Jump to content
IGNORED

What measurements correlate with X?


Recommended Posts

can someone provide background or links regarding which types of measurements correlate with some specific attributes of stereo?

In particular: what helps/harms the image stability and specificity of instruments/vocals?  when i hear (or think i hear) an instrument in a certain location (relative to width or depth of the music), is there a measurement that correlates with this?  or the decay of a note or a vocal line as it diminishes?

 

if i were to speculate, i wonder if something in the correlation between left and right channels is critical?  is correlated noise/distortion worse than random noise/distortion in a stereo reproduction?

 

cheers

 

Link to comment

One problem with these types of measurements, is finding hi-fi components in (ears only) tests that have these attributes.  Some audiophiles are quick to say that this unit has this and that unit has that. But they are never interested in demonstrating those differences.

Now loudspeakers in rooms are another ballgame.

And components with euphonic colorations are yet another ballgame.

Link to comment
3 hours ago, cat6man said:

can someone provide background or links regarding which types of measurements correlate with some specific attributes of stereo?

In particular: what helps/harms the image stability and specificity of instruments/vocals?  when i hear (or think i hear) an instrument in a certain location (relative to width or depth of the music), is there a measurement that correlates with this?  or the decay of a note or a vocal line as it diminishes?

 

if i were to speculate, i wonder if something in the correlation between left and right channels is critical?  is correlated noise/distortion worse than random noise/distortion in a stereo reproduction?

 

cheers

 

have you read

 

Sound Minds Mind Sound

 

 

Link to comment
29 minutes ago, Audiophile Neuroscience said:

 

My understanding is that soundstaging, where the image is on the soundstage, is determined by the recording. The quality and qualities of the image is determined by the room/speaker interaction. Eg Omnidirectional speakers in a highly reflective room produces huge images locked between the speakers. The more common directional speakers produce more focused  realistically sized images, provided room early reflections are absorbed.

 

I agree that the soundstaging is determined by the recording, but not that quality and qualities of the image are determined by the room/speaker interaction.

 

29 minutes ago, Audiophile Neuroscience said:

 

Our perception of the location of the stereo central phantom image relates to amplitude and delay panning (with phase also affected but differently for each frequency). Psychoacoustically the brain perceives that the image is located away from the time delayed speaker and towards the not time delayed speaker and similarly for loudness differential. Delay will result from the difference in the paths between listener and each speaker. Amplitude will change also and in this respect the angle we have to the speaker is important and influenced by the directivity of the speaker (of highs and mids)  and baffle shape.

 

Yes.

 

29 minutes ago, Audiophile Neuroscience said:

 

Soundstage and central image can be made to move in concert as you move left or right, which is to say it follows the listener. Each position will have its own amplitude and delay panning effect, a blend of left and right channels.

 

Yes.

 

29 minutes ago, Audiophile Neuroscience said:

 

With lots of early reflections you will have a larger image and sound stage and the image will occupy a larger part of that sound stage. It will float between the speakers and tend to stay fixed irrespective of whether you get up and move around the room. This is how more omnidirectional speakers work in reflective rooms. In most acoustically treated rooms which dampen first reflections, when you get up and move to the left or right the image and stage shift smoothly to the left or right. In problem rooms, the image collapses to the loudspeaker that is closest to you.

 

This is where I strongly depart ... the floating behind the speakers and tending to stay fixed irrespective of whether you get up and move around the room is a function of the SQ emerging from the speakers; reflections from within the room may help in some borderline SQ situations, but aren't necessary.

 

The "when you get up and move to the left or right the image and stage shift smoothly to the left or right. In problem rooms, the image collapses to the loudspeaker that is closest to you." situation is not problem room dependent - rather, doing a lot of things with room acoustics is a workaround to diminish the impact of less than stellar SQ, giving the brain a better chance of making sense of the presentation.

Link to comment
18 minutes ago, Audiophile Neuroscience said:

Yep, we disagree.

 

soundstaging, where the image is on the soundstage, is determined by the recording. Obviously, the recording has to have the cues in the first place, so to that extent yes, sound "quality" in the recording is important. The rest is a function of your speakers and their interaction with the room.

 

It was first described to me reading Art Noxon's (Master of Science degrees in Mechanical Engineering/Acoustics and Physics). Try it, it works 

One thing that really mangles soundstaging is stereo dynamic range compression being used on an entire recording.  I'd suspect that problem is seldom recognized, but it doesn't do the kind of damage that one might initially think...  Usually, dynamic range compression on an entire recording is done with L and R getting the same gains.   When doing the DRC, it tens to flatten (decrease 'height') of the image, and makes a physically impossible image.  I do know that the compression has to be pretty fast to really flatten it, but I can hear the flattening  on a certain 1-20msec attack, 20-40msec release compression system.   Since the system that I am speaking of does normally compress L and R separately, the scheme used for FA actually does the compression so that ch1 is L+R and ch2 is L-R instead of the more normal DA operation of ch1 being L and ch2 being R.   So, in this case of FA, it is similar to what would happen on a normal stereo DRC.

 

Of course, I am speaking of 'stereo' DRC, and not DRC that might be used on each item in a mix.  (Of course, in a complex pop music mix, reality is rather arbitrary anyway :-)).

 

John

 

Link to comment
15 minutes ago, John Dyson said:

One thing that really mangles soundstaging is stereo dynamic range compression being used on an entire recording.  I'd suspect that problem is seldom recognized, but it doesn't do the kind of damage that one might initially think...  Usually, dynamic range compression on an entire recording is done with L and R getting the same gains.   When doing the DRC, it tens to flatten (decrease 'height') of the image, and makes a physically impossible image.  I do know that the compression has to be pretty fast to really flatten it, but I can hear the flattening  on a certain 1-20msec attack, 20-40msec release compression system.   Since the system that I am speaking of does normally compress L and R separately, the scheme used for FA actually does the compression so that ch1 is L+R and ch2 is L-R instead of the more normal DA operation of ch1 being L and ch2 being R.   So, in this case of FA, it is similar to what would happen on a normal stereo DRC.

 

Of course, I am speaking of 'stereo' DRC, and not DRC that might be used on each item in a mix.  (Of course, in a complex pop music mix, reality is rather arbitrary anyway :-)).

 

John

 

 

Thanks John, I didn't know that but will listen to see if I can perceive it.

 

On a related topic there was a huge debate on AS as to whether height could be encoded on a stereo recording in the first place. With a conventional microphone, an object moving sideways or vertically is simply moving away from the mic and it cannot encode anything but  L-R differential response. I always pondered that there must be something, some cue in the signal that subserves the illusion of height perception.Perhaps dynamic range??

Sound Minds Mind Sound

 

 

Link to comment
12 minutes ago, opus101 said:

When an object is moving away from the microphone what changes is the ratio of direct to reflected sound. So I'd guess if there's a floor bounce there's a chance that height could be decoded from the relative amplitude of direct vs reflected from the floor?

 

Yeh, I wondered that too, but I guess that it still only tells you distance away not height......I found mansr' original post...

<quote>

Two microphones in, say, a standard XY stereo configuration clearly separate sounds on a left/right axis. This is so blatantly obvious that I neglected to explicitly mention it in that post you refer it. Picture a sound source at centre, some distance directly in front of such a microphone pair. The sound arrives at both microphones 45° off-axis, giving equal response from both. If the source moves sideways, it comes closer to the axis of one microphone and further off-axis of the other. As these are cardioids, their responses will now be unequal. The lateral position is thus readily recorded. I doubt anyone disputes this.

 

Now let the sound source move upwards from its original centre position. As it rises, the source moves further off-axis, by equal amounts for both microphones. The resulting response is somewhat diminished but otherwise unchanged. Due to symmetry, the exact same difference obviously is produced if the sound source moves down rather than up. Moving the sound source, still centred, away from the microphones again drops the amplitude of the response. From the recorded signals, there is no way of determining from which point in a plane bisecting the angle between the microphones the sound originated. For an off-centre source, the geometry becomes a little more complicated and dependent on the exact pattern of the microphones, but in general, for any point in space, there are infinite other points where a sound source would produce the same relative response from the microphones. Mathematically speaking, we are dealing with a projection of three dimensions onto two, and this always entails loss of information.

 

The reason we are able to locate sounds is that our ears are not symmetrical. Their frequency response varies with direction, and the resulting change in sound as our head moves, even by a small amount, helps us deduce the location of the source. For this reason, locating the source of a pure tone can be quite difficult compared to a spread of frequencies. In fact, sirens on emergency vehicles have in many places been altered to take this into account. Likewise, electric carts seen in places like airports often emit bursts of pink-ish noise as alerts since this is more easily localised than a tone.......

 

Slight correction. Although there are two microphones, their responses are not independent. The projection is onto one dimension, the left/right axis, although it ends up encoded using two values. This might seem wasteful, but we anyway need two values in order to represent both position and amplitude. Using a mid/side microphone configuration, their outputs more closely resemble this separation. Calculating the sum and difference of the XY signals gives a similar result.

 

 

</quote>

Sound Minds Mind Sound

 

 

Link to comment
16 minutes ago, Audiophile Neuroscience said:

Mathematically speaking, we are dealing with a projection of three dimensions onto two, and this always entails loss of information.

 

I'm not sure this is correct, the situation seems worse (i.e. more difficult) for the ears than for the eyes. With our eyes, we construct 3D space from 2 (left and right retina) 2D sources of information. With the ears, there seem to me to be 2 1D sources of information.

 

ISTM that Daniel Levitin has it right with his ping pong balls and pillowcase analogy : https://www.newscientist.com/article/mg19726441-500-music-special-the-illusion-of-music/

Link to comment
3 hours ago, Audiophile Neuroscience said:

Yeh, I wondered that too, but I guess that it still only tells you distance away not height.


Stereo do not have height information encoded in them. We localize height based on changes in the frequency response shaped by the pinna. This still requires a priori knowledge of the original event. Having said that, it is still possible to perceive height with speakers by having effects with increasing ( decreasing ) frequencies from 5 to 10 KHz. An example is the LEDR tones. 
 

Make a recording of contradicting placements of the source at various height. Play them to young kids and ask them to localize the sound and you will see that they actually localize based on what they already knew. 
 

To others, I can post a recording of five or six animals sound spread about “x” from left to right and at different elevation.  Anyone want to guess?

Link to comment
7 hours ago, Audiophile Neuroscience said:

 

Thanks John, I didn't know that but will listen to see if I can perceive it.

 

On a related topic there was a huge debate on AS as to whether height could be encoded on a stereo recording in the first place. With a conventional microphone, an object moving sideways or vertically is simply moving away from the mic and it cannot encode anything but  L-R differential response. I always pondered that there must be something, some cue in the signal that subserves the illusion of height perception.Perhaps dynamic range??

I might have mis-stated the term, as I am not familiar with the non-technical language.  Instead of height, perhaps most accurately what I am tryingt to suggest is that the high speed compression distorts the directions between far/rear left and forward left &  far/realr right and forward right.   It creates a kind of gap in the stereo image.  One time, someone used the term 'height', and since sometimes the more common language is abstract to me -- I might have misused the term.

 

So -- right forward and left forward are only distorted a little bit with fast DRC.   Also, there is still some perception of 'rear' -- but the 'in-between' is 'disappeared', gone.

 

It kind of makes sense considering what the compression does -- it amplifies the ambiance -- perceptively pushing it forward.

 

So, I might have misstated 'height', but at least one or two people had mentioned it to me -- but I tend not to percieve height myself.  (I am totally missing the mental ability to consiously image things in my mind, so my perception of stereo imaging might be very primitive.)

 

John

 

Link to comment
10 minutes ago, John Dyson said:

what I am tryingt to suggest is that the high speed compression distorts the directions between far/rear left and forward left &  far/realr right and forward right.   It creates a kind of gap in the stereo image.  One time, someone used the term 'height', and since sometimes the more common language is abstract to me -- I might have misused the term.

Well, semantics aside, thank goodness we have people like yourself open minded enough to consider the possibilities and the technical skills to make the abstract actually materialize.

Cheers

David

Sound Minds Mind Sound

 

 

Link to comment
12 hours ago, Audiophile Neuroscience said:

Yep, we disagree.

 

soundstaging, where the image is on the soundstage, is determined by the recording. Obviously, the recording has to have the cues in the first place, so to that extent yes, sound "quality" in the recording is important. The rest is a function of your speakers and their interaction with the room.

 

It was first described to me reading Art Noxon's (Master of Science degrees in Mechanical Engineering/Acoustics and Physics). Try it, it works 

http://www.about-audio-mastering-software.com/stereo-imaging-mastering.html

 

 


Link to comment
12 hours ago, Audiophile Neuroscience said:

 

Thanks John, I didn't know that but will listen to see if I can perceive it.

 

On a related topic there was a huge debate on AS as to whether height could be encoded on a stereo recording in the first place. With a conventional microphone, an object moving sideways or vertically is simply moving away from the mic and it cannot encode anything but  L-R differential response. I always pondered that there must be something, some cue in the signal that subserves the illusion of height perception.Perhaps dynamic range??

 

Perhaps our knowledge that some instruments are on the floor whilst others are at waist height and others still at head height.

"Science draws the wave, poetry fills it with water" Teixeira de Pascoaes

 

HQPlayer Desktop / Mac mini → Intona 7054 → RME ADI-2 DAC FS (DSD256)

Link to comment
19 hours ago, Audiophile Neuroscience said:

 

soundstaging, where the image is on the soundstage, is determined by the recording. Obviously, the recording has to have the cues in the first place, so to that extent yes, sound "quality" in the recording is important. The rest is a function of your speakers and their interaction with the room.

 

It was first described to me reading Art Noxon's (Master of Science degrees in Mechanical Engineering/Acoustics and Physics). Try it, it works 

 

Yes, the soundstage is determined by the recording - the fascinating thing is that even recordings where attempts were made to reduce the captured acoustics, there is still enough leakage of cues to provide a sense of space.

 

That soundstage is a function of your speakers and their interaction with the room is probably the greatest myth in audio - I can create soundstage without touching either of the latter, and ruin it as well, merely by altering the SQ of the electronics prior to that part of the chain. What different speakers do, is help compensate for inadequacies of the sound being fed to them ...

Link to comment
1 hour ago, fas42 said:

 

That soundstage is a function of your speakers and their interaction with the room is probably the greatest myth in audio -

 

Disconnect one speaker and see what happens to the soundstage and imaging ! You can tweak your electronics all you like but it won't fix the problem

 

 

1 hour ago, fas42 said:

I can create soundstage without touching either of the latter, and ruin it as well, merely by altering the SQ of the electronics prior to that part of the chain. What different speakers do, is help compensate for inadequacies of the sound being fed to them ...

 

Frank, your magic touch is legendary 😉

 

Think of it like his Frank. Soundstaging and Imaging cues are baked into the recording.The electronics in the playback chain can manipulate those cues (we grok it) (but there is some controversy here). The speakers and room interaction alter the quality of subsequently perceived soundstage and image.

Sound Minds Mind Sound

 

 

Link to comment
14 minutes ago, Audiophile Neuroscience said:

 

Disconnect one speaker and see what happens to the soundstage and imaging ! You can tweak your electronics all you like but it won't fix the problem

 

Did that effectively, 35 years ago, 😝 ... just use a pure mono recording ...

 

What that gives you is a soundstage that is all about depth; lateral positioning is hinted at, but the absence of pure panned information is not missed. Think of listening to a live performance, through a doorway, where you are positioned some feet back from the opening, but still have direct visual sight of some of the performers, in the centre.

 

14 minutes ago, Audiophile Neuroscience said:

 

Think of it like his Frank. Soundstaging and Imaging cues are baked into the recording.The electronics in the playback chain can manipulate those cues (we grok it) (but there is some controversy here). The speakers and room interaction alter the quality of subsequently perceived soundstage and image.

 

How I find it works is like this ... the lower the the SQ of the electronics driving the speakers, the more the choice of speaker type, brand, model, and room acoustics matters, subjectively; the higher the SQ of that earlier part, the more one becomes oblivious of the speakers and room, and you only hear "what's on the recording" ... every time.

 

Each combo of highly distinctive components I tweak go through the same process - they start by sounding like themselves, and end up sounding like the recording ... mission accomplished, 😉.

Link to comment
1 hour ago, fas42 said:

How I find it works is like this ... the lower the the SQ of the electronics driving the speakers, the more the choice of speaker type, brand, model, and room acoustics matters, subjectively; the higher the SQ of that earlier part, the more one becomes oblivious of the speakers and room, and you only hear "what's on the recording" ... every time.

 

 

What he said. Electronics SQ can to some degree be characterized by how well it deals with the lowest level sounds - reverb tails, ambience cues. Introducing noise modulation (IMD by another name) corrupts these lowest-level details first (most particularly at the lowest audio frequencies) and the subjective result becomes : a recording which is known to contain plenty of ambience sounds less 'wet' or less 'warm'.

Link to comment
9 hours ago, opus101 said:

 

What he said. Electronics SQ can to some degree be characterized by how well it deals with the lowest level sounds - reverb tails, ambience cues. Introducing noise modulation (IMD by another name) corrupts these lowest-level details first (most particularly at the lowest audio frequencies) and the subjective result becomes : a recording which is known to contain plenty of ambience sounds less 'wet' or less 'warm'.

I am glad that you mentioned IMD as a source of damage to a stereo image -- I have been developing software that in it's natural, most degenerate state will produce IMD very easily.   Even a slight amount of IMD has a strong effect on the ambiance/clarity of a recording.   Clarity has somewhat of an effect on imaging.   IMD damages something good about audio that I have problems actually describing, and it doesn't take much of the IMD to do the damage.

 

IMD is insidious...  IMD isn't just insidious, it is develishly insidious.   Given a minor about of IMD, I have difficulty in detecting it unless doing a direct A/B comparison.   Some forms of IMD appear as a 'fog' over the sound (if I can get my act together when I am less busy, I might be able to create an example.)  It is really interesting what happens to audio when it has been processed in certain ways.  Those forms of processing were ALMOST thought to be benign, but really were/are not benign at all.   Some methods (unfortunately sometimes necessary to do), produce very stealthy, but damaging IMD type distortions.

 

Since ambiance is significantly affected by certain kinds of IMD behaviors, then it can have some effect on the image.

This is actually one of my own frustrations against alot of commercial recordings.

 

John

 

Link to comment

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...