What makes the sensation of "sound stage"?

[audio2design]

Our room is exceptionally non reflective. Soft furniture and very heavy drapery on all windows. Whatever it is it's an interesting difference

What about the walls, floor, ceiling? Heavy drapery does not work on an overly wide frequency range. Good at high frequencies, so so at lower frequencies. It can be good for frequency tuning of mids-highs, but that is about it.

 

[MakeMineVinyl]

Interesting discussion...manufactured illusion or encoded effect? Is it possible that there are some localization clues (reflections etc.) captured by the mic’s that our brain can decode?
I’m sure that the engineers have tools available as well to enhance the effect. Listen to “Dogs” on Roger Water’s In The Flesh album. You will swear the barking is coming from behind you!

Plug-ins such as this use HRTF processing to enable apparent panning all around the listener. Water's "Amused To Death" uses Q Sound processing very heavily (and very well) to achieve this. While its easier to hear the effect of processing like this (and true binaural) with headphones, speakers which are highly directional and are well away from walls where there are minimal first reflections can pull off the effect very well also.

I'm so much an advocate for hearing as much as possible only the direct sound from the speakers - and thus only what is actually in the recording - that I rate any secondary sound from the room as distortion - just as bad a thing as gross harmonic or intermodulation distortion. The problem is that with the vast majority of speakers being relatively non-directional, this distortion of the intent of the recording will continue to exist. I'm a hard ass in that way.

 

[audio2design]

Plug-ins such as this use HRTF processing to enable apparent panning all around the listener. Water's "Amused To Death" uses Q Sound processing very heavily (and very well) to achieve this. While its easier to hear the effect of processing like this (and true binaural) with headphones, speakers which are highly directional and are well away from walls where there are minimal first reflections can pull off the effect very well also.

I'm so much an advocate for hearing as much as possible only the direct sound from the speakers - and thus only what is actually in the recording - that I rate any secondary sound from the room as distortion - just as bad a thing as gross harmonic or intermodulation distortion. The problem is that with the vast majority of speakers being relatively non-directional, this distortion of the intent of the recording will continue to exist. I'm a hard ass in that way.

When I listen to those samples for speakers, they don't ever sound anything like in front of me, and only extending about to the speakers, though if I pull my panels off a wall (add back first reflections) the width grows. The headphone ones work better.

 

[MakeMineVinyl]

When I listen to those samples for speakers, they don't ever sound anything like in front of me, and only extending about to the speakers, though if I pull my panels off a wall (add back first reflections) the width grows. The headphone ones work better.

The conventional speakers in my office don't pull off the effect either, but its obvious (and at times pretty wild) on my directional horns extending up to and sometimes beyond the room boundaries in all directions. I remember even as a child preferring to listen to my stereo speakers in extreme near field, like 1 foot away, and preferring the much more enveloping imaging.

 

[Blaspheme]

Good thinking, and I almost clarified this point in my previous post, but I felt I was already getting too far out into the physics weeds that don't really matter for audio. Important to not confuse the "points" that I am talking about with "coordinates".

A plane(3 points) is indeed 2 dimensions, and the fact that you can identify any point on that plane with 2 coordinates(x,y) is a great way to think about it, as you mention. Similarly, you can fully define the location of any object on a line(2 points) with just 1 coordinate(x), hence it(and 2 speakers) being just 1 dimension.

There was a really great youtube video I watched 5+ years ago describing points, lines, and planes as a means to understand the 11 dimensions. I looked for it a few days ago, but couldn't find it. There was one that was similar, but no where near as good as the one I remember.

Yes, x alone positioning a point in a line completes the explanation, very cool.

Thinking about stereo, I'd say we have two points representing the speakers plus a third representing the listener (simplified model, obviously). Two-channel stereo is the minimum requirement for representing a three-dimensions sonic image/soundstage. We confirm that three dimensions can be represented by using that audiocheck LEDR test (certainly via headphones, otherwise when adequate gear is correctly set up in-room, assuming a non-confounding room) for example. So I agree with the "stereo is 3D" argument. I'd also assume modern object-oriented multichannel can do it more comprehensively/reliably (but I don't have that set up, and availability of music in that format is limited).

Shame you couldn't find that video! Something I really like about these forums are the fabulous tangents. The Sorli and Fiscaletti article (linked upthread, on time as 4d—or not) was very cool also.

 

[Blaspheme]

The conventional speakers in my office don't pull off the effect either, but its obvious (and at times pretty wild) on my directional horns extending up to and sometimes beyond the room boundaries in all directions. I remember even as a child preferring to listen to my stereo speakers in extreme near field, like 1 foot away, and preferring the much more enveloping imaging.

I hadn't thought all that deeply on implications of wide vs narrow directivity in sonic image/soundstage and consequent loudspeaker preference terms before reading these threads.

The only horns I've spent any time listening to were Oswald Mills Mini, which were fun. But I realise retrospectively I was likely comparing wider directivity (and wide baffle) Focal with narrow(er) directivity (and narrow baffle) Audio Physic when I bought my last speaker pair (Scala vs Codex, specifically). Focal did a better job with (my go-to hardcore test track) BMTH The Comedown but a bit flat with (also post-hardcore, but more intricate) La Dispute Such Small Hands/Said the King to the River. Codex was way more fun with (very electronic) FKA twigs Figure 8 ... yada yada (other people's taste can be a bit f*cking boring I know).

Audiophile-review-style subjective listening track citations aside, I'm making analytical assumptions: no Klippel-style spins for either speaker in sight, but there are some on and off-axis FR charts in German HiFi press for Codex (confirming treble is a bit hot on-axis and drops away more as you turn: AP use their cone tweeter of course) and for Focal I'm simply guessing wider dispersion from dome tweeter. AP's original design philosophy was to position speakers well into the room and listen almost near-field. And a lot of stuff about relative timing of direct vs first refections and all that. Which makes sense in the context of this thread. They are notable for soundstage/imaging and disappearing act in the subjective reviews, for what it's worth. I had more gear-lust for the Scala though.

But I guess I was leaning to narrow dispersion sonically, as it was more fun for most (but not all) of the music I like listening to. Comparison was in moderately but not maximally treated rooms (some bass traps and wall panels). Note that my context for soundstage and imaging is studio-assembled sound, not acoustic performers-on-a-stage live. In this case narrow dispersion setup appeared to give wider/deeper stage and more specific imaging. The impression carries over to my home setup (I bought the Codex). There are way too many other variables in play to generalise though. That said, next time I want to listen to single-speaker mono with a blindfold on I'll absolutely go Revel.

 

[MakeMineVinyl]

I'd also assume modern object-oriented multichannel can do it more comprehensively/reliably (but I don't have that set up, and availability of music in that format is limited).

For what its worth, with even the most elaborate Atmos type setups, I hear the surround/height speakers oftentimes as discrete sources which destroys the illusion. In rooms with this type of setup which are more reverberant, the overall soundstage is destroyed because I'm hearing more of what the room wants to say than what is in the recording.

Overall, I get the most natural sounding and integrated 'as a whole' soundstage with two directional stereo speakers (full range horns), well spaced from any walls so that the room contributes as little as possible. As I mentioned previously, with the right recordings the sense of there being a room with walls completely disappears.

 

[richard12511]

Yes, x alone positioning a point in a line completes the explanation, very cool.

Thinking about stereo, I'd say we have two points representing the speakers plus a third representing the listener (simplified model, obviously)

I can kinda agree with this. I'm not sure I would call it another "dimension", simply because it's a variable that is unknown to the mix engineer. It's not a tool the mix engineer(other than by guessing) can work with. Maybe they can make educated guesses, though, so maybe I can see an argument. They also have the real 2nd dimension of "time". It's not a spatial dimension, but time is often referred to as the fourth dimension in our 3D world . IMO, this is where the 2nd dimension of stereo comes from. Mix engineers know that the sound will spread out in 3 dimensions as a function of time. They can use this to make certain sounds come from "behind" you. The third dimension, "height" is something that's really interesting to me. From what I've learned recently, it seems that it comes from frequency response manipulation designed to trick the brain into thinking a sound is coming from above.

We confirm that three dimensions can be represented by using that audiocheck LEDR test (certainly via headphones, otherwise when adequate gear is correctly set up in-room, assuming a non-confounding room) for example. So I agree with the "stereo is 3D" argument.

It works here really well on speakers . I would make the distinction between real dimensions and simulated dimensions(illusions). From the perspective of our brain, the sound really is coming from above us. From the speaker's perspective it's coming from the same "point" in the z plane. It's why I agree with both sides. From our brain's perspective(which is what matters, ultimately), 2 speakers are capable of moving images in 3 dimensions.

I'd also assume modern object-oriented multichannel can do it more comprehensively/reliably (but I don't have that set up, and availability of music in that format is limited).

Indeed! Multichannel (with surrounds and heights) is capable of real (not simulated) 3D. The problem is(as you say) the lack of adoption, and by consequence, lack of material. It's just not practical, imo. No doubt multichannel sounds better than stereo ever can, but good stereo gets you 90% of the way there, and clutters the room up 500% less. Luckily, I live alone(hopefully not forever), so I can experiment with multichannel. True multichannel is a small step above stereo, but it's rare(especially outside of classical), so 90% of the content I listen to is still stereo. One great tool I've found is upmixing to Auro3D. I prefer Auro3D over stereo on all but just a few songs(even with crap surrounds). Problem is it requires 4 height channels, and Auro2D or Dolby Atomos or DTS are all worse than stereo.

Shame you couldn't find that video! Something I really like about these forums are the fabulous tangents. The Sorli and Fiscaletti article (linked upthread, on time as 4d—or not) was very cool also.

I'm gonna keep looking. The concept that really helped me to understand higher dimensions is the concept of compressing 3 dimensions down into a single point. So:

0th dimension: a point
1st dimension: two points, connected with a line
2nd dimension(plane): three points connected by lines(think triangle)
3rd dimension: 4 points connected by lines(think triangular prism)
4th dimension: compress that triangular prism down into a single point, now draw another point, and then draw a line to connect those "points". That's 4 dimensions
5th dimension: draw a 3rd "triangular prism point" and connect it with a line.
etc.

 

[richard12511]

I hadn't thought all that deeply on implications of wide vs narrow directivity in sonic image/soundstage and consequent loudspeaker preference terms before reading these threads.

The only horns I've spent any time listening to were Oswald Mills Mini, which were fun. But I realise retrospectively I was likely comparing wider directivity (and wide baffle) Focal with narrow(er) directivity (and narrow baffle) Audio Physic when I bought my last speaker pair (Scala vs Codex, specifically). Focal did a better job with (my go-to hardcore test track) BMTH The Comedown but a bit flat with (also post-hardcore, but more intricate) La Dispute Such Small Hands/Said the King to the River. Codex was way more fun with (very electronic) FKA twigs Figure 8 ... yada yada (other people's taste can be a bit f*cking boring I know).

Audiophile-review-style subjective listening track citations aside, I'm making analytical assumptions: no Klippel-style spins for either speaker in sight, but there are some on and off-axis FR charts in German HiFi press for Codex (confirming treble is a bit hot on-axis and drops away more as you turn: AP use their cone tweeter of course) and for Focal I'm simply guessing wider dispersion from dome tweeter. AP's original design philosophy was to position speakers well into the room and listen almost near-field. And a lot of stuff about relative timing of direct vs first refections and all that. Which makes sense in the context of this thread. They are notable for soundstage/imaging and disappearing act in the subjective reviews, for what it's worth. I had more gear-lust for the Scala though.

But I guess I was leaning to narrow dispersion sonically, as it was more fun for most (but not all) of the music I like listening to. Comparison was in moderately but not maximally treated rooms (some bass traps and wall panels). Note that my context for soundstage and imaging is studio-assembled sound, not acoustic performers-on-a-stage live. In this case narrow dispersion setup appeared to give wider/deeper stage and more specific imaging. The impression carries over to my home setup (I bought the Codex). There are way too many other variables in play to generalise though. That said, next time I want to listen to single-speaker mono with a blindfold on I'll absolutely go Revel.

IME, narrower dispersion tends to image better, and be more clear. Horns are great for this, and also benefit by being more dynamic. Wider dispersion tends to throw a wider soundstage, and be more enveloping. The experts(from what I can tell) are divided on this. Toole heavily favors wide dispersion, while Geddes favors more narrow dispersion.

 

[Pennyless Audiophile]

Interesting discussion...manufactured illusion or encoded effect? Is it possible that there are some localization clues (reflections etc.) captured by the mic’s that our brain can decode?
I’m sure that the engineers have tools available as well to enhance the effect. Listen to “Dogs” on Roger Water’s In The Flesh album. You will swear the barking is coming from behind you!

Manufactured. The position of instruments and singers is built in the mix, The musical event as you hear it in the recording, it never happened in reality, save very few cases of super taliban purist recording. And those often look real but do not sound well...