Jump to content
IGNORED

The problems with stereo recordings


Recommended Posts

I've been thinking a lot about stereo recordings lately. It seems obvious that stereo recording techniques have their good and bad points. Good stereo recordings, such as of a live orchestra, use multiple microphones to capture the music directly and indirectly (room resonance), but studio recordings often play around with instrument placement, making music sound artificial. Even some live electric recordings do this. I've been listening to the Grateful Dead June '76 box set recently, and, while the mixes are beautiful, Jerry's guitar is very far to the left and Keith's keyboards all the way to the right. I saw enough Grateful Dead shows to know that no one ever heard it like that. (Yes, Jerry stood on the left and Keith's piano was on the right, but the sound didn't just come from where they were on stage.)

 

This exaggerated stereo separation is prominent when I listen while working with two speakers on my desk, and is even more prominent on headphones. In a more classic setup, further from the speakers, there is less directionality, but it's still quite obvious. 

 

So I've been wondering why this has become common. Notwithstanding recordings like Sgt Pepper or Dark Side of the Moon, where the stereo effect is used as an effect, why are so many recordings mastered with this separation? Sometimes I listen to jazz piano trios where the piano is very far to the left and the bass and drums to the right, and it sounds artificial. With all this talk about wanting recordings to sound like music as it is performed (yes, I know, the catchphrase is usually "as the artist intended," but I would argue that the goal is more to make it sound like it was performed), and griping about mastering and other alterations to music, does no one care about this exaggerated use of the stereo effect?

I write about Macs, music, and more at Kirkville.

Author of Take Control of macOS Media Apps

Co-host of The Next Track podcast.

Link to post
Share on other sites
45 minutes ago, kirkmc said:

I've been thinking a lot about stereo recordings lately. It seems obvious that stereo recording techniques have their good and bad points. Good stereo recordings, such as of a live orchestra, use multiple microphones to capture the music directly and indirectly (room resonance), but studio recordings often play around with instrument placement, making music sound artificial. Even some live electric recordings do this. I've been listening to the Grateful Dead June '76 box set recently, and, while the mixes are beautiful, Jerry's guitar is very far to the left and Keith's keyboards all the way to the right. I saw enough Grateful Dead shows to know that no one ever heard it like that. (Yes, Jerry stood on the left and Keith's piano was on the right, but the sound didn't just come from where they were on stage.)

 

This exaggerated stereo separation is prominent when I listen while working with two speakers on my desk, and is even more prominent on headphones. In a more classic setup, further from the speakers, there is less directionality, but it's still quite obvious. 

 

So I've been wondering why this has become common. Notwithstanding recordings like Sgt Pepper or Dark Side of the Moon, where the stereo effect is used as an effect, why are so many recordings mastered with this separation? Sometimes I listen to jazz piano trios where the piano is very far to the left and the bass and drums to the right, and it sounds artificial. With all this talk about wanting recordings to sound like music as it is performed (yes, I know, the catchphrase is usually "as the artist intended," but I would argue that the goal is more to make it sound like it was performed), and griping about mastering and other alterations to music, does no one care about this exaggerated use of the stereo effect?

Some of the distorted image can happen because of the kind of compression/processing used on a recording.

I have a little concept/tool in my toolkit where I modify the width of the stereo image when needed.   Distortions in the stereo image are usually by ratios like sqrt(sqrt(2)), or 1/sqrt(sqrt(2)).   Sometimes these ratios are done before compression so the numbers might not be so exact.   Depending on material, I sometimes narrow the stereo image by 0.8409, 0.707 or widen it by 1.19 or 1.414.   How do I do that?   Well I use tools that do that for me -- some that I have written, some that are available as a program.  For ad-hoc testing and simple changes I use the 'SoX' program, but it has the tendancy to erase metadata in copies that it makes.

 

* Off topic:  I do have some gratis software that can bring the recordings close to their natural state -- but usage is a science project, and not really suggested unless REALLY PICKY about the sound and willing to deal with a fairly long learning curve.

 

To widen the stereo image in sox, by 1.19 one might do this:

>  sox infile.wav outfile.wav remix 1,2v1.0 1,2v-1.0 remix 1,2v1.19 1,2v-1.19

 

To narrow the stereo image in SoX by 0.707, which is relatively strong:

>  sox infile.wav outfile.wav remix 1,2v1.0 1,2v-1.0 remix 1,2v0.707 1,2v-0.707

 

I create a shell script (I use Linux 90% of the time) that effectively does this work.  There are probably 'better' programs that do the widen/narrow more explicitly without it being a 'science project'.   If you use Linux, here is my shell script: (I scale the input by 0.50 to avoid clipping and to maintain certain aspects of the signal level...

 

Script 'wide <width>'

#!/bin/bash
spread=$1
/usr/bin/sox -v 0.50 - --type wav - remix 1,2v1.0 1,2v-1.0 remix 1,2v${spread} 1,2v-${spread}

 

 

I use the script like this:

sox infile.wav --type=sox - | wide 0.8409 | sox - outfile.wav

 

It almost looks like it is almost as much typing to use the script, but remembering the sequence in the script is trickier than just remembering the simpler SoX commands.

 

John

 

Link to post
Share on other sites
50 minutes ago, kirkmc said:

What does compression have to do with stereo imaging? Imaging is essential a factor of the volume applied to each element of a recording. 

It changes the shape of the spatial relationships -- very wll known issue.   Basically, simple stereo compression flattens the image.  You still get right/left, but it flattens the image depth.   The real stereo image of recordings before the ubiquitious CD compression (orginally called the 'digital sound' back in the middle 1980s) is quite different.

 

The attack/release time of the compression used is SO FAST that it distorts the timing relationships -- therefore the subtle imaging resulting from time (the depth) is strongly compressed.   The L and R are mostly just mildly distorted because of the super high speeds and the gross amount of compression.  The compression is a VERY AGGRESSIVE scheme.  Dont even try to listen for pumping.

 

John

 

Link to post
Share on other sites
2 minutes ago, kirkmc said:

I’m afraid you’re missing the point. I’m talking about exaggerated stereo separation, which is done in mixing and mastering. I’m not talking about “flat“ imaging.

And, then refer to my suggestion about modifying the width -- you know, the width part of the discussion.   Don't be distracted by the general comments about image.   L+R are only a part of the image, and just focusing on L,R is missing a lot of what is going on WRT the whole image.

 

Timing is a significant part of the image.  

John

 

Link to post
Share on other sites

I’m not interested in discussing arcane technical matters that may or may not affect how I perceive the width of the imaging. I’m simply discussing the way certain recordings of certain instruments way off to one side. I honestly don’t care about how this is done technically, or whether it is different now than before CDs. I’m really only talking about what I hear.

I write about Macs, music, and more at Kirkville.

Author of Take Control of macOS Media Apps

Co-host of The Next Track podcast.

Link to post
Share on other sites
3 minutes ago, kirkmc said:

I’m not interested in discussing arcane technical matters that may or may not affect how I perceive the width of the imaging. I’m simply discussing the way certain recordings of certain instruments way off to one side. I honestly don’t care about how this is done technically, or whether it is different now than before CDs. I’m really only talking about what I hear.

Then -- look at my suggestion to help CORRECT the complaint that you have.   Are you just complaining, or do you want help?   BTW -- my technical comments are NOT arcane -- it is necessary to know the things that I mention to understand the issues that you might run into.

None of this is brain surgery -- and I sometimes don't understand those who are offended by information intended to help.

Link to post
Share on other sites
11 hours ago, kirkmc said:

I’m not interested in discussing arcane technical matters that may or may not affect how I perceive the width of the imaging. I’m simply discussing the way certain recordings of certain instruments way off to one side. I honestly don’t care about how this is done technically, or whether it is different now than before CDs. I’m really only talking about what I hear.

Maybe you (and one other) doesn't understand want I am talking about.   I live in a technical world -- some people don't like that world -- so I'll try to explain in non technical terms (also acknowledging the some are mixed for effect also.)

 

Almost every consumer POP CD (and a lot of Jazz and Classical) have a kind of processing added that DISTORTS the stereo image.   It is possible to clean up some of the damage, and I am NOT suggesting that for you.  It is just a matter of fact that the stereo image IS damaged one way or another.   I have seen this SAME damage on PREMIUM Telarc disks -- so unless you get something directly from Diament, or someone like that -- the recording IS damaged goods.   I can explain the real details of the processing and why the damage occurs -- but sometimes the actual technical details are not important.   Here, I am just talking about what IS.


Let me explain -- part of the stereo depth comes from the timing information.   When they apply the common CD/digital processing, it SQUEEZES the stereo image into hard left and hard right.   The little bits that are inbetween and infront are pushed fully to the side.   It is like 45degree part of the image has a hole in it.   I hear this effect on almost all recordings that should be stereo.

 

When I view the image in my mind -- the processing creates holes in the image -- front is okay, hard right and hard left are okay.  But hte bits of left and right that should be between front and side are pushed to the limits.   Sometimes there is a combination of hard effect from mastering mistakes and sometimes it is the common processing, and sometimes it is REALLY in the mix.


When you have the combination of the 'processing' and the image manipulation, a lot of recordings SOUND different after mastering than when they are mixed.

 

====

Example of the common stereo image damage.

 

For a comparison of the stereo image distortion that I am speaking of -- listen to the piano being 'hard right' on the 'Image-dist' version, whilie the piano is more forward on the 'Image-full' version.   (please ignore the frequency response difference -- not a focus of this discussion.)   There are myriad of other examples -- but the problem of distorted, overly wide stereo image -- everything RIGHT OR LEFT OR FRONT, but very little in-between is ENDEMIC to consumer digital material.

 

 

John

 

Image-dist.flac

Image-full.flac

Link to post
Share on other sites
11 hours ago, kirkmc said:

I’m afraid you’re missing the point. I’m talking about exaggerated stereo separation, which is done in mixing and mastering. I’m not talking about “flat“ imaging.

 Perhaps some of this was done originally in the Vinyl era , due to the poor channel separation of most cartridges, with around 35dB being close to as good as it gets back then ? Some may have also been due to the fascination with Quad sound back then also ? (Enoch Light etc.) 

How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 28-06-2020

Link to post
Share on other sites

Interesting that this whole conversation is about left and right - what happened to back and front? 😉

 

Much stereo playback is flat as tack - there is hardly any perceptible depth to it ... it could be like watching an old Cinerama move, when sitting close to the screen. Which is something that did happen to me ... I saw "It's Mad, Mad, etc ... World" in the very front row, 'cause I was last in - my neck muscles got a solid workout, as I had to swing my head left to right, and back again, constantly, to follow everything 😁.

 

Good playback doesn't have left and right, and middle - it has soundstage, in the true sense of the word. Everything in the recording occupies part of the space from the line of the speakers back to as far as the acoustic cues tell one where things are - in figures, this could be hundreds of square feet of area where various sounds in the mix present as originating from - the left and right aspect of things barely rates a mention, as being subjectively of great importance.

Frank

 

http://artofaudioconjuring.blogspot.com/

 

 

Over and out.

.

 

Link to post
Share on other sites
12 minutes ago, sandyk said:

 Perhaps some of this was done originally in the Vinyl era , due to the poor channel separation of most cartridges, with around 35dB being close to as good as it gets back then ? Some may have also been due to the fascination with Quad sound back then also ? (Enoch Light etc.) 

Separation at 20dB or so doens't make such a huge difference.   The problem isn't always mixing, but is more in mastering and distribution changes.  Sometimes it IS in mixing thoough.    Quad processing can do WEIRD things though.

 

* We consumers are usually NOT getting what was mixed in the studio, but a very damaged copy.   Along with dynamic range, the stereo image is the next to go.

 

Listen to the difference of the two versions of the SAME mixing, but different LAST STEP in mastering, that I showed above.   That difference is much more than just cartridge, unless we are talking 10dB or less of separation.   The version most likely on original vinyl is more likely close to the 'full' version than the 'dist' version...   The 'full' nor 'dist' version have NEVER touched vinyl.

 

John

 

Link to post
Share on other sites
5 minutes ago, fas42 said:

Interesting that this whole conversation is about left and right - what happened to back and front? 😉

 

Much stereo playback is flat as tack - there is hardly any perceptible depth to it ... it could be like watching an old Cinerama move, when sitting close to the screen. Which is something that did happen to me ... I saw "It's Mad, Mad, etc ... World" in the very front row, 'cause I was last in - my neck muscles got a solid workout, as I had to swing my head left to right, and back again, constantly, to follow everything 😁.

 

Good playback doesn't have left and right, and middle - it has soundstage, in the true sense of the word. Everything in the recording occupies part of the space from the line of the speakers back to as far as the acoustic cues tell one where things are - in figures, this could be hundreds of square feet of area where various sounds in the mix present as originating from - the left and right aspect of things barely rates a mention, as being subjectively of great importance.

A lot of recordings DO have depth, until they are mastered or prepared for distribution.   The difference can be small, or the difference can be profound.  But there can be a LOT of damage to the stereo in the last step before consumer copy.

 

 

Link to post
Share on other sites
1 minute ago, sandyk said:

 Speak for your own setup Frank .😋

 

These days it's quite remarkable what $AU500 worth of kit can deliver - sound almost completely lifts out of the speakers, they only reveal themselves when fairly close - depth could be better, but the word "flat" is the last adjective one would use ...

Frank

 

http://artofaudioconjuring.blogspot.com/

 

 

Over and out.

.

 

Link to post
Share on other sites
3 minutes ago, John Dyson said:

A lot of recordings DO have depth, until they are mastered or prepared for distribution.   The difference can be small, or the difference can be profound.  But there can be a LOT of damage to the stereo in the last step before consumer copy.

 

 

 

Damage could always be done to the mix, but I find it hard to believe that this has been a universal, intentional practice.

Frank

 

http://artofaudioconjuring.blogspot.com/

 

 

Over and out.

.

 

Link to post
Share on other sites

I think you are comparing a large venue where the goal is simply to make sure everyone hears everything and stereo image is secondary whereas in the studio it can be a much more intimate setting and having hard left or right panning could be what they heard in the studio.  Or it could be a ham fisted engineer.  Honestly I’ve never been impressed with Greatful Dead recordings but I’m not a fan either so I might not have paid as close attention as I should have.

Link to post
Share on other sites
4 hours ago, John Dyson said:

Separation at 20dB or so doens't make such a huge difference.  

 It may not in your setup, but with the dual mono type construction (even the PSU area) used in my DIY gear which has way better than normal channel separation, the differences are very obvious with good material.

How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 28-06-2020

Link to post
Share on other sites
2 minutes ago, sandyk said:

 It may not in your setup, but with the dual mono type construction (even the PSU area) used in my DIY gear which has way better than normal channel separation, the differences are very obvious with good material.

I use headphones most of the time -- separation is about as obvious as possible.

Anyway -- the MAJOR differences that the OP was speaking about wouldn't be manifested by 25dB vs 20dB of separation.

 

He was talking about differences on the level of matrixing differences or worse (an actual mix.)   However, many of those differences can be created by the kind of example that I showed -- almost pure left and pure right vs a mix that gives a spatial relationship of sorts.  (My example was NOT perfect, as I ran it off in less than 1-2 minutes -- better examples can be shown with other material.  I just happened to be in my Brasil'66 directory at the time.)

 

So, the cartridge differences might be 20 vs 25dB, but the differences that we can easily see from other sources (mixing, compression, etc) can effectively be similar to creating fake separation or losing a *big* part of the stereo image.   The extreme amount of damage is a LOT more than can be manifested by almost any properly working cartridge.


Sure, on a more fine level -- 15dB vs 25dB can make a difference, and pretty sure that I could detect it.

As it is -- my digital setup probably has 90dB separation.

 

John

 

Link to post
Share on other sites
13 minutes ago, John Dyson said:

 


Sure, on a more fine level -- 15dB vs 25dB can make a difference, and pretty sure that I could detect it.

As it is -- my digital setup probably has 90dB separation.

 

John

 

And it's more likely around 70dB at 20kHz 😉

How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 28-06-2020

Link to post
Share on other sites
5 minutes ago, sandyk said:

And it's more likely around 70dB at 20kHz 😉

I cannot hear 20kHz :-).   Geesh, who can hear -70dB at 20kHz? :-).

 

BTW -- I just learned BY EXPERIENCE, not metaphysics, something about the audibility of distortion -- you'd like the hear about it.   Not appropriate to go into details here.   Still trying to figure out the details.  I just might be wrong (you know how difficult/frustrating that doing valid experiments can be.)

 

John

 

Link to post
Share on other sites
51 minutes ago, John Dyson said:

cannot hear 20kHz :-).   Geesh, who can hear -70dB at 20kHz? :-).

 Nether can I, but I can hear the difference between genuine Hi Res material such as from Barry Diament and other well recorded albums where most of the genuine musical content of >20kHz is well down, and 16/44.kHz as I have reported back  to you on quite a few occasions where you have made available the >48kHz versions of the same files. So has fellow A.S. member Chris with the same material.

How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 28-06-2020

Link to post
Share on other sites
5 hours ago, sandyk said:

 Nether can I, but I can hear the difference between genuine Hi Res material such as from Barry Diament and other well recorded albums where most of the genuine musical content of >20kHz is well down, and 16/44.kHz as I have reported back  to you on quite a few occasions where you have made available the >48kHz versions of the same files. So has fellow A.S. member Chris with the same material.

It is really great (really, truly) that there is some material that has some detail above 20kHz.   Most material just has noise modulation WRT NR systems (which can be more mitigated than I thought), and we can discuss that offline.   On most pop material, the vast majority of the spectogram ghosts above 20k are just that -- 'ghosts' or NR noise modulation.

When there REALLY IS good material that reaches above 20k, that is a good thing but on limited material.

 

John

 

Link to post
Share on other sites
9 hours ago, Killahertz said:

I think you are comparing a large venue where the goal is simply to make sure everyone hears everything and stereo image is secondary whereas in the studio it can be a much more intimate setting and having hard left or right panning could be what they heard in the studio.  Or it could be a ham fisted engineer.  Honestly I’ve never been impressed with Greatful Dead recordings but I’m not a fan either so I might not have paid as close attention as I should have.

 

I'm comparing any venue of pretty much any size (other than very small), where no one really hears individual instruments on one side or the other, with the approach of mixing discrete channels to a sided dominance. Grateful Dead recordings are very well mastered, but there is this odd tendency to isolate some of the instruments. But they're far from the only ones: some Brad Mehldau live recordings are like that, with the piano way off to the left, and there are plenty of others like that. Again, this is not about a conscious choice in the studio, such as by Pink Floyd, but rather a choice when mixing live recordings. 

I write about Macs, music, and more at Kirkville.

Author of Take Control of macOS Media Apps

Co-host of The Next Track podcast.

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×
×
  • Create New...