[safulop]

OK now, since I'm actually interested in knowing people's preferences and why they turn out that way, I've decided to turn from the theoretical arguments about why vinyl might sound superior to digital, and start up a little contest and see if it actually does sound better.  Some folks around here have posted that CDs were a gift from the Gods and they sold all their records shortly after.  I beg to differ. Eventually I'd like to publish a study on this, although audio listener preferences is not really my usual area of research.
 
To kick us off, I am posting links to two sound files of Beethoven's 5th Symphony, 1st movement.  I figure it's public domain music and I am not even posting the complete work so I probably won't get sued.  Luckily I happen to have both a CD and a vinyl record of the famed 1962(?) recording by Herbert von Karajan and the Berlin Philharmonic.   The CD was released in the 1990s as a digitally remastered reissue, part of a box set of Beethoven's symphonies.  The vinyl was released very recently, it is some kind of direct metal master cut from the original master tapes.  I won't tell you which is which, I assume people could figure out which is the vinyl.  I did my best to level match the files, but it was not that easy because the two masters are not equivalent so they have different dynamic ranges.
 
Anyway here are the links, happy listening and I am eager to hear everyone's opinion on this.  For the real study I plan to put a survey on QuestionPro but since this isn't my usual research I don't really know what ought to be asked.
 
 
http://zimmer.csufresno.edu/~sfulop/Vinyl_vs_digital/Beethoven5th_A.wav
 
http://zimmer.csufresno.edu/~sfulop/Vinyl_vs_digital/Beethoven5th_B.wav

 

[arnyk]

Anyway here are the links, happy listening and I am eager to hear everyone's opinion on this.  For the real study I plan to put a survey on QuestionPro but since this isn't my usual research I don't really know what ought to be asked.
 
 
http://zimmer.csufresno.edu/~sfulop/Vinyl_vs_digital/Beethoven5th_A.wav
 
http://zimmer.csufresno.edu/~sfulop/Vinyl_vs_digital/Beethoven5th_B.wav

 
Figuring out which is the vinyl by ear is pretty easy as the following spectral analysis of the first few seconds of each demonstrates:
 

 
Beethoven - LP
 

 
Beethoven CD
 
 
The noise in the LP is a disappointingly and clearly audibly just 27 dB below reference level, while the CD is nothing special, but is still at least 20 dB more quiet.
 
This is why about just 99% of everybody abandoned the LP tarting about 30 years ago.

 

[nick_charles]

  OK now, since I'm actually interested in knowing people's preferences and why they turn out that way, I've decided to turn from the theoretical arguments about why vinyl might sound superior to digital,
 
You are asking the wrong question - what you must ask for is preference 'sounds superior' is meaningless since it is just an opinion, superior in measurable terms is a valid question but we already have plenty of data on that for both perspectives
 
To kick us off, I am posting links to two sound files of Beethoven's 5th Symphony, 1st movement.  I figure it's public domain music 
 
The music may not be copyrighted  but the recording very well may be due to the bizarre 1996 GATT provisions and US anti-piracy laws meaning that in certain states even non-us recordings between 1923 and 1972 cannot be public domain until 75 years after publication even if they had become public domain in the original country of publication,  go figure...
 
But there is also fair use which give some slack if used for educational non-commercial purposes but when youtube hound people for 30 second clips of their brats dancing to a song in the background.....
 
I used to design surveys for a living I can help you with the one question you need 

 

[Leporello]

What are the possible outcomes of this study? Here we have two slightly different sounding snippets. I might prefer snippet A over B -  or vice versa. Then what? It is already a well known fact that many people prefer the sound of vinyl just as there are many who prefer cd. That fact in itself is not a great scientific mystery.
 
Even if 99 % of the population preferred the vinyl snippet it would only show that plain jane digital is perfectly capable of reproducing all the glories of vinyl sound.
 
Snippets are quite easy to tell apart. Tiny pops and crackles are easily heard on the vinyl version. There is also a certain amount of pre-echo on the vinyl snippet.
 
FWIW, the cd version seems to sound a little softer than my own cd copy of the Karajan -62. 
 
 
Regards,
 
L.

 

[XenHeadFi]

 
Snippets are quite easy to tell apart. Tiny pops and crackles are easily heard on the vinyl version. There is also a certain amount of pre-echo on the vinyl snippet.
 

And the vinyl is faster and has obvious wow and flutter. I guess the wow and flutter emulates vibrato that was not placed there by the musician.

 

[safulop]

  And the vinyl is faster and has obvious wow and flutter. I guess the wow and flutter emulates vibrato that was not placed there by the musician.

The vinyl is faster -- an interesting subjective impression because it's incorrect. **Edit -- the vinyl is indeed 3.4 seconds shorter**   I don't hear any wow an flutter; obviously if there is flutter it is due to the mastering process since a belt-drive turntable cannot introduce flutter.

 

[safulop]

  What are the possible outcomes of this study? Here we have two slightly different sounding snippets. I might prefer snippet A over B -  or vice versa. Then what? It is already a well known fact that many people prefer the sound of vinyl just as there are many who prefer cd. That fact in itself is not a great scientific mystery.
 
Even if 99 % of the population preferred the vinyl snippet it would only show that plain jane digital is perfectly capable of reproducing all the glories of vinyl sound.
 
Snippets are quite easy to tell apart. Tiny pops and crackles are easily heard on the vinyl version. There is also a certain amount of pre-echo on the vinyl snippet.
 
FWIW, the cd version seems to sound a little softer than my own cd copy of the Karajan -62. 
 
 
Regards,
 
L.

I thought that what is "a well known fact" is simply that some people prefer vinyl because they like to play records, or they have dumb ideas about digital sound, etc.  I don't believe it is established what proportion of people, when confronted with a forced choice, would actually assert that "this vinyl track sounds better to me than the corresponding digital track."  Personally I think that compared to the vinyl on this one the CD track sounds hideous.  I will never listen to it again.  And there are technically measurable reasons for this.  The vinyl has a much greater dynamic range, due to the CD master being compressed.  Always the difference will come from different treatments of the master, and so the idea is to pursue the question, does vinyl really sound better because the masters are better?  In this case my answer is, yes it does.

 

[Leporello]

  I thought that what is "a well known fact" is simply that some people prefer vinyl because they like to play records, or they have dumb ideas about digital sound, etc.  I don't believe it is established what proportion of people, when confronted with a forced choice, would actually assert that "this vinyl track sounds better to me than the corresponding digital track."  Personally I think that compared to the vinyl on this one the CD track sounds hideous.  I will never listen to it again.  And there are technically measurable reasons for this.  The vinyl has a much greater dynamic range, due to the CD master being compressed.  Always the difference will come from different treatments of the master, and so the idea is to pursue the question, does vinyl really sound better because the masters are better?  In this case my answer is, yes it does.

To my ears neither of them sounds particularly good compared to many modern digital recordings. Perfectly enjoyable, though. Do you have any dynamic range measurements wrt these two versions? I would be hard pressed to say that the cd version has been compressed.
 
I somewhat share the impression that the vinyl version might be faster, like it plays at higher pitch. But I am not sure.

 

[safulop]

Arny can you tell me how you did the spectrum?  Because the long-term average spectra I'm seeing look like this here.  This shows the power spectra of the first 6 seconds of music, essentially the first 'da da da duh' .
 
Which spectrum is which in your viewpoint?  As you can see here they have almost identical music spectra.
 

 

[safulop]

  To my ears neither of them sounds particularly good compared to many modern digital recordings. Perfectly enjoyable, though. Do you have any dynamic range measurements wrt these two versions? I would be hard pressed to say that the cd version has been compressed.
 
I somewhat share the impression that the vinyl version might be faster, like it plays at higher pitch. But I am not sure.

Yes here are the waveform plots of the entire movement, CD on the top.  You can see that it doesn't preserve the crescendo toward the end, it has all been "ramped up" to equal volume throughout.
 

 

[XenHeadFi]

  The vinyl is faster -- an interesting subjective impression because it's incorrect.  The two musical tracks are approximately 429 seconds in length, give or take a couple of tenths, so you couldn't possibly hear a speed difference.  I don't hear any wow an flutter; obviously if there is flutter it is due to the mastering process since a belt-drive turntable cannot introduce flutter.

The long lead track is the vinyl based on the pops and clicks. Trying to ABX the 2 files, the only place they match up is at the end. So long lead before music starts and syncing up towards the end means the vinyl is faster by that extra lead time. Yeah, I cannot hear that change in pitch as I am not pitch perfect, but knowing is going to bug me like it did with cassette tapes.
 
The CD mastering is definitely clipping. ABX around 25 s and was 10/10.
 
I was wrong about the wow/flutter. I could not ABX it on the first try 6/10.
 
I think both choices are "poison": Pops & Clicks/Bad Timing/Good Mastering vs. Bad Mastering. My choice would be digital with a better master.

 

[safulop]

  The long lead track is the vinyl based on the pops and clicks. Trying to ABX the 2 files, the only place they match up is at the end. So long lead before music starts and syncing up towards the end means the vinyl is faster by that extra lead time. Yeah, I cannot hear that change in pitch as I am not pitch perfect, but knowing is going to bug me like it did with cassette tapes.
 
The CD mastering is definitely clipping. ABX around 25 s and was 10/10.
 
I was wrong about the wow/flutter. I could not ABX it on the first try 6/10.
 
I think both choices are "poison": Pops & Clicks/Bad Timing/Good Mastering vs. Bad Mastering. My choice would be digital with a better master.

With you on the clipping!  I thought those loud parts were too loud.  But the waveform does not show "digital clipping" so I'm not sure how it got introduced.  This CD is labeled "ADD", which was supposed to mean that it was digitally re-mixed from the multitrack masters.  Anything could have happened.
 
On the timing you're still not with me; I measured the length from first note to last using audio editing software; it is 429.x seconds in each, so don't worry about it
 
And I agree that a better digital master on a CD would be the best.  But the point of my study will be to show that a great proportion of recordings from across history, selected essentially at "random," will have lousy digital masters in comparison to their vinyl counterparts, and therefore a great proportion of naive listeners could well agree.

 

[XenHeadFi]

 
On the timing you're still not with me; I measured the length from first note to last using audio editing software; it is 429.x seconds in each, so don't worry about it
 
And I agree that a better digital master on a CD would be the best.  But the point of my study will be to show that a great proportion of recordings from across history, selected essentially at "random," will have lousy digital masters in comparison to their vinyl counterparts, and therefore a great proportion of naive listeners could well agree.

Line them up from the beginning. The vinyl is a full second off at 2:45. By the end of the piece it is off by 3 seconds.
 
Well, if the point is that vinyl is usually mastered less hot than CDs, I don't think you will find very many people that will argue against you. So the choice still becomes mechanical issues versus bad mastering.

 

[safulop]

  Line them up from the beginning. The vinyl is a full second off at 2:45. By the end of the piece it is off by 3 seconds.
 
Well, if the point is that vinyl is usually mastered less hot than CDs, I don't think you will find very many people that will argue against you. So the choice still becomes mechanical issues versus bad mastering.

My bad!  Wasn't reading my instruments correctly.  You're right, the vinyl is 3.4 s less in overall length.  But since you can't hear the difference why would it bother you?  It seems strange to lead with your logic when we're concerned about subjective sound.

 

[safulop]

My intensity measurements show the CD track has about 15 dB less noise floor; the noise there is due to tape hiss.  So overall the vinyl has only about 30 dB dynamic range in practice from loudest peak to noise floor -- pretty amazing that that is sufficient to get an impression of great sound.