Soundstage, what is it exactly?

[MattHooper]

Yes, I've read this stuff about about cables as well. And cable risers. And little rocks that go under your sofa to radically smooth the quantum resonances in the room so obviously that even the reviewer's wife and her boyfriend can hear it from the other bedroom.

Yup. But that’s typically under sighted listening conditions.

That’s why I’m more interested when these impressions survive testing that controls for imagination.

 

[john61ct]

I I would say something like "you can still hear a good soundstage even from off to the side".

3-channel stereo, using a physical center rather than relying on a "phantom" one

is supposed to help deliver that last.

But does it help or hurt what strict-stereo fans refer to as soundstage generally?

 

[Duke]

3-channel stereo, using a physical center rather than relying on a "phantom" one

is supposed to help deliver that last.

But does it help or hurt what strict-stereo fans refer to as soundstage generally?

Good question!

Using a physical center channel probably always helps maintain a good soundstage for off-centerline listeners with a conventional stereo set-up.

The approach I use is something unorthodox called "time-intensity trading", and it calls for speakers that have fairly narrow and well-controlled radiation patterns, toed-in aggressively such that their axes criss-cross in front of the center of the listening area. The ear uses both arrival time and intensity to determine the direction of a sound source. So with a time-intensity trading configuration, for off-centerline listeners the near speaker obviously "wins" arrival time, but the far speaker "wins" intensity, because an off-centerline listener is on-axis (or very close thereto) of the far speaker, and very far off-axis of the near speaker. Thus the far speaker is louder in the frequency region where most of our directional cues come from. The secret to time-intensity trading working well is this: The off-axis response of the near speaker must fall off SMOOTHLY and RAPIDLY as you move off-axis.

No center channel needed, as long as none of the listeners have a serious hearing imbalance. I have customers with home theater systems who have sold their center channel speakers because they prefer to use time-intensity trading. I can post a photo or two if you'd like to see what it looks like. [edited to add: I had one customer with a trinaural music system who also preferred time-intensity trading and sold his center-channel speaker, so it's not just for home theater.]

The advantages include, zero potential timbral mis-match between the mains and center channel speaker, and the soundstage depth is better. Ime, and in the experience of my beta-testers and customers, using a center channel speaker constrains the soundstage depth on sounds that come primarily from the center channel speaker. This can be detected when you close your eyes. This greater center-of-the-soundstage depth it's probably of academic interest only for home theater, but arguably is a worthwhile improvement for music videos, as well as for two-channel playback. Also, you don't have to switch between systems for music vs home theater.

So back to the word "soundstage"... when I demo this type of system, at some point I like to ask the person to sit well off to one side of the centerline, maybe in front of one of the speakers or even a bit further over. With most systems, the sound would almost all seem to come from the near speaker. When asked, people consistently tell me they can still hear an enjoyable soundstage from that far-off-center location. The tonal balance is still good over there as well, so it's not only the spatial quality that benefits. Imo.

Credit to Earl Geddes for teaching me about this.

 

[john61ct]

Wow interesting and food for thought.

Makes 3-channel almost seem simplistic

 

[Keith_W]

The approach I use is something unorthodox called "time-intensity trading", and it calls for speakers that have fairly narrow and well-controlled radiation patterns, toed-in aggressively such that their axes criss-cross in front of the center of the listening area.

That sounds like Ishiwata toe-in?

 

[xanalog]

'Soundstage' is like pRonography!
U.S. Supreme Court Justice Potter Stewart's concurrent opinion in the 1964 pRon case: "I know it, when I see [hear] it"

 

[Duke]

That sounds like Ishiwata toe-in?

I think the set-up geometry is pretty much the same, but Earl Geddes' approach adds the aforementioned radiation pattern requirement. Unlike Ken Ishiwata, I have not found a time-intensity trading setup geometry to work well with the typically wide-and-fairly-variable off-axis directivity of "conventional" speakers.

Here is my contribution to that thread.

Wow interesting and food for thought.

Makes 3-channel almost seem simplistic

Well the overwhelming advantage of adding a third channel to an existing stereo system is this: You can keep the speakers you already have, instead of starting all over again and being limited to the few-and-far-between models likely to work well in a time-intensity trading configuration.

 

[john61ct]

Well the overwhelming advantage of adding a third channel to an existing stereo system is this: You can keep the speakers you already have, instead of starting all over again and being limited to the few-and-far-between models likely to work well in a time-intensity trading configuration.

Exactly what I meant!

I am firmly in the camp of making do with my already-collected gear, not going to start replacing anything until my base system is very well optimised at least as far as my skillz allow.

 

[KSTR]

I’ve never thought of XTC as something meant to widen the stage. Although, it is possible to be added with that purpose. The spatial cues are already in the recording — removing crosstalk just lets the brain perceive them without the masking and comb‑filtering that normally occur

Well, the majority of older recordings are produced for speaker playback as-is, hence full speaker crosstalk removal creates the very unnatural rendering we know from headphones, which can't do any form of soundstage I could enjoy. When there is a fully left-panned signal in the recording, I don't want to hear that on the left ear only like if it were a real source directly at the ear. With full XTC is not as bad as with headphones but still bad for my tastes. OTOH, headphone feed with added crosstalk to mimic the incidence of real speaker signals doesn't fully cut it for me either, although it makes many recordings much more listenable than without.

 

[KSTR]

3-channel stereo, using a physical center rather than relying on a "phantom" one

is supposed to help deliver that last.

But does it help or hurt what strict-stereo fans refer to as soundstage generally?

Trinaural (or its refined extension, TriField as designed by the late Michael Gerzon) works the opposite way, it removes all discrete sound sources and replaces them by phantom sources. All three speakers always get signal (except for one rare corner case where one of L or R speakers goes silent).
This can create comb-filtering right in those three physical sources depending on how the stereo signals are encoded in the recording and while that sounds like a bad idea, it actually helps to render a better, more see-through, soundstage, with the physical speakers disappearing better. It is also less prone to image shift with head movement/rotation, which helps the impression of realism big time.

 

[john61ct]

Trinaural (or its refined extension, TriField as designed by the late Michael Gerzon) works the opposite way

Very interesting! Are there relevant implementations that can be inserted into ordinary systems at a reasonable cost for testing the concept?

Ideally not requiring a PC as player...

 

[KSTR]

See here for the basic and very simple matrix equations which can be realized with a few opamps. Later in that thread I describe a fully passive version that one can use with fully signal balanced outputs of a source/preamp.
EDIT: you absolutely need three identical speakers, or at least speakers with compatible crossovers so that the acoustic output of all three speakers is in phase. Otherwise you get a very fancy but totally screwed-up frequency-dependant soudnstage (L and R content swapping sides, etc).

 

[john61ct]

See here for the basic and very simple matrix equations which can be realized with a few opamps. Later in that thread I describe a fully passive version that one can use with fully signal balanced outputs of a source/preamp.
EDIT: you absolutely need three identical speakers, or at least speakers with compatible crossovers so that the acoustic output of all three speakers is in phase. Otherwise you get a very fancy but totally screwed-up frequency-dependant soudnstage (L and R content swapping sides, etc).

Very interesting! But I would have no clue how to implement this, never soldered, dunno what opamps really do.

So maybe a little more expensive?

I will be getting a small cheap miniDSP just for learning, if that helps.

I also have a Schiit SYN if you're familiar, or at least my son does

 

[KSTR]

I'm not personally familiar with either of these, but once you can have any kind of stereo width control (analog or digital) that allows out-of-phase mix (to widen the stereo width) you can create the new L and R channels, and the center channel simple is the sum.

Theoretically, a miniDSP 2x4 HD would fully do, just that those idiots forgot to add a "channel invert" switch in the matrix mixer, meh :-(

 

[thcdru2k]

Imaging precision and soundstage size are not always correlated

 

[Suono]

by definition they should be, then what they do after the registration could change everything.

 

[dannut]

those idiots forgot to add a "channel invert" switch in the matrix mixer, meh :-(

The newer Flex8 had channel inversion for matrix mixing. So they were listening.

The ubiquitous ADAU board w. SigmaStudio is also capable. But strictly DIY and hassle for the end user. I don't think Thomann's cheap DSP unit had channel inversion.

 

[KSTR]

The newer Flex8 had channel inversion for matrix mixing. So they were listening.

But most likely it doesn't have a proper input field to enter a simple (signed) factor, just a dB field and invert checkbox, right? If so, failed attempt in an effort to make the operation "user-friendly" by imposing restrictions without need.

 

[KSTR]

Imaging precision and soundstage size are not always correlated

Definitely. You can have a large soundstage but filled with only blurred images. What we (or I) want is pinpoint 2D imaging of the discrete sources plus large and 3D-ish imaging of the diffuse sources (ambience, reverb tails, early reflections).

 

[Duke]

What we (or I) want is pinpoint 2D imaging of the discrete sources plus large and 3D-ish imaging of the diffuse sources (ambience, reverb tails, early reflections).

Agreed.

Ime it can be challenging to get both pinpoint imaging and ambience/envelopment/immersion simultaneously. Painting with broad strokes here, I've found that having a fairly long time gap between the first-arrival sound and the strong onset of reflections can help. What approach(es) have you found to be helpful?