Two unresolved issues

[Bruce B]

Again, I'm really not trying to be dense. But you need to spell out what you're showing. Did you record from your own SACD player? Do we know which of the SACD players used for the M&M tests can and cannot output above 20 KHz? I honestly don't understand what you're showing.--Ethan

Yes, Ethan... now you are being dense.

1. I gave a link to the equipment used.

2. I gave a link to the analysis of one of the discs that were used.

3. I did an analog capture of 2 of the players used and compared them to the original file.

3. The players they used are not much better than Redbook and that anyone would have a problem identifying hi-rez from RBCD. Yes, they do output measurements above 20k... but not much.

4. I gave a link to the analysis of a player here in the studio to show you what a good SACD player should look like. Using this player and inserting a bottleneck would allow you to correctly identify, blindly, which is hi-rez and not.

5. Ethan, you're arguing for the sake of doing so when the information is provided before you. So unless you have anything more to contribute to this discussion as far as my data is concerned, then I'm moving on. I'm not feeding the troll. So if you're "honestly don't understand what I'm showing", then I think you should seek another line of work.

Please answer John's questions. We want to see data. I showed you mine.

 

[Bruce B]

Thanks for the link Bruce. BTW, yesterday at the CSA store in NJ, Peter McGrath played me the 24-bit/96kHz files of some of his analog Audiofon piano recordings that you had restored and transferred to DSD. Excellent job! I am looking forward to these recordings being available from HDTracks.

John Atkinson
Editor, Stereophile

Thank-you John. We used a modified Studer A80RC MkII into a Sonoma DSD system. The tapes were 30ips and had some of the best sonics we've ever heard. Would anyone expect less from Peter?

 

[microstrip]

Thank-you John. We used a modified Studer A80RC MkII into a Sonoma DSD system. The tapes were 30ips and had some of the best sonics we've ever heard. Would anyone expect less from Peter?

Bruce,
Can we ask you what were the modifications carried on the Studer?
Although I also own an A80, I will have to wait for the release of your transfer.

 

[Bruce B]

Bruce,
Can we ask you what were the modifications carried on the Studer?
Although I also own an A80, I will have to wait for the release of your transfer.

Ki is the one who rebuilt all the boards with new caps and such. I'm sure he could tell you everything that was done. NOS pulleys, guides, etc.. were purchased as well. The headblock was sent to JRF who replaced the heads with Flux Magnetics ER heads. This was initially a 1/2" machine with low hours. A new meter bridge was purchased from ATAE.

 

[JackD201]

Ah... that one. I can't quite get there, though in practice, particularly with judging electronics, I think we're very, very close.



Actually, that was me, though Lee may have said it too. We may have a similar grasp of the obvious .



Is it really so different? Audiophiles claim to hear that which the existing science does not acknowledge. More than that, they claim it as evidence of their scientifically unsupportable "truth." The virgin Mary in a piece of toast. It should be very easy to understand why it is so suspect to those who don't hear it.



I own a 1965 Fender Deluxe Reverb. I also own a digital amp simulator. The simulation is remarkably close, but not right on. The simulator sounds more like a really good recording of my Deluxe, than my Deluxe. "Everything" is a really big word.

Back on topic so we don't get deleted: The argument against measurement is a bit of a stretch, I think, but I'll take it; truly comprehensive measurements are pretty rare and there might be something there that we can't yet measure. I won't get into AB/X because it's difficult and expensive to do them well and no matter how well they're done, they're complex enough that those who don't want to believe the results can deny them; they always do.

But that leaves us with one of Ethan's questions that I don't think anyone has an answer for: The null test. Those things that digital leaves behind, those extra unmeasurable goodies that analog preserves -- in a proper null test, would they not be the only thing left? And if audiophiles hear them, would they not clearly be audible? And shouldn't that address the question whether we're talking about the audible difference between a Pioneer and a Krell, or the musicality gap between 80 g vinyl on a $20k tt vs. lossless files from my MacBook? Couldn't, shouldn't that be an end to the conflict, an answer to the questions? If not...why not? how is that audible something that makes it through digital to analog conversion, pre amplification, amplification, cabling and transducers disappearing in the null test?

Tim

Hi Tim,

Please refer to Amir's post #119. He expressed it a million times better than I ever could.

Jack

P.S.

Don't be surprised if the thread gets shifted to the audibility of subtle differences all of a sudden.

 

[Bruce B]

But that leaves us with one of Ethan's questions that I don't think anyone has an answer for: The null test. Those things that digital leaves behind, those extra unmeasurable goodies that analog preserves -- in a proper null test, would they not be the only thing left? And if audiophiles hear them, would they not clearly be audible? And shouldn't that address the question whether we're talking about the audible difference between a Pioneer and a Krell, or the musicality gap between 80 g vinyl on a $20k tt vs. lossless files from my MacBook? Couldn't, shouldn't that be an end to the conflict, an answer to the questions? If not...why not? how is that audible something that makes it through digital to analog conversion, pre amplification, amplification, cabling and transducers disappearing in the null test?
Tim

A null test like Ethan is talking about only tests 2 digital files. What you want is a null test to see if the digital is REALLY capturing all the signal. So.. here is how to do it.

Say.. for instance, you're recording an orchestra. Take a signal in from the microphones and monitor on headphones or a separate remote truck. Record the event in the digital domain. Now you have 2 signals... one analog and one digital. Now the tricky part. Put an analog delay on the analog signal that compensates for the time it takes the recorder to record in digital and reverse the polarity (probably a few hundred ms). Then you can see if digital IS capturing all the analog signal. Agreed?

 

[Scott Borduin]

A null test like Ethan is talking about only tests 2 digital files. What you want is a null test to see if the digital is REALLY capturing all the signal. So.. here is how to do it.

Say.. for instance, you're recording an orchestra. Take a signal in from the microphones and monitor on headphones or a separate remote truck. Record the event in the digital domain. Now you have 2 signals... one analog and one digital. Now the tricky part. Put an analog delay on the analog signal that compensates for the time it takes the recorder to record in digital and reverse the polarity (probably a few hundred ms). Then you can see if digital IS capturing all the analog signal. Agreed?

Yes, this would be the "real" null test - compare the A/D - D/A chain to a piece of wire. A piece of expensive wire, if you insist

But of course, we know for a fact that a digital recording chain like this will not pass a null test, and I assume that is not in dispute. We know for a fact that digital recorders and playback devices add noise and distortion; we know for a fact they add a certain amount of ringing at the frequencies of their anti-alias filters. What is actually in dispute is whether these artifacts are audible, and how we can know whether they are. This is as much a philosophical question as a scientific one - I'm putting together a post for tomorrow that goes more into this.

 

[Phelonious Ponk]

A null test like Ethan is talking about only tests 2 digital files. What you want is a null test to see if the digital is REALLY capturing all the signal. So.. here is how to do it.

Say.. for instance, you're recording an orchestra. Take a signal in from the microphones and monitor on headphones or a separate remote truck. Record the event in the digital domain. Now you have 2 signals... one analog and one digital. Now the tricky part. Put an analog delay on the analog signal that compensates for the time it takes the recorder to record in digital and reverse the polarity (probably a few hundred ms). Then you can see if digital IS capturing all the analog signal. Agreed?

I don't know, Bruce, I've never done anything like that. Not sure I'm even certain that I quite understood what Ethan was talking about now. What I thought I understood he was saying was that you could subtract one signal from the other, leaving behind the difference between them. That way we could understand what digital doesn't capture, for example. Did I get that wrong?

Tim

 

[Bruce B]

I don't know, Bruce, I've never done anything like that. Not sure I'm even certain that I quite understood what Ethan was talking about now. What I thought I understood he was saying was that you could subtract one signal from the other, leaving behind the difference between them. That way we could understand what digital doesn't capture, for example. Did I get that wrong?

Tim

Right... in the configuration I mentioned, you reverse the polarity (subtract) one of the signals, either the digital or analog signal and you'll hear what's missing. Subtract the analog to see what digital is missing or subtract the digital chain to see if the analog is missing anything. This is easily done in our studio.

 

[Phelonious Ponk]

Right... in the configuration I mentioned, you reverse the polarity (subtract) one of the signals, either the digital or analog signal and you'll hear what's missing. Subtract the analog to see what digital is missing or subtract the digital chain to see if the analog is missing anything. This is easily done in our studio.

OK...and you can do this with analog? So you can actually hear what was on the analog medium that wasn't captured by digital?

Tim

 

[JackD201]

......and vice-versa Tim. The thing is, we cannot discount what Scott said that the recording processes add as well as omit. The likelihood of a perfect null is thus extremely unlikely. Analysis of the data now becomes the focus of attention as Amir stated in 119.

 

[Phelonious Ponk]

......and vice-versa Tim. The thing is, we cannot discount what Scott said that the recording processes add as well as omit. The likelihood of a perfect null is thus extremely unlikely. Analysis of the data now becomes the focus of attention as Amir stated in 119.

If you can do this with analog, as I just asked, I wouldn't think the analysis of the data would be all that difficult. What is it? In what frequency range does it lie? Is it linear? Is it noise? Is it musical? The first question I'd ask if someone did such a test is cool...what does it sound like?

Tim

 

[JackD201]

Yeah, that's what I'd ask too. To answer our question the data would have to be correlated with the parameters and conditions you mentioned plus perhaps some others. That is the tough part because even the manner in which the measured parameters are set up by the tester affect the output data. HOW something is measured is as important as WHAT is measured and the manner in which the output data is synthesized is crucial to the validity of any conclusion.

 

[Bruce B]

Microphone -> Mic preamp -> Console -> Digital Recorder -> Headphone/Speakers
/
/
Orchestra
\
\
Microphone -> Mic preamp -> Console -> Delay -> Headphone/Speakers


Then you can invert polarity on each to see what either analog or digital is missing.

You'll need a remote truck outside the building to isolate you from the music.. or use headphones

 

[JackD201]

Hi Tom,

Think of it as pre vs post recording monitoring with a twist. In a two file null test, both have been subjected to ADC so you really aren't comparing analog to digital you'd be comparing digital to digital, strictly speaking. Bruce's method keeps one signal purely in the analog domain and that is what is compared with the digital copy. The delay and inversion is necessary to allow for the cancellations to happen (you know that already so we can skip the details of that). One could scope them too instead of having a remote station. Here is one possible variation of Bruce's concept.

The scope should flatline at full null.

 

[microstrip]

Hi Tom,

Think of it as pre vs post recording monitoring with a twist. In a two file null test, both have been subjected to ADC so you really aren't comparing analog to digital you'd be comparing digital to digital, strictly speaking. Bruce's method keeps one signal purely in the analog domain and that is what is compared with the digital copy. The delay and inversion is necessary to allow for the cancellations to happen (you know that already so we can skip the details of that). One could scope them too instead of having a remote station. Here is one possible variation of Bruce's concept.

The scope should flatline at full null.

Jack,

Most probably the analog delay box "distortions" will make the test invalid. First you would have to prove that this box is "transparent" .

 

[JackD201]

....or use exactly the same cables. Hahahaha!

 

[Phelonious Ponk]

This is hurting my head. So can someone tell me simply, could this eliminate the commonalities between digital and analog reproductions of the same recording, leaving us with the unmeasurable stuff that analog preserves and/or adds? And once you've done that, could you then look/listen and understand what that stuff is? IE: Is it signal digital loses or distortion analog adds?

Tim

 

[fas42]

Eveything so far is pointing to the fact that a simple null test has one very key attribute: it is not simple! The gymnastics that will need to gone through to get a meaningful and accepted result will virtually guarantee the uselessness of this as something practical, something straightforward to apply. We have just seen the impractically of null testing for comparing before and after treatments of an LP, as a simple example.

Far better, I would suggest, to use various testing procedures to generate comprehensive measurements in the areas of spectrum analysis and the like, done at a summary, and on an instance by instance level of detail. That is probably at least one decent way of getting a handle on what's going on ...

Frank

 

[fas42]

This is hurting my head. So can someone tell me simply, could this eliminate the commonalities between digital and analog reproductions of the same recording, leaving us with the unmeasurable stuff that analog preserves and/or adds? And once you've done that, could you then look/listen and understand what that stuff is? IE: Is it signal digital loses or distortion analog adds?

You're right, Tim. At the moment, as I have just said, it's too bloody hard!

Frank