[old tech]

I was wondering if anyone from a technical or production background can resolve a discussion I had with a colleague regarding stereo imaging and the extent that it can be influenced by a recording format.

The claim is that LP playback imparts a better stereo image and a wider soundstage than CDs or digital playback generally. I understood that imaging is more of a product of mixing and mastering engineers, though of course speakers and their positioning has a large effect. However all things equal I would have thought digital playback would have superior imaging because of the virtual lack of cross-talk across L and R channels. If anything, vinyl playback would be worse than tape due the bleed between channels.

My colleague claims that there must be something more to it as he believes his vinyl playback provides better imaging. He is not the only one that claims this, I remember a thread on the Steve Hoffman site where many of their members claimed the same thing, though some of the claims are way over the top, eg 3D, holographic etc.

The thing is I also notice this effect with some vinyl records but I don’t think it is real in the sense of comparing it to the pin point imaging accuracy of digital playback. My theory (I really don’t know) of what is going on is that the wide soundstage of vinyl playback is just the effect phase shifts, which is an unavoidable artefact of vinyl playback. I don’t believe it is anything which the producer intended or what exists on the source masters.

Any thoughts or explanations?

 

[bigshot]

How could poorer crosstalk create better soundstage? That seems like a no brainer to me. If someone wants to claim that LP records have better channel separation / phase, I would ask them what planet they are residing on.

 

[Strangelove424]

I can't think of a single technical superiority to vinyl. Whatever benefits there are, if any, have to do with mastering decisions made to suit intended markets. And separating the format from the mastering for comparison purposes is impossible without your own custom pressings.

When they choose to take full advantage of CD, it's a wonderful experience.

 

[71 dB]

Horizontal movement of stylus = channel difference L - R
Vertical movement of stylus = channel sum L + R

Horizontal movements are physically different from vertical movement => horizontal movements cause different kind of distortion than vertical movement. In other words the mono part of music is distorted differently from the "stereo/side" part. This might cause an illusion of improved sound imaging.

Elliptic filtering used with vinyl to reduce horizontal stylus movements at low frequencies is effectively cross-feeding which allows better imaging when listened with headphones. Yes, reduced channel separation can benefit imaging, in this case by reducing spatial distortion which renders imaging worse.

 

[ev13wt]

Since playing back vinyl requires physical manipulation, often the listening is also more focused. This is equated to better sound.

Not many people can evaluate sort of objectively.

 

[old tech]

Horizontal movement of stylus = channel difference L - R
Vertical movement of stylus = channel sum L + R

Horizontal movements are physically different from vertical movement => horizontal movements cause different kind of distortion than vertical movement. In other words the mono part of music is distorted differently from the "stereo/side" part. This might cause an illusion of improved sound imaging.

Elliptic filtering used with vinyl to reduce horizontal stylus movements at low frequencies is effectively cross-feeding which allows better imaging when listened with headphones. Yes, reduced channel separation can benefit imaging, in this case by reducing spatial distortion which renders imaging worse.

So is it that the one stylus is being compromised through trying to track both L and R channels which causes phase shifting between the channels? That could explain the illusion of a wider sound stage.

I noticed something similar when playing my Moody Blues Days of the Future Past CD the other night. There is one part on The Night (nights in white satin) track where for a split second a sound seems to float past the right hand wall. Sounds impressive but in this case, given it is a CD, I think it is a glitch in the recording.

 

[bigshot]

To be fair, a well cut LP and a well aligned cartridge will have channel separation of somewhere around 30 dB, which although not spectacular, is probably enough to do the job pretty well.

 

[castleofargh]

So is it that the one stylus is being compromised through trying to track both L and R channels which causes phase shifting between the channels? That could explain the illusion of a wider sound stage.

I noticed something similar when playing my Moody Blues Days of the Future Past CD the other night. There is one part on The Night (nights in white satin) track where for a split second a sound seems to float past the right hand wall. Sounds impressive but in this case, given it is a CD, I think it is a glitch in the recording.

how could we assume anything about the cause of a subjective impression given everything that is in nature different with vinyls? distos of many sorts reach audible levels. crosstalk does too. even the noise floor can usually be audible on silent or quiet passages. the signature might not be the same for various reasons, which will change the perceived image. the stuff often done to the low end or the trebles, how non linear the trebles can be in amplitude, and of course there are all the times when it's simply a different mastering.

those are only the objective stuff, trying to look into how we adapt to, and interpret those changes is going to lead us to a long list of maybes, but I don't see how we can hope for more than guesses that will stay guesses. someone could decide to undergo the gruesome challenge of testing each variable independently while trying to recreate something as close as what happens to that variable on vinyls, but given how most variables measure at an audible levels of low fidelity, it seems reasonable to expect that studying only one variable will not inform us about what we're getting with the all package of low fidelity.
for all I know the crosstalk coming from a mechanical movement, isn't constant and creates a specific signature that the brain might in some ways identify as HRTF cues?
or the brain senses something too mono and decides to expend it the same way I get used to crossfeed after a moment and feel again like I'm getting full panning? who knows when so much wrong is going on all at once. we know a little about normal interpretation of audio cues, but we have pretty much no data about the sort of wrong that can still feel right or how the brain can get better convinced by one sort of bad instead of another one. those are questions for Walter Bishop and Fringe division.

 

[spruce music]

You might also think about the fact one can digitally record an LP, and the sound is that of LP. The imaging differences, which are real, are captured in the digital recording. Which means LP imaging can incidentally create some pleasing imaging artefacts. Same thing can happen with tube power amps. They really can sound more dynamic and 3D in imaging, but it is a coloration that happens to sound great to humans. Not a case of increased fidelity.

 

[castleofargh]

of course, I'm just saying that too many variables are impacted at an audible level, to just take one aside and check if "that's it".

 

[pinnahertz]

Horizontal movement of stylus = channel difference L - R
Vertical movement of stylus = channel sum L + R

You have it exactly inverse. lateral is L+R, vertical is L-R.

Horizontal movements are physically different from vertical movement => horizontal movements cause different kind of distortion than vertical movement. In other words the mono part of music is distorted differently from the "stereo/side" part. This might cause an illusion of improved sound imaging.

Nope. That's not what happens, and that is completely wrong. If that were true there would be almost no channel separation at all.

Elliptic filtering used with vinyl to reduce horizontal stylus movements at low frequencies is effectively cross-feeding which allows better imaging when listened with headphones. Yes, reduced channel separation can benefit imaging, in this case by reducing spatial distortion which renders imaging worse.

Again...your "elliptic filtering" is largely myth, probably comes from a rather poorly researched wiki article. An elliptic filter is simply a type of filter, just like a Chebyshev or Butterworth filter, but has little to do with the actual mechanism of bass summing. The function with the legend "elliptic filter" appeared on a mastering console or two, and has been absorbed into audio mythology as something standard. It's not, it wasn't, it isn't.

And since the issue is that vertical modulation has hard limits of the aluminum substrate of the lacquer and the top surface of the lacquer, vertical modulation must be more controlled, which is why it's used for L-R in the first place. However, there is no independent vertical/l-r processing in mastering, and the use of elliptic filter is completely incidental to the process anyway. Low frequency summing is not done in mastering, it's done in the mix, and typically not done by summing, just done by panning bass-heavy modulation tracks to the center.

No, reduced separation doesn't benefit imaging. Channel separation in vinyl is frequency dependant...highly. Mostly it's pretty bad especially at the high end.

Vinyl doesn't have better imaging in and of itself. Any perception of that comes from either specific mastering choices or the extremely powerful expectation bias coupled with the owning, handling and playing of vinyl.

 

[old tech]

You have it exactly inverse. lateral is L+R, vertical is L-R.

Vinyl doesn't have better imaging in and of itself. Any perception of that comes from either specific mastering choices or the extremely powerful expectation bias coupled with the owning, handling and playing of vinyl.

So phase shifts between L and R channels do not impact imaging, artificially and random as it may be?

 

[pinnahertz]

So phase shifts between L and R channels do not impact imaging, artificially and random as it may be?

Phase shifts between channels could affect imaging but that degree of phase shift doesn't exist in any professional recording system, analog or digital. Loss of channel separation in vinyl also does not entail a large degree of phase shift.

 

[71 dB]

You have it exactly inverse. lateral is L+R, vertical is L-R.

Thanks! Checked it out. you are right.

Nope. That's not what happens, and that is completely wrong. If that were true there would be almost no channel separation at all.

My self-confidence is gone…

 

[Arpiben]

Phase shifts between channels could affect imaging but that degree of phase shift doesn't exist in any professional recording system, analog or digital. Loss of channel separation in vinyl also does not entail a large degree of phase shift.

By curiosity phase shifts between L&R are kept under which value, 90 degrees?
Some studies dealing with sound localization cues often provide IPD=90 as Interaural Phase Difference boundary, at least at around f=850Hz single tone or signal envelope (ITD...Burghera & al 2013).