Soundstage, what is it exactly?

[Hayk]

People complain about the Soundstage of an audio equipment whether it is speakers, amplifiers, DACs, anything that makes part of reproduction. Personally, I listen to THE music to be in mono but I understand when people speak about the wide or narrow stereo that can be channel separation, but it exceeds in most cases 90db.
Some speak even the height and the dept.
How it can be measured or related to the design?

 

[NTK]

It is in the recording/signal.

For example, try BACCH. Same hardware, just some added signal processing and you get completely different "soundstage".

 

[klettermann]

I'm not sure I truly know what it is, but I know it when I hear it.

 

[Timcognito]

Placing where sound emanates in a stereo playback image. Left center right, some say front to back but that is harder to discern. I starts in the recording and requires at least two separate channels. When more microphones are added mixing of the recording can synthetically augment it with L/R (im)balance and volume changes. So guitar mid-left, drums center, Horns right, etc. It's product of the recording, the accuracy and placement of the speakers and room effects, such as reflections, bass nodes etc. It has little to do with the speaker design intent.

 

[john61ct]

But it is impacted by transients performance

what some call "Speed" as in "fast vs slow" bass, smeared mushy boomy bloated, as opposed to tight clean & crisp

If not a complete myth, there are objective measurable factors (e.g. attack / decay) that may be at play, just because terms are currently loosey goosey

does not mean the science won't be evolving to more precisely explain in future, what are currently mostly subjective descriptors.

I'm not sure I truly know what it is, but I know it when I hear it.

Amen

 

[kemmler3D]

I read "soundstage" as perception of spatial cues in a recording, or when describing gear, ability to reproduce those cues.

Soundstage can be quantified in a more or less straightforward way when it comes to the direction a direct sound is meant to be coming from in a recording. You can (in principle, I think) look at interaural timing and level differences in the recording to say objectively where the sound will appear to be coming from.

What is not easy to quantify as far as I know is front / back, height, or indirect/reflected spatial cues. These depend on reverb and EQ in the recording that's basically idiosyncratic.

And then when it comes to speakers, it's hard to say much outside of observing the beam width and inferring what's reflecting from the sidewalls.

So overall I think it's hard to speak objectively about spatial cues beyond direct sound.

 

[olieb]

I read "soundstage" as perception of spatial cues in a recording

I would agree that the term seems to relate to spatial characteristics of what is heard, mainly something encoded (signal/channel count), but what can be altered by speakers and room.
I do struggle with the distinction between "soundstage“ and “image“, though.

If the soundstage widens, is this independent from the acoustic positions of the instruments (the “image“?), or not?
The way I understand it, ‘soundstage’ would primarily describe the characteristics of the spatial setting (such as its extent) and “image“ would relate to the perceived positions of the instruments.
Perhaps pictures can make clear what I mean.
Spatial presentation = image + soundstage

=

+

In this distinction, image is mainly a result of direct sound and soundstage comes from reflections (in recording venue and listening room).
However, I might be wrong about the use of these words/terms.

 

[RexrothPigeon]

whether it is speakers, amplifiers, DACs, anything that makes part of reproduction

The only things that contribute to what we perceive as soundstage are:
1. the recording
2. the speakers
3. the room
4. the listener's position in relation to the speakers and the room

Not DACs or amps.

 

[john61ct]

Soundstage is to me includes depth

There are effects that purport to "widen" the image, so some of the aspect is mathematical.

What is not easy to quantify as far as I know is front / back

Schiit SYN - all analog - has not only a Width control, but part of what the Presence knob is supposed to do - when its "Hafler-type simulated Surround upmix" feature mode is enabled is move the vocals forward or back.

Whether it does anything greatly depends on the content, and how it was mixed.

I do have car audio components that separate Front and Back channel pairs, and my Dad's old Quadraphonic Marantz receiver (still going strong), and have to admit I enjoy the effects of messing around with those controls!

 

[KSTR]

For me, this is a function of the content (the recording) and the speakers+room, basically how much of the encoded content can be rendered and even some fake clues being generated. The quality of the projected auditory illusion.

- 0D (point) : any kind of playback with one single speaker.

- 1D (line) : the basic stereo setup with two speakers, equilateral triangle and all, moderate acoustic room quality. The sound stage is a line between the speakers (or even slightly beyond) with any kind of stereo source material. This level is reached by almost any quality of speakers. DAC/Amp etc play a role but quite minor at this point.

- 2D (plane, adding depth to the above) : Most setups can do this as well, with the soundstage now also extending to behind the speaker (and in front of the speakers as well, with a certain class of speakers in good rooms). Discrete parts of the (phantom-)images are rendered very differently than diffuse parts (reverb tails, "ambience" etc). A trinaural speaker setup (an simple static LCR upmix of stereo content) does this extremely well IME, with the image sort of fully detached from the speakers (like, you open your eyes and wonder if it is really just those two -- or three -- speakers that produce the stunning illusion that even stays intact when you turn/move your head). Does never work well in an acoustically bad room.

- 3D (adding height) : this is very fragile because it is based on fake clues as there are no vertically distributed sources the could render any height aspect precisely, but sometimes things fall into place and all the different phantom image locations in a scene do a have a different source height or distinct vertical extension, making it much easier to follow different sources in dense scene.

Think of choral music with many voices, like 20 or so. Perfect soundstage is what happens when you can easily pinpoint and follow each voice without even trying hard with a stunning feeling of being there, and have a palpable impression of the geometric setup of the choir, like where exactly people are standing. When this happens it is pure magic.
I always compare that to the difference between watching a movie on a cell phone (the 0D variant) vs. seing a 3D version in a good movie theater.

 

[john61ct]

Verging on the Envelopment concept with that latter

 

[Vuki]

3D & music

 

[Keith_W]

Of course you know what it is. I am sitting here in my living room. I can hear the clothes dryer in the laundry, the patter of rain outside, my wife drying her hair upstairs, the chirping of crickets, and my music playing in the background. If I close my eyes, I can point to exactly where all these things are. Soundstage refers to how well your speakers/room/recording can reproduce that.

 

[Hayk]

I as mono amp designer, the only technical aspect of soundstage I can master, is the dept. I can bring the voice forward to be expressive in melody or throw it at the backstage to be soothing in harmony.
I read some people complaining the DAC or the amp to lack soundstage, with over 90db channel separation, but never heard someone complaining about pickup cartridge that are less than 30db.
There are amplifiers that reproduce big sound, large, generous in mono and others loud but small sound as a powerful transistor radio. Maybe this aspect is what people describe as soundstage, if so, bring the two speakers side by side to test for soundstage to appreciate how big it sounds.

 

[KSTR]

I read some people complaining the DAC or the amp to lack soundstage, with over 90db channel separation, but never heard someone complaining about pickup cartridge that are less than 30db.

If anything, too much channel separation is bad for sound-stage. Crosstalk helps to generate fake HRTF-triggered directional cues by creating slight comb filter patterns with A/B type content in the recording which contain time delta between almost identical signals in L/R channels, other than X/Y type which is only about magnitude differences. This is also (part of) the magic behind Trinaural playback setup.

 

[kemmler3D]

I would agree that the term seems to relate to spatial characteristics of what is heard, mainly something encoded (signal/channel count), but what can be altered by speakers and room.
I do struggle with the distinction between "soundstage“ and “image“, though.

If the soundstage widens, is this independent from the acoustic positions of the instruments (the “image“?), or not?
The way I understand it, ‘soundstage’ would primarily describe the characteristics of the spatial setting (such as its extent) and “image“ would relate to the perceived positions of the instruments.
Perhaps pictures can make clear what I mean.
Spatial presentation = image + soundstage
View attachment 535838 = View attachment 535853 + View attachment 535844

In this distinction, image is mainly a result of direct sound and soundstage comes from reflections (in recording venue and listening room).
However, I might be wrong about the use of these words/terms.

I get your drift and to be honest I haven't consulted a glossary to be sure that my interpretation is proper.

I think where it gets very tricky is room interactions interaction with room interactions on the recording...

 

[STC]

Soundstage is a term whose meaning has shifted significantly over time, which often leads to people talking past each other without realizing it.

In the original, “old school” sense of stereo (1950s–1960s), soundstage referred specifically to the phantom image and sound scene created between the two loudspeakers. This was the core promise of stereo: stable, convincing lateral placement of instruments and voices along the line connecting the left and right speakers, achieved mainly through amplitude panning and basic phase/ITD cues. That is what classic stereo was engineered to do.

As audio technology, recording techniques, and listening environments evolved, the term soundstage expanded in audiophile circles. It came to include the ability to create illusions of sound outside or beyond the physical speakers — wider than the speaker span, with added depth layering, and sometimes even vague height or “air.” A recording or system praised for “great soundstage” today is often valued more for this extended, spacious presentation than for a tightly focused image strictly between the speakers. As a result, a well-engineered classic stereo recording that excels at precise internal imaging might be rated lower by modern listeners than one engineered (or played back) to push sounds outward. Think 8D remastered.

This evolution has created real confusion. The term is now frequently mixed up with spatial audio — a more recent concept involving fuller 3D reproduction through multi-channel formats (like Dolby Atmos), advanced binaural processing, or room correction. What began as a straightforward description of two-channel stereo capability has become broader and more subjective, leaving many (including long-time listeners) unsure what exactly someone means when they say “soundstage

ST

 

[Geert]

Soundstage is what your brain zooms into when placing a heavy doorstop on your DAC or when swapping opamps.

 

[RandomEar]

The only things that contribute to what we perceive as soundstage are:
1. the recording
2. the speakers
3. the room
4. the listener's position in relation to the speakers and the room

Not DACs or amps.

Agree mostly. In addition, you can influence it to a large degree with DSP (think "virtual 7.1 surround" and that stuff) and to a lesser degree with EQ (see "Blauert's Bands"). So some bad amps like tube amps with high output impedance might be able to unintentionally create kind of a forced EQ with specific speakers which influences the sound stage. This will depend on the speaker impedance curve and on how bad the amp's output impedance is. From my point of view, this is a defect in the signal chain and shouldn't happen, because an amp should not audibly alter the frequency response of a HiFi setup. It's not impossible, though.

Based on what many fellow members have said, it is pretty clear that the concept of sound stage itself is somewhat subjective. For me, it's mostly related to how wide and deep the stereo image appears to be, which is 80% information encoded in the recording: Time of flight differences between instruments and the microphone, the delays and magnitude of reflections, the influence of damping factors from wall and floor/ceiling materials or furniture in a room reducing the frequency response of direct and indirect sound reaching the microphone, as well as differences between the recorded L and R channel for all this stuff.

Speakers and room acoustics as well as our own HRTF will then alter this info from the recording again during playback and add the remaining 20% to the impression.

 

[Hayk]

I could understand by Google AI about the height. It gives the idea of how big are the instruments that you can visualize. I did remember having read once upon a time about Bose 901. The reviewer was complaining that he sees gigantic violins, being inside a guitar.
I had once a very strange experience. It was a shop in electronic market in Bangkok that sold tube related articles but the fellow was a speaker gouru built on order. He was trying out a speaker under development, with concave ceramic woofer and a dome tweeter probably SEAS. The enclosure was made of stacked layers of medium wood forming a shape of half ellipse. What strange phenomenon I heard, I could hear the sound coming only from the tweeter where the very tiny instruments I could visualize upon an invisible half dome about 30cm in diameter covering the tweeter. I even could sense that the instruments are slightly in floating movements.