAC decoding/encoding stages. I can ear the difference between a 1st and a 2nd gen track, and I can assure you that the W1 does 1st gen copies.

So here's the question: if I can do the same thing by uploading SP to WAV and transforming back to SP, would you believe my assertion about the mathematics of it?

At the moment this is a thought-experiment, as there is no software encoder WAV->SP, although there IS a software decoder SP->WAV

From what I can see, there is NO magic whatever in the hardware circuitry. And there should be.

Link to post
Share on other sites

So here's the question: if I can do the same thing by uploading SP to WAV and transforming back to SP, would you believe my assertion about the mathematics of it?

At the moment this is a thought-experiment, as there is no software encoder WAV->SP, although there IS a software decoder SP->WAV

From what I can see, there is NO magic whatever in the hardware circuitry. And there should be.

Oh now I got you (sorry, I did not understand what you meant earlier). Again, observing the W1 as a black box, one can distinguish three different copy modes between deck A and deck B:

Fast (4x) lossless W1 MOVE:

ATRAC data read|digital copy|ATRAC data write

Real-time lossy digital copy:

ATRAC data read|ATRAC decompression|digital copy|ATRAC compression|ATRAC data write

Real-time lossier analog copy:

ATRAC data read|ATRAC decompression|digital to analog conversion|analog copy|analog to digital conversion|ATRAC compression|ATRAC data write

How this is implemented in hardware, I do not know. From what I have read, ATRAC compression/decompression is asymmetric and lossy, by design. Therefore, I am not comfortable with your hypothesis.

Link to post
Share on other sites

Right. You are saying there are three modes.

I see nothing in the hardware that convinces me that more than 2 modes exist. But I have asked a hardware-expert friend this question too.

One more point: your statement "ATRAC compression/decompression is asymmetric and lossy" I totally accept. Otherwise you'd be breaking the second third law of Thermodynamics, more information in than out.

But I believe the statement "ATRAC decompression/compression is lossless" may well be (and should be) true.

Consider Sound C on CD. C->A->B is clearly going to be a lossy process where B is once again on CD.

But if you take B and transfer to MD(->D) and back to CD, it should be lossless.

Just as A(MD)->E(CD)->F should yield F is identical to A, give or take a few bits (as in your experiment). The key in this case is that you have two identical decks with identical chips and identical encoding characteristics.

Stephen

Link to post
Share on other sites

Right. You are saying there are three modes.

I see nothing in the hardware that convinces me that more than 2 modes exist. But I have asked a hardware-expert friend this question too.

The author of the minidisc.org MDS-W1 page http://www.minidisc.org/part_Sony_MDS-W1.html seems to confirm my perception. I quote:

Unusual "Inter Disc Move" function will move a song from one disc to another at 4 times realtime, defragmenting the copy and deleting the original. The move is done losslessly in the digital domain with raw ATRAC data.

Link to post
Share on other sites

The author of the minidisc.org MDS-W1 page http://www.minidisc.org/part_Sony_MDS-W1.html seems to confirm my perception. I quote:

Unusual "Inter Disc Move" function will move a song from one disc to another at 4 times realtime, defragmenting the copy and deleting the original. The move is done losslessly in the digital domain with raw ATRAC data.

OK, maybe it's time for me to get brutal.

1. You can't prove that my previous explanation is wrong

2. You can't point (neither, so far, can anyone else) to something in the hardware showing any physical connection how this "fairy-tale" version can be accomplished. There are plenty of (more, or less, wildly) inaccurate statements in that website, so this would not be an isolated occurrence. Whoever did it (Eric W?) based their information on the best that was available at the time (and there were probably no service manuals when that page was written).

3. The few errors you see may well be related to something like frame completion/roundoff/skipping/padding, as you yourself admit. That's exactly what you would expect with some timing or other errors associated with a digital (coax or optical) transfer.

I imagine the 4x can be done by changing the frequency of the clock, correct? 176.4 is not exactly unknown these days, there are plenty of hardware devices that support it. Even in 1998.

Stephen

Link to post
Share on other sites

I am curious and will look into the hardware (service manual).

However, I would revise the experiment to an improved experiment with RH1.

Take MDa and copy to MDb, MDc.

File compare MDb and MDc.

They should be exactly the same.

Then copy MDb to MDd and compare.

The outcome of this result could settle the question here.

Link to post
Share on other sites

I imagine the 4x can be done by changing the frequency of the clock, correct?

Yes, Stephen, you must be right. I am not competent in hardware, and I had not thought of that. I could only imagine 4x less data being transfered. Thank you for your explanations!
Link to post
Share on other sites

Yes, Stephen, you must be right. I am not competent in hardware, and I had not thought of that. I could only imagine 4x less data being transfered. Thank you for your explanations!

Finally maybe you're starting to see my point (and it's only a point). Now, please, don't take it as read, I am not a H/W expert either. I am still hoping to get some more answer from my hardware expert friend.

Couple more points:

1. See page 52 (of the MDS-W1 service manual). I am not sure what this is but I see some frequency dividers, that seems rather suspiciously like what we are talking about.

2. If I had 2 decks with coax (less sure about optical because I believe it's not so accurate) I could probably rig something that looked like the MDS-W1. Unfortunately I don't. There's also the problem of SCMS, which we KNOW Sony can solve by tweaking the chip the way they want to (see their Pro Decks which use all the same basic hardware).

So mine is in essence still a "thought-experiment". But Schrodinger and Heisenberg relied on same or less :)

I wonder if converting one of the CD-MD decks would be possible. I don't see any sign of the divider circuit in there but that maybe because 18-bit oversampling has been routine for CD's forever (and that's all you need, right? 4x is just 2 more bits). So one might be able to clone the idea, just not with x4 speed. Presumably the difference between highspeed and normal on the CD-MD decks is something about this, where you reduce the oversampling (perhaps from 20 to 18?).

Stephen

PS were it not for your pointing out of the remarkable properties of this unit maybe no one would have ever figured this out.... thank YOU!

Link to post
Share on other sites

On page 12 of the user manual there is a title copy function md to md. Does this not imply ATRAC to ATRAC direct data copy? Does this mode run at 4x? Can titles be copied through SPDIF? This could rule out one of the modes?

That's the function we are talking about. Unless you mean cloning of the TOC entry (copying the title), which is indeed possible because it doesn't go through the S/PDIF interface at all.
Link to post
Share on other sites

I think the long discussion above basically tries (maybe unsuccessfully) to make the point that the external properties of a device do not necessarily give an infallible guide to its construction. At this juncture I believe the properties Philippe is talking about arise out of some inherent properties of ATRAC rather than from some fundamental trickery done by the Sony engineers in this deck.

Stephen

Link to post
Share on other sites

I think I understand thanks for the explanation.

"FROM ATRAC to CD is lossless" is lossless the best way to describe this transformation?

The point I made (and still subject to verification/proof) is that (maybe):

- starting from ATRAC you CAN get losslessly back to ATRAC (via CD quality)

- if you start from CD you cannot get losslessly back to CD (via ATRAC), because ATRAC must throw away information in order to compress the data and store it.

Clear as mud?

Link to post
Share on other sites

The point I made (and still subject to verification/proof) is that (maybe):

- starting from ATRAC you CAN get losslessly back to ATRAC (via CD quality)

ATRAC compression/decompression is definitely lossy: sound quality audibly degrades with generations. That is true for digital and analog copies. I assume that this was an early Sony requirement, precisely to avoid identical copies of copyrighted sound material. Ironically, the W1 meets the requirement, as it slightly damages the last 1/86s of the track :-) Well done.
Link to post
Share on other sites
  • 10 years later...

I am a dj and great fan of this format. It was a great idea. On the topic of longevity of the discs. I have recently experienced failure to record on sony " old stock" new discs. The coating on the underside seems to have degraded and shed its top surface causing Spotting.. I had others from the same stock I had recorded on previously which still play & record fine. I suspect by recording a chemical reaction takes place which stabilises the surface and strengthens it making it useable time and time again. So if you sitting on new ones you bought years ago. Record to them & rerecord when needed. Trust me , they will last longer.

Link to post
Share on other sites

Recording spot-heats the disc, so this probably drives off any moisture that could cause oxidation of the disc surface.

The discs you experienced this with, were they still in their intact wrappers prior to the use you describe?

I've only seen this (or similar) on one disc - a TDK MXR that I'd had for many years (well since ~1999). On one part of the disc it had what looked like small rust spots on it. It would not read.

Link to post
Share on other sites

This is a very olde thread? But one thing not mentioned by anyone is that the top surface has a lubricant to cut down on record head wear and tear. Maybe a md will stil be ok to play in 50 years time but will it still be recordable, will theĀ  lube have dried up and damage the record head?

Link to post
Share on other sites

I tend to lean towards that "lubricant" is more of a solid form than of a liquid or gel-like. Something similar to the teflon coatings on the "non-stick" pans.

Certainly even that could deteriorate over time, although I still owe myself an explanation, how - besides when exposed to excess heat and/or damp. (That coating on the back of the discs were claimed as a sort of silicone-based compound, and those are generally fairly stable stuff.)

I second to Kevin, none of my MD discs I have ever bought (first one back in early 1998 or so) had any trouble with recording onto, even these days, but one or two (out of ~1400-ish).

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

×
×
  • Create New...