Skeptoid: Digital vs Vinyl, Podcast casts its skeptical eye on digital audio Options

[rick.hughes]

Skeptoid: Digital vs Vinyl

[Fandango]

This thread just reminded me that I haven't listened to this episode yet! Actually I wanted to post about the episode here, too. But forgot about it.

What I can see from skimming through the transcript is that he doesn't go into all the technical details, i.e. the false claim that the dynamic range of vinyl goes beyond that of the CD is not technically refuted in all detail, if it was mentioned at all. edit: Now, that I've listened to it, I'm not so sure that vinyl and cd masters in the 80s were identical. But I have no clue actually, maybe someone can clarify that.

Nevertheless Skeptoid is really a great podcast, highly recommended. It's only ~10 minutes long so he can't get into much detail, but he still manages to pack it full with surprisingly rich content. Just look at the length of the transcripts.

But best of all is his persistence, Brian Dunning stoicly rips through every myth and conspiracy theory with ironclad scientific scepticism.

This post has been edited by Fandango: Mar 28 2012, 21:50

[Woodinville]

Hmm, we shall see how he copes with my comment that points out the effects of distortion on bandwidth and loudness (as opposed to intensity).

[Fandango]

Didn't he dodge the big question anyway? He said that because the mediums are so different there's no point in ABX'ing them. I doubt he will challenge the theoretical limitations of the vinyl format. He simply claims it's inaudible and not ABX'able anyway, so it doesn't matter. The thing that I'm sceptical about is the first one, sure the effects of vinyl cannot be properly tested with actual vinyl because there are too many giveaways that have nothing to do with the way the format works, crackle, pops and noises feed into from the turntable for instance, but what one could do is simulate the theoretical physical limitations of vinyl records along with the post-production techniques digitally and take such processed samples and compare them to the same source samples treated as if they were produced for a CD. To answer the question he raises right from the start, if the limitations of vinyl alone are even audible, I think is doable, if one can eliminate all interfering factors, from the mics to the mastering and beyond.

Of course the results will hardly be applyable to the real world, because the overall real-world setups of vinyl listening and CD listening and the real-world preparations before each mediums are pressed are as he clearly pointed out too distinct from another.

I mean isn't it unscientific to claim one can do a proper ABX test between vinyl and CD under real-world conditions and say it is likely people can hear a difference because of how vinyl works and not because of all the other factors that also shape the sound that goes in a studio mic and come out a home speaker, them being different for vinyl and CD? Well, I would be very curious if the characteristics of vinyl playback can be simulated as a whole and whether it has been done before, maybe even just for certain aspects of the format.

This post has been edited by Fandango: Mar 29 2012, 03:59

[hlloyge]

I doubt he will challenge the theoretical limitations of the vinyl format.

Aren't those limits practical limits, not just theoretical? S/N, crackling, wow&flutter...?

[Nessuno]

I doubt he will challenge the theoretical limitations of the vinyl format.

Aren't those limits practical limits, not just theoretical? S/N, crackling, wow&flutter...?

Maybe the sense is that the practical limits of common hardware prevent you to reach the theoretical ones of the format?

This post has been edited by Nessuno: Mar 29 2012, 12:11

[krabapple]

Didn't he dodge the big question anyway? He said that because the mediums are so different there's no point in ABX'ing them. I doubt he will challenge the theoretical limitations of the vinyl format. He simply claims it's inaudible and not ABX'able anyway, so it doesn't matter. The thing that I'm sceptical about is the first one, sure the effects of vinyl cannot be properly tested with actual vinyl because there are too many giveaways that have nothing to do with the way the format works, crackle, pops and noises feed into from the turntable for instance,

In the real world, an LP doesn't necessarily acquire new crackles and pops and noises between playing it once to record it digitally, and playing it again to compare it to the digital copy. So that is not necessarily a giveaway.

This post has been edited by krabapple: Mar 29 2012, 17:22

[Fandango]

Right, besides one could analyze the recording for such artefacts. In fact many aspects of vinyl playback can be analyzed and identified with software, or even simulated. But I dare think no-one has ever done an elaborate ABX test between vinyl and CDs where they have checked for vinyl playback giveaways. Would be interesting, even though I also have made my choice based on more practical reasons years ago.

[mzil]

In the real world, an LP doesn't necessarily acquire new crackles and pops and noises between playing it once to record it digitally, and playing it again to compare it to the digital copy. So that is not necessarily a giveaway.

Although there are other causes for it, audible wow in record playback due to fluctuations in the platter's speed, will occur in random locations for each playback, so one couldn't even fairly ABX an LP to the same LP (a clone with identical pops/ticks) on the same model of turntable! There's a giveaway, at least for music where the wow might be audible. [Sustained piano chords and flute are often used to look for this.]

Synchronizing the two is also a nightmare, or an LP to a CD for that matter. So even if there are indeed no new pops or ticks on a second playing of an LP, the "giveaway" problem still exists.

What I'd do is the Meyer and Moran A/D/A loop trick (where they compared SACD to their CD recorder's quick A/D/A version of it); I would take the output of the LP and digitize it, but then bring it back to analog, almost instantaneously . This gets rid of the "giveaway" problems of subsequent LP playbacks due to all these factors. The wow (including flutter) components and the pops/ticks are the same, plus the two are synchronized.

This post has been edited by mzil: Mar 29 2012, 22:46

[Woodinville]

I doubt he will challenge the theoretical limitations of the vinyl format.

Aren't those limits practical limits, not just theoretical? S/N, crackling, wow&flutter...?

There is a theoretical limit to surface noise elimination on vinyl. It's not insubstantial, either.

[knutinh]

What I'd do is the Meyer and Moran A/D/A loop trick (where they compared SACD to their CD recorder's quick A/D/A version of it); I would take the output of the LP and digitize it, but then bring it back to analog, almost instantaneously . This gets rid of the "giveaway" problems of subsequent LP playbacks due to all these factors. The wow (including flutter) components and the pops/ticks are the same, plus the two are synchronized.

Or, do a high-resolution digital recording of vinyl, play it back at high-resolution and CD-resolution in an ABX test.

-k

[knutinh]

There is a theoretical limit to surface noise elimination on vinyl. It's not insubstantial, either.

Assuming that cost and impractical conditions does not pose theoretical limits, what is this based on?

If I was to make a "vinyl" record out of any concievable material, to be played back in strictly specified lab conditions, I would expect that the granularity (and mechanical "memory") of the atomic structure of that record to be the most fundamental limit?

I think that this presentation had some interesting background information on analog media degradation, funky italian-english aside:
http://www-dsp.elet.polimi.it/ispg/images/...tion_slides.pdf

[mzil]

Or, do a high-resolution digital recording of vinyl, play it back at high-resolution and CD-resolution in an ABX test.

-k

That way wouldn't be comparing direct analog LP sound to digital, though; it would be comparing two kinds of digital sound.

[greynol]

Ignoring CDDA, if you are attempting to suggest that high-res digital is unable to capture vinyl (or the master used to create it either before or after RMAA equalization was applied) 100% transparently, what is your basis for this?

This post has been edited by greynol: Mar 30 2012, 16:03

[mzil]

Ignoring CDDA, if you are attempting to suggest that high-res digital is unable to capture vinyl (or the master used to create it either before or after RMAA equalization was applied) 100% transparently...

I'm not.
---

New topic for the general audience: I'm unable to read the original poster's link, because, at least through my browser/settings, when I click on it it brings me to a page that looks like I got the correct article, I think, however there seems to be many, many of the opening paragraphs (over a full page) completely obscured with a large, green table of contents starting with "About Skeptoid". If I click on any of the listings in it I still get this large, obscuring block, preventing me from seeing the actual page underneath. [Later, an invitation widow to subscribe also appears, however that I am easily able to close].

I don't really want to subscribe to skeptoid.com, but is that the only way to read the full article without this superimposed listing of articles? Or could some settings of my popup blocker, cookie blocking, or other settings be what's stopping me? I'm not well versed in how to deal with this sort of problem so any help would be appreciated. I am using Internet Explorer 9, if that matters.

This post has been edited by mzil: Mar 30 2012, 18:10

[dhromed]

I'm unable to read the original poster's link, because, at least through my browser/settings [...] I am using Internet Explorer 9, if that matters.

Try Compatibility Mode. It should be a broken-page icon to the right of the address bar. Activating it will be remembered for the entire skeptoid.com website.

Failing that, you can hit F12, and a utility window will appear. The top bar of that particular window shows a menu button labeled "Browsermode: XYZ". Try selecting IE8 or IE7 from that menu. This is a one-time setting and will be forgotten when you navigate away.

If nothing seems to work, proceed to quietly mutter curses at IE, and download Firefox or Chrome, which should be a five-minute task.

[greynol]

I'm not.

That's good. Perhaps it's only me, but I think sight has been lost on the target.

I feel the argument reverts to the inevitable, "as a delivery format, is there anything that vinyl can provide over CDDA that is audible to humans?" Right now it appears that one can only answer yes when the the order of the two is reversed.

[mzil]

Dhromed, Thanks! Great respone, but when I went back to the site to try your suggestions, it now loads just fine, as is! Maybe there was an oddity at their site the moment I had tried them before? Anyways, I wont complain any more, now that it works.

This post has been edited by mzil: Mar 30 2012, 21:36

[splice]

If I was to make a "vinyl" record out of any concievable material, to be played back in strictly specified lab conditions, I would expect that the granularity (and mechanical "memory") of the atomic structure of that record to be the most fundamental limit? ...

I have an extensive physical library of articles about the "vinyl" process. Being physical, it's hard to find things in it... Somewhere in it there is a quote which stuck in my mind, that at the highest recorded frequencies when the signal level approaches the noise level, the amplitude of the "wiggle" in the groove approaches the size of a hydrogen atom. I'll keep searching...

[krabapple]

Although there are other causes for it, audible wow in record playback due to fluctuations in the platter's speed, will occur in random locations for each playback, so one couldn't even fairly ABX an LP to the same LP (a clone with identical pops/ticks) on the same model of turntable! There's a giveaway, at least for music where the wow might be audible. [Sustained piano chords and flute are often used to look for this.]

Good point.

Synchronizing the two is also a nightmare, or an LP to a CD for that matter. So even if there are indeed no new pops or ticks on a second playing of an LP, the "giveaway" problem still exists.

What I'd do is the Meyer and Moran A/D/A loop trick (where they compared SACD to their CD recorder's quick A/D/A version of it); I would take the output of the LP and digitize it, but then bring it back to analog, almost instantaneously . This gets rid of the "giveaway" problems of subsequent LP playbacks due to all these factors. The wow (including flutter) components and the pops/ticks are the same, plus the two are synchronized.

Identifying A vs B by 'before/after' switch comparison requires synchronization -- and that can work great. But another way is to compare the same segment of A and B (resetting back to start at each switch), which one can do in foobar's ABX comparator, for example. This is my preferred method of comparing short but 'telling' bits of audio. It would be even more of a nightmare to apply to LP vs CD than synchronization -- and the wow/flutter problem would also pertain. ;>

This post has been edited by krabapple: Mar 31 2012, 22:20

[Woodinville]

Just want to point out that the biggest component of wow is decentering, and that will be consistent from play to play.

Having said that, playing vinyl twice inside of a short period of time is bad for the vinyl and you will not get the same signal off the record.

[mzil]

But another way is to compare the same segment of A and B (resetting back to start at each switch), which one can do in foobar's ABX comparator, for example. This is my preferred method of comparing short but 'telling' bits of audio. It would be even more of a nightmare to apply to LP vs CD than synchronization...

Actually I think that method would be impossible. Foobar2000 has no analog incoming signal source option (that is, to be used as the A or B signal) , as far as I know. Both A and B have to be digital files. Isn't that right? [I've only played around with it briefly.]

To do an ABX comparison using at least one purely analog medium, like LP, one would have to get a mechanical ABX comparator, like QSC's, or do manual swapping.

[Arnold B. Kruege...]

I doubt he will challenge the theoretical limitations of the vinyl format.

Aren't those limits practical limits, not just theoretical? S/N, crackling, wow&flutter...?

Maybe the sense is that the practical limits of common hardware prevent you to reach the theoretical ones of the format?

Many of the limits are not due to the use of common hardware, but due to the fact that even the best hardware ever made couldn't do anything about them.

[Arnold B. Kruege...]

Or, do a high-resolution digital recording of vinyl, play it back at high-resolution and CD-resolution in an ABX test.

-k

That way wouldn't be comparing direct analog LP sound to digital, though; it would be comparing two kinds of digital sound.

So what?

We already know from other experiments that reasonably good digital is sonically transparent, and the very best digital bests the thresholds of human hearing by orders of magnitude.

[Arnold B. Kruege...]

Just want to point out that the biggest component of wow is decentering, and that will be consistent from play to play.

Agreed, which leads to the obvious question:

Since decentering is often the largest source of wow, and decentering is easy enough to reduce significantly by fairly simple mechanical means, why aren't all the vinylphiles doing that?

BTW, after decentering, its probably warps and warp wow induced by the tone arm that comes next.

Then there is the FM distortion caused by the interaction of bass and offset angled tone arms (IOW not straight line).

Same question.

Having said that, playing vinyl twice inside of a short period of time is bad for the vinyl and you will not get the same signal off the record.

Agreed, but again how many vinylphiles have transcribed their records to digital to eliminate that effect? Probably a lot more LPs are transcribed for other reasons, many not even audible per se but rather related to convenience.