What's the deal with vinyl? (2 Viewers)

[Lee Scoggins]

Musical instruments generate both even and odd order harmonics so even odd order can be pleasant.

True and a good tube design can also be good on odd order harmonics.

It's time to bust this "euphonic" myth on tube amps.

 

[Chu Gai]

It's time to bust this "euphonic" myth on tube amps.

Then why have you tube-rolled? Different degrees of accuracy?

 

[Steve_AS]

A rough calculation suggests that the surface roughness should be of the order of 3 nanometers, or 30 Angstroms, to give the known level of surface noise. This is of course pretty darned smooth by most measures, and is about five times the diameter of a vinyl molecule! The minor radius of a modern "fine line" stylus such as that of the Clearaudio Insider is about 5 micrometers, more than 1,000 times
larger.

 

[Saurav]

I believe that if you get accuracy, you get closer to the recorded performance and this translates to more listening pleasure since you hear more of the detail and nuance and the right tonality of the musician's craft.

*shrug* Not in my experience, not always. I have many recordings where I enjoy the music more if the sound can be tweaked a little from as-accurate-as-it-can-be. Most of my Beatles albums would fall into this category. I dunno, that's just me - tonality, detail, soundstage, imaging, yada yada yada - those are usually very low on my scale of priorities. I know, not a card carrying audiophile at all

Have a nice day, guys.

 

[Steve_AS]

Steve, you've left a very critical part fo the equasion out.

*Sample Rate*

Compare sample rates on digital recordings to a *truly continuous* analog waveform.

Or would you contend that sample rate is unimportant (since you didn't mention it)? Or, if you believe that common sample rates are sufficient to replicate a truly continuous waveform. If so, which one? DSD, DVD-A, and CD use different sample rates.

Sample rate's important, and the Nyquist/Shannon theorem explains why it has to be at least twice the highest expected frequency in the signal, in order to *completely* capture the useful information in that signal. In essence, they proved that 'common sample rates' *are indeed* sufficient to replicate a 'truly continuous waveform' in digital form.

(The bigger problem isn't the sampling rate, it's regenerating the analog waveform from the digital version, and AIUI *that's* where information loss is likely to occur, *if* it does (and it needn't).)

I believe that the 44 kHz sample is rate is sufficient to do this for audio, as evidenced by the existence of numerous excellent digital releases on CD. Higher sampling rates in *recording* are useful not because 20-20kHz information is lost by by 44 kHz sampling per se, but because they make implementing noise filters easier, or provide more digital 'headroom' for multiple rounds of processing in the digital domain.

 

[Steve_AS]

And that's totally besides the point. IMO, anyway. No engineer (AFAIK) argues that tubes are more accurate than solid state, or vinyl is more accurate than CDs. They're not, period. And the point is, that that has no bearing on anything. Most people who listen to vinyl and/or tubes do so because they prefer the sound, not because they're more accurate. This is a purely subjective preference, and so by definition, there is no right or wrong to it (again, IMO).

Some people dismiss it as "oh, they're just looking for euphonic distortion". Well, I've never been convinced that accuracy is a more important goal for a stereo system than musical enjoyment. Our hearing isn't linear by any stretch of the imagination. For instance, we react differently to different kinds of distortion, even if they're of the same magnitude. So, it seems logical to me to build a system that is subjectively the least distorted, instead of having the lowest distortion as measured by an instrument (which is almost always linear). So, if you gave me two components A and B, and A sounded less distorted to me than B even though B was measured to have lower distortion than A, I would always pick A. That's a no brainer to me. Why should I pick the one that's more 'accurate', if it sounds worse to me? Doesn't make any sense at all.

This seems to me a mostly sensible point of view. Bravo! Listen to what you think sounds subjectively best. BUT: just be prudent about making claims as to *why* it does so. It could be better accuracy. It could be something added -- ie.., distortion -- that sounds pleasant to you. We are indeed more sensitive to midrange frequencies -- that accounts for the 'smiley EQ' that's so popualr on boomboxes with EQ sliders -- and on some Mobile Fidelity LPs that people love.

 

[Saurav]

It could be something added -- ie.., distortion -- that sounds pleasant to you.

Or it could be something missing, i.e. phase errors from poorly implemented analog filters in DACs

Just pointing out that there's always two sides to any coin. And we don't really have a very good handle on how we hear, what we are sensitive to, and so on. Clubbing various phenomena into one big "distortion" lump and measuring its absolute level is erroneous, IMO. This is because we find different kinds of distortion distasteful to different degrees, and so evaluating all "distortion" equally does not correlate well with how we react to that distortion.

Anyway, I believe I've made that point already. So now I'm rambling.

 

[Chu Gai]

Oh let's not go into phase errors now! I think when the order of upper harmonics gets up high and are also audible then it might be appropriate to call the device broken. That just might wreak some havoc with tweeters! I've a feeling most of the differences that you & I may have Saurav would be washed out after a few beers anyways. Krall is cute though. Have you seen her 'unpublished videos'? My my!

 

[Saurav]

Krall is cute though. Have you seen her 'unpublished videos'? My my!

Nope. Makes me wonder how you got to see them if they're unpublished though

 

[Lee Scoggins]

'accuracy' as is generally understood today is usually measured by a certain set of performance parameters, and in those aspects, SS amps come out ahead.

Respectfully disagree...Some tube amps are as flat in FR or low distortion as solid state amps.

 

[Chu Gai]

Define audible. Some people will tell you that the "cold and harsh" sound that they hear from SS amps is precisely that, tiny amounts of upper order harmonics.

an example would be of interest here along with relevant distortion measurements. It is of course a complicated question bringing to bear levels and masking phenomena. However it may be more relevant to my mind to examine the tube source and the very real issues of output impedance and the resultant FR contouring that occurs. It is of course dependant upon the tube manufacturer. That could be ameliorated by using negative feedback but then tubes would lose their 'character' if you will.

Lee your thinking that there is more information on vinyl and that it is a more faithful reconstruction of the master is is without merit and is a weak footing upon which to base your preference which is being elevated to idolatry. It may behoove you to seek out recording engineers skilled in both arts or to pen your query to J. Atkinson. Various techniques known to those involved in the making of vinyl have been mentioned. Interchannel crosstalk is a real phenomena and cannot be elimated. The very real ability of vinyl to mask imperfections in the recording process due to its inherant higher noise floor is real and not imaginary. It is these quite real phenomena that bear closer examination if you're looking for the reasons why vinyl sounds different.

 

[Saurav]

Respectfully disagree...Some tube amps are as flat in FR or low distortion as solid state amps.

True, but if you average out most tube amps against most S amps, you'll probably find the tube amps to have slightly higher THD. If you take the very best of both types, I would expect them to be close.

 

[Chu Gai]

Still have a solid state unit laying around Saurav?

 

[Saurav]

All my sources (CD player and phono stage for now, DAC probably sometime in the not-too-distant future) are solid state, and probably will be for a while. My linestage is passive, and it's just easier for me to get low noise/high output/low output impedance with solid state devices. Some day when I'm bored with tinkering with other parts of my system, I'll probably look into an all-tube analog signal chain. For now, the opamps and transistors are doing just fine, and I don't want to tackle the challenge of building a tube phono stage just yet

 

[Steve_AS]

Again, the burden of proof is on you since you made the statement. All this talk about Nyquist and Shannon is great but we still don't have any solid evidence presented that 16/44 has more data than an LP!

(sigh) Lee, you're just flailing away now, and you really shouldn't use terms like 'data' if you can't do better than this. What *evidence* do you require? Refernce to a an information science or digital audio textbook? Nyquist/Shannon has been verified and accepted for *decades* now -- Nyquist first formulated it in the back in the late 1920's IIRC, when information theory was being born -- so the burden of proof that it's *wrong* about information retention/loss during digital sampling is clearly on *your* side.

Your differnces with Mr. Pinkerton's presentation of the facts -- to the very limited extent you engaged it, because you failed to comment on an *awful* lot of it -- should properly be directed to him, and man, would I like to see *that*. However, *I* will offer these rebuttals based on my more limited expertise: 1) saying that 'noise floors vary with the recording' does not speak to the *inherent physical limits* of the respecive media that Stewart laid out. 2) Your thoughts about resolution belie a lack of understanding of what resolution *means* . By the very definition of resolution, it cannot help be 'unbreakably tied' to dynanmic range (and S/N) as Stewart indicates. 3) You also fail to understand that the digital reconstruction *is* a 'continuous wave'. All of the information is *there*. THAT'S WHAT THE NYQUIST/SHANNON THEOREM MEANS. 4) As for 'problems areas' imperfect implementation is far less a 'problem' for D/A conversion than it is for tube amps, in practice.

You keep repeating audiophile folklore that have been debunked again and again -- and you'd know that if you actually posted to places like RAHE instead of places where such folklore goes generally unchallenged.

 

[Steve_AS]

Or it could be something missing, i.e. phase errors from poorly implemented analog filters in DACs

I'd call that added distortion (even if it results in 'missing' information). But in any case, it's not *inherent* to DAC technology, whereas some forms of LP/turntable/cartridge 'euphonic distortion' *are* inherent to *that* technology.

Personally, I'd rather have a system that transmits the audio signal as *accurately* as possible, in technical terms, and lets *me* do the fiddling to make it sound more subjectively 'real' if I think that's needed. (Didn't some audionauts actually market a CD player with a button you could push to make it sound like an LP, some years ago?)

 

[Saurav]

Personally, I'd rather have a system that transmits the audio signal as *accurately* as possible, in technical terms, and lets *me* do the fiddling to make it sound more subjectively 'real' if I think that's needed.

I'm curious, what kind of fiddling would you do if your system is as accurate as possible? In other words, what do you have to fiddle with? All I can think of is room treatments and speaker/subwoofer positioning. Or are you talking about tone (or other) controls? Some purists consider the mere presence of tone controls to be detrimental to the osund (and depending on how they are implemented, this assumption can be true).

I build/mod almost all of my gear, so I tweak the sound to be the way I like it by changing components, design, topologies, layout, and so on. That, in a sense, is fiddling with the sound to fine-tune it to my preferences.

Anyway, like I said, there's no right or wrong to personal preferences, and everyone has the right to prefer whatever they prefer

 

[Chu Gai]

Well Mr. Shane Morales...this is another fine mess you've gotten us into! What are your thoughts on this matter?

I say it's time for a beer. Any takers? Maybe a Bud...maybe some Busch?

 

[LanceJ]

Maybe I should have clarified myself better concerning the Nyquist Theorem.

Why is it many listeners (including myself) & pros alike report better mids/highs with 96kHz & 192kHz recordings compared to the 16/44.1 version? And most were using NON-remastered master tapes, or using their own test recordings, like this detailed MIX Magazine test report carefully describes. And I still personally believe better sound can be recreated by using more samples to capture more parts of that sound wave—it sure as hell can’t hurt.

Sorry, but I’m just getting tired of hearing Nyquist’s theorem used to completely discount what others have found from empirical research. And from doing a search, I’ve found out that while Mr. Nyquist has a great mathematical theory, when implemented in the real world, things don’t look so rosy. Quite a few things have to be done to create good sound from 44.1kHz sampling rates. 96kHz and 192kHz make things much easier, particularly concerning the filters in the DAC which actually reconstruct the analog wave itself:

http://www.earlevel.com/Digital%20Au...talAudio1.html

http://www.futureproducers.com/site/...inition/id/279

http://www.mindprint.com/english/96info.htm

As many of you know, I’m no anal-retentive listener (heck, I’m one of the few Technics/Boston Acoustics owners here); I believe in DBTs; and I don’t believe in $500 interconnects. But I’ve heard enough hi-res recordings to know something good is going on, even if it is subtle. It’s not just marketing.

LJ

 

[Saurav]

Lance, you're right, of course, and I think someone already mentioned that higher sampling rates make the D-to-A conversion easier. Also, some people believe that we respond to frequencies higher than 20KHz in some way or the other, hence the need to accurately reproduce those frequncies. However, that doesn't invalidate Nyquist's theorem - IF you accept that 20KHz is the highest frequency you need to to reproduce, THEN 40KHz is the highest sampling rate you need. And if you have theoretically perfect D-to-A converters and filters, you can recreate that 20KHz frequency perfectly too.

So... the theory is correct, but as you said, when applied to real world scenarios, some ifs and buts come in. That doesn't negate the theory though, and it's also not a case of "Nyquist's theorem doesn't apply because music is a complex waveform and not a pure sine wave".