My SHM-CD results...

Hi lukpac,

Oh. For some reason, I though crapfromthepast was referring to some process I hadn't heard of.

To answer his question then, no I haven't tried this.
I'd be curious if someone else does though.

I know when I extract the data from the disk, to find out if it is truly intact, I get a match with the master. My concern in these situations is to show the data that got pressed is in fact the same data I sent the plant. So far, this has been the case.

This is a tough one to explain to folks who haven't heard this.
My mastering colleagues - at least all the ones I've spoken with - have been pretty unanimous in making the same observations.

It would be easy to have someone visit the studio and hear playback of a master vs. various pressings made from that master. Then again, anyone can do this with pressings made at different plants but sourced from the same master. The problem for the average (i.e. non-industry) person is knowing for sure the two disks are sourced from the same master. There are plenty of these around though. In the early days, especially, a lot of big name acts' CDs were often sent to many plants because no single plant could accommodate the amount required by the label.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

I think I know what you're trying to say, but that isn't what you wrote.

If the jitter is so bad that values are being incorrectly read, yes, your statement is correct. But I haven't seen any evidence of that happening. That would be something like this:

Code:

Input:
    0      1     0       1     1
Clock:
|     |     |     |     |     |
Output:
0 1 0 0 1

Note that the fourth sample should have been a 1, but was instead read as a 0. Again, I don't think that's what we're talking about.

Now, in *theory*, as long as the data is read correctly compared to the clock, the exact timing shouldn't matter. Example:

Code:

Input:
  0        1  0        1  1
Clock:
|     |     |     |     |     |
Output:
0 1 0 1 1

The data is read correctly, and would thus show up in a file correctly, but it isn't steady from a time perspective. It shouldn't matter during playback, but...does it?

 

Right. One theory that keeps getting brought up is errors are introduced during real time playback that aren't seen during digital audio extraction (ripping). I doubt this is the case, but it would be nice to try it anyway.

The problem is finding people that can reliably tell the difference, and/or equipment where it makes a difference. I have pressings that some folks have claimed are very different, but I have not yet detected any differences. Nor do I have the equipment to analyze either a digital stream in real-time (as opposed to writing to a file and comparing that) or the analog output of a DAC.

 

True I have never heard an SHM-CD and even if I had one my "system" wouldn't be able to confirm whether it would make a difference or not anyway. Thus, I go by what I know and can find out about how all of this works. I am a physicist, I accustomed to trying to find solid data backed up by solid theory. I've already read about various listening tests showing that people really can hear what they want to hear despite what is actually being produced by their speakers so to be perfectly honest I am more inclined to believe theory than experience in this case. I always want to learn though so this is interesting to me.

See, here's the problem I have with all of this. The laser is continually reading the disc and "noting" where pits start and finish, then deciding whether it the pit it has just read is a 0 or 1 based on the length of the bit, correct? The data must then be passed to the DAC where it is "held" until it has all 32 bits (two channels) and it can perform an analogue conversion. However, the entire system is based on some sort of clock - the DAC doesn't just convert to analogue whenever it feels like, so where does the pits not being exactly in the right place affect anything (other than read errors which I've admitted from the beginning might exist).

Exactly, it shouldn't affect playback. If it does then we missing something.

If you rip a CD with EAC you sometimes get occassions where it has to read the same sector multiple times because an error has been detected (e.g. a parity or correction bit is not what it is expected to be). On a normal CD player, there's two things that could happen. One, the error is not detected or detected but outputted anyway (this would be a "playback error" that would usually be "fixed" by a secure ripper) or two, the player is reading the data quickly enough and has a large enough buffer (as I'd expect from a modern system, perhaps not an older one though) to perform the same sort of "brute force" error correction as the secure ripper.

There must be some sort of unsecure ripping program that can rip data at a fixed speed, e.g. 1x, that can be used to simulate real time playback?

 

Hi lukpac,

Maybe we're talking about two different things but what I wrote is exactly what I meant. The bit might be a 1 but if it happens at the wrong time, it alters the reconstruction of the analog waveform and while it is right in being a 1, it is effectively wrong because that isn't where the 1 belongs.

Does it? To my ears it certainly seems to.
Speaking only for myself of course, any theory that says it doesn't matter is a theory I would cast out.

Or, as Yogi Berra aptly put it:
In theory, there is no difference between theory and practice but in practice, there is.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

Been away, sorry, too many to answer here are a few:

Look at the sheer number of SHM-CDs coming out, it's a veritable flood at this point, mainly intended for the consumption of a small island nation in the pacific. How many people believe they are taking the actual master tape and remastering it for any of these discs? At what expense? And yet the packaging makes no such claim, just goes on and on about the superiority of the MEDIA (not the mastering). If so, this would represent the single largest CD remastering effort in history. Yet, they spent all that money and don't trumpet NEW IMPOVED REMASTERING! I doubt it...

You can usually tell if there are bonus tracks equivalent to the ones on the latest remaster. The booklets also contain mastering info, yes, they may have just repeated the original credit and left the superior Japanese engineer out in the cold unknown, but he should contact his agent...

The Talking heads catalog, however, is the one that really intrigues me, as there are 2 SHM-CD versions. One is the latest Ted Jensen remastering in the LP-style art as evidenced by the same bonus tracks as the latest remastering. The other is the just the original album in regular CD packaging. Is that the original mastering put on SHM-CD? I'm almost crazy enough to order both to try to find out.

OK, great, What is your current system? How many SHM-CDs have you done A/B's with?

It's true, I was snotty and disrespectful in advance of the inevitable snotty and disrespectful posts from flat-earthers I anticipated from doing battle for centuries in the Cables section at another site. Yep ya got me, and I do apologize for that it's bad form to insult first.

 

To my knowledge, real-time playback is a completely different process from DAE, regardless of speed.

 

Hi DragonQ,

I understand and admire the search for the "how" and "why".

As to theory and experience, if theory tells me something say, doesn't hurt and I experience it as painful, I look for a new theory.

I think much about listening depends on the individual listener. As to folks hearing what they want to hear, while this may be true in some instances, I've found truly experienced listeners (there are some folks with good hearing acuity) are pretty reliable.

Perhaps where we differ here is in our expectation of just how much the buffer is going to accomplish. Even if I were to say I thought the output of the buffer is in perfect time (which I'm not saying), something is still wrong.

As I said earlier, I don't pretend to have all the answers. I can only report on my experience. And I find it of interest that every other experienced mastering engineer tells me they experience exactly the same thing.

It took a few years after the introduction of mass distributed digital audio (i.e. CD) for someone to quantify jitter. I am confident time will reveal other characteristics that affect digital playback, particularly realtime decoding of eight to fourteen modulated digital audio from disk.

In the meantime, I have two options:
1. I can accept the theory that suggests pressings from all plants, made from the same master and shown to be bit identical to it "should" sound exactly the same.
or
2. I can accept the response my ears have when I listen back to something I've been working on so long I know it inside out and I wonder, "What happened to the tone of that guitar that opens the album?"

I have to send my clients to a replicator I feel will give them back something reasonably close to what they hear in the mastering room. Interestingly (to me at least), when I was searching for a plant to do the Soundkeeper disks and narrowed it down to a few, the one I ultimately chose (because I feel they do the best work I've heard) was also the only one who told me straight out the disks they send me won't sound exactly like the masters I send them. (The plants that do some of the cheesiest work also happened to be the ones that said, any differences are impossible.)

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

FYI, that process isn't correct. The data on a CD is not simply stored linearly. It is interpolated and there is redundancy involved.

http://en.wikipedia.org/wiki/Eight-to-Fourteen_Modulation
http://en.wikipedia.org/wiki/Cross-Interleaved_Reed-Solomon_Coding

 

I think this is key.

If you rip two differently manufactured CDs with the same digital data to a hard drive and play it back from the hard drive and compare the data bit-for-bit, that will tell you something.

Anyone do this to take real-time playback out of the equation?

The difference that would explain everything could be faults in real-time playback.

 

I agree 100%. That's jitter.

Jitter can manifest itself in only two places: (1) the A-to-D converter that's used to convert the analog 2-track mixdown tape to digital, and (2) the output D-to-A of the CD player.

It's my understanding that there's a buffer in every CD player between the 0's and 1's as they're read off the CD by the laser and the output D-to-A converter. I'm not sure how big the size of the buffer is, but it's certainly there. It was probably quite tiny in the early CD players, when the slightest bump would cause the thing to skip. Later on, the buffer was huge, like in the walkman-style portables that stored 20 or 40 seconds of audio in its buffer.

Even if the incoming data stream, as read off the disc, had plenty of jitter, then all that should happen is the buffer is fed too quickly or too slowly. The clock in the output D-to-A converter runs with a mind of its own, so to speak, and should be unaffected by the input to the buffer.

Somewhat related:

Barry - I appreciate your discussion of eye patterns - I used to work in the data storage industry and I know all about those things. For those that don't know, an eye pattern is what you'd see on an oscilloscope if you just looked at the raw output from the CD player intensity detector. All the pits on the disc correspond to particular discrete lengths, so on the scope you'll see spikes - one spike for each allowable pit length. Each spike has a little noise to it, corresponding to real variations in the pit length. You want an "open" eye pattern - the openness means that unwanted variations the pit edges are small enough so that a particular pit isn't mistaken for a longer or a shorter pit.

I also 100% agree that if your eye pattern is more closed, then you've increased the likelihood that a particular pit length will be mistaken as too long or too short.

It sounds like you've seen plenty of variation in the eye patterns from CD to CD. I would think that this would directly affect the raw bit error rate, so that one bit in every 10^6 (10^8? 10^10?) bits would have the wrong value. (Don't remember the exact spec for audio CDs; it was around 10^12 when I worked with computer hard drives.) Even if it's as low as one bit in every 10^6 bits, that boils down to one 16-bit sample being off by only one level every 11 seconds (2 channels, 44,100 samples per second), which isn't nearly enough to cause the kind of effects that you're hearing.

 

Not that I care, since I don't use a CD player, but I don't believe for a second that there is any physical difference with these CDs (never mind whether it would matter). It's not like putting poison in baby milk...nobody will go to jail because of a little lie about a CD.

 

Based on not listening and no evidence. Thanks for playing. Don't let your pre-conceived beliefs (based on nothing) interfere with evidence of your ears. Come back when you've done some actual A/B's.

 

Like I said, I don't own that 1980s technology. But think about it from a business point of view. It costs nothing to put a marketing claim on a package. Asian bootleg CDs are often sprinkled with "audiophile" logos like HDCD, DSD, etc. They are playing with your head.

 

IIRC, Barry has said before that if he rips the factory CD to his hard disk it sounds just like the master. So, wouldn't it be a matter of comparing the playback of the CD with that from the hard disk going into the same DAC?

 

The contradiction here is that if the SHM discs did in fact bring a new mastering people would say that the difference in sound is due to it and not the material or process used (as in Blu-Spec CD).

 

In other words, no, you have no clue if the discs you're A/Bing are in fact bit-identical or not? You're just guessing?

 

Yep.

I wondered in a previous thread if one couldn't analyze this at the analog stage?

I mean, if we assume for argument's sake that this effect is real. That two discs with the exact same bits - one a carefully manufactured SHM, one from a slipshod el-cheapo plant - can sound different when played in real-time on a CD player. Don't we need to assume the resulting analog signal traveling down the interconnect has to be different as well?

Aren't we at the stage where most folks with a decent soundcard and software can record an SHM and a plain-vanilla CD playing on a few different players and compare the difference? Seems simple enough, but of course I haven't tried it. . . .

 

Barry - I believe you may be thinking of a Robert Harley quote from his Stereophile article on jitter in 1992:

"Now, look what happens if the samples are reconstructed by a DAC whose word clock is jittered (fig.3). The sample amplitudes—the ones and zeros—are correct, but they're in the wrong place! The right amplitude at the wrong time is the wrong amplitude. A time variation in the word clock produces an amplitude variation in the output, causing the waveform to change shape. A change in shape of a waveform is the very definition of distortion. Remember, the word clock tells the DAC when to convert the audio sample to an analog voltage; any variations in its accuracy will produce an analog-like variability in the final output signal—the music."

http://www.stereophile.com/reference/193jitter/

 

my thoughts exactly, 16bit/44,1kHz is 1980s standard, it doesn't matter if you put that data on super clear plastic or on your hard drive it's still 16/44,1kHz, not even 1Hz more

SACD's and 24/96 flac downloads are a different story

 

It doesn't have to be a full-blown Hoffman/Gray jobbie for each and every one, no.

Say they wanted to add a little extra sparkle, just so folks are really wowed by these new SHM discs. (They have a vested interest in doing this, mind you.) So, they decide to boost 12kHz by a decibel or so using the current mastering. (Japanese masterings are famous for this, remember.) That would take your "superior engineer" what, 30 seconds?

Considering they are already there preparing for a new transfer the net cost would be nil.

 

If such an approach was being taken, how would we have instances where a SHM CD cancels out with a non-SHM CD originating from the exact same mastering?

We wouldn't.

Btw, I'm no fan of SHM. I think the whole idea is an exaggerated joke.

 

You're talking my comment out of context.

I asked markl if he was certain the CDs he was comparing were bit identical.

His reply was that he is not sure at all, he just assumes since no new mastering engineer is touted on the tray insert, and the bonus tracks are the same, that they must be bit-identical copies of the CDs he already has.

I'm saying that no way is that enough evidence for the claim he started this thread with; that the digital content on his SHM discs is the same. Got it?


Now, what are you saying in reply? That you've tested all 20 SHM discs markl has, and all his regular issue counterpart CDs, and you can confirm they are bit-identical? If so, fair enough. . . .

 

If this is a problem, why are there still multi-thousand dollar CD players made? Wouldn't it better just to securely rip everything, and play it back from a hard disc? Either CD playback works or it doesn't, if it doesn't work, then high end CD players shouldn't exist. They should all be CD rippers that store everything on hard discs.

 

That would make us all sad and lonely individuals, according to markl.