Qualities correlated to Soundstage or ASW? Looking for papers.

[Lord Victor]

Spoiler: I'm looking for papers/objective measures, don't yet have them. This takes offset in subjective observation/experience.

I'm currently trying to decipher what properties of speakers or drivers, might be correlated with particular aspects of perceived soundstage - based on having heard certain speakers consistently perform well in this regard (fx. Dynaudio Audience 10), regardless of which room/setup they've been in, compared to other speakers.

Since "soundstage" is a nebulous, apparently bordering on inflammatory, term, I'll attempt to define specifically what I'm looking for:
1. Scale: Some speakers tend to create a stage much larger in scale (width, depth and height) compared to others - making the music/musicians sound "bigger" (possibly less defined imaging).
2. Localization: Some speakers are difficult to make "disappear" in the sense that its generally easy to hear that the sound is coming from them, and the staging tends to be smaller and collapse very easily if you move (fx. very directive ribbons). Conversely some tend to very easily disappear, making a stronger center image, and in general rarely calling attention to themselves - and these tend to retain this staging more when you move around, comparatively.

So I'm looking to make something which creates large imaging, and which sonically 'disappears' readily.

As mentioned specific speakers like Dynaudio Audience 10, Evoke, Special 40 and Contour 1.4 have had this property in my experience, consistently across rooms - being far better at this than the about 15 other speakers I've owned and more that I've heard. Only close contender might be an Infinity Kappa 6.2i series ii.

What little papers/documentation I've found about Apparent Stage Width, Toole and from various other literature lead me to think this might tie to the reflected sounds relation to the direct sound from speakers, and hence possibly relate to the speakers directivity - which as far as I cant tell is largely determined by the tweeter. My assumption being that off axis response being as close to on axis would be best. However looking at directivity plots of old Dynaudio tweeters, their off axis response appears to roll off significantly, albeit in a fairly smooth/consistent manner.

Beyond this I know the aforementioned Dynaudio speakers tend to be noted for their exceptional bass apart from the staging. It's been a suspicion of mine that to some extent bass output can psychoacoustically lead to a sense of things being "bigger" - thought that is even more guesswork/intuition.

In my current attempts I haven't found common FR components amongst the speakers in measurement.

I'm hoping anyone in here have any thought on objective parameters which might influence this, or know of research papers out there that discuss this, as I've been unable to find much so far.

 

[steve59]

I've posted this question and another asking for measurements that would display resolution and so far there hasn't been one presented for either. I think bass could be the link to the foundation that creates a soundstage. I had a pair of speakers (my amps were unable to drive the woofers in those speakers to spec) that could image exceptionally well, but the presentation was eerie because there was no soundstage, just separate instruments in space, pretty cool, but impossible to develop the moving picture behind my eyes music usually creates. Sometimes I wonder if the scientific mind even makes that translation?

 

[Lord Victor]

Not quite sure what you mean there, but lack of large stage certainly hampers the musical enjoyment greatly imo... That said I have built speakers using Dyn drivers which stage well and quite large, with barely any bass (due to being small and sealed speakers) - but a subwoofer definitely enhances the sense of space further to me. - But the property would appear to not be reliant on it.

And though I'd prefer to stay on topic, I'll say resolution suffers even worse from the issue of being hard to define clearly, is would be hard to separate "real" resolution if that exists, relating to the quality of the driver in some way, from perceived resolution caused by either high frequency energy drawing attention to "details" or uneven FR similarly highlighting certain things more than one is used to.

With regards to soundstage so far this paper supports the theory of reflected sound to some extent

"The proposals also imply that the perception of spaciousness (and envelopment) will depend on

1. 1. the absolute strength of the reverberant field at least 120 to 170 ms after the ends of sound events;
2. the spatial properties of the reverberant field in the same time period; and
3. the proportion of gaps in the musical source material which allows the background stream to be heard."

However from the Abstract, I'd say it only accounts for effects caused by the room and source material. Nothing that could be caused inherently by the speaker itself.

 

[steve59]

Consider this experience of mine. My first revel speaker was the F52, their first waveguide tweeter (i think). It was shallow and deep and soundstaged like no speaker I'd had before. I thought more has gotta be better and sold those for the salon 1's, soundstage went from precise and intimate to larger than life 'garage band' sound(no waveguide). I then moved to the salon 2(shallow, wide waveguide)hoping to get the intimate, precise stage back and got halfway there with amazing image presentation along with precise placement, but the soundstage was mia? Resolution from the f52 to the salon 2 was improved, but bass was much better from the f52's (In room measured)leading me to believe the bass is necessary to create the stage and had my amp been better at driving the more demanding salon 2 I probably would have realized my goal. All my speakers land in the same room. The soundstage is an illusion created by the mind, but in my case assisted by a bassline.

When you consider Dr Toole only does his work with a single speaker and he seems to be the last word in loudspeakers I don't expect any progress regarding soundstaging from that camp. Isn't the soundstage an illusion initially created by stereo?

 

[steve59]

I have no academic background so can provide nothing intellectual to your thread. not trolling so much as providing personal experience.

 

[Lord Victor]

Consider this experience of mine. My first revel speaker was the F52, their first waveguide tweeter (i think). It was shallow and deep and soundstaged like no speaker I'd had before. I thought more has gotta be better and sold those for the salon 1's, soundstage went from precise and intimate to larger than life 'garage band' sound(no waveguide). I then moved to the salon 2(shallow, wide waveguide)hoping to get the intimate, precise stage back and got halfway there with amazing image presentation along with precise placement, but the soundstage was mia? Resolution from the f52 to the salon 2 was improved, but bass was much better from the f52's (In room measured)leading me to believe the bass is necessary to create the stage and had my amp been better at driving the more demanding salon 2 I probably would have realized my goal. All my speakers land in the same room. The soundstage is an illusion created by the mind, but in my case assisted by a bassline.

When you consider Dr Toole only does his work with a single speaker and he seems to be the last word in loudspeakers I don't expect any progress regarding soundstaging from that camp. Isn't the soundstage an illusion initially created by stereo?

Harman’s research paper on the matter showed it is better evaluated with a single speaker. Speakers in general are better evaluated as singles.
The stereo effect, whilst nice, masks performance.

I was deliberately keeping personal experience to a minimum to make the question as concise as I could - since personal experience usually just results in grumbling and no answer in these topics.

 

[tmuikku]

Hello, see David Griesinger work, a lot of it is on his website also some lectures on youtube. Even though it is mostly for live sound in big venues, it is very relevant for perception in general, how you can develop your listening skill in this regard

 

[Lord Victor]

Hello, see David Griesinger work, a lot of it is on his website also some lectures on youtube. Even though it is mostly for live sound in big venues, it is very relevant for perception in general, how you can develop your listening skill in this regard

Thank you, I'll give it a look.
I've also found this paper. Which would seem to reinforce the notion that it has to do with off axis response and room reflections, similar to Toole's assertion iirc.

I also points out the obvious - that rolled off off-axis response decreases the impact/involvement of the room. Given that the speakers I'm using as a goalpost actually all appear to roll off significantly above 10k off axis, and retain their staging ability in spite of varying rooms; this might actually suggest that for normal listening rooms, without significant treatment, maybe it might in fact be desirable to have a fairly directional speaker, so long as the off-axis rolloff's are even and smooth?
So that rather than having a wide dispersing speaker, bouncing all over the room causing havoc and glare, the "poor" off axis response calms this, causing a much nicer result.

That said, in an optimally treated room, I'd still assume a wide off axis would be ideal.

This does however clash with the fact that these speakers tend to sound good even as you walk around, and don't suffer from the head-in-a-vice issue that some narrow dispersing monitors or electrostats fx. do.
But maybe that's in fact still down to that rolled off >10k meaning there is less directional info to place the sound source based on, giving the illusion of the stage remaining detached from the speakers...

 

[Timcognito]

It's 90% in the recording process and has little to with speakers that test good and a lot to do with the room they are in. Look into recording techniques and you can learn much about soundstage.

 

[tmuikku]

Thank you, I'll give it a look.
I've also found this paper. Which would seem to reinforce the notion that it has to do with off axis response and room reflections, similar to Toole's assertion iirc.

I also points out the obvious - that rolled off off-axis response decreases the impact/involvement of the room. Given that the speakers I'm using as a goalpost actually all appear to roll off significantly above 10k off axis, and retain their staging ability in spite of varying rooms; this might actually suggest that for normal listening rooms, without significant treatment, maybe it might in fact be desirable to have a fairly directional speaker, so long as the off-axis rolloff's are even and smooth?
So that rather than having a wide dispersing speaker, bouncing all over the room causing havoc and glare, the "poor" off axis response calms this, causing a much nicer result.

That said, in an optimally treated room, I'd still assume a wide off axis would be ideal.

This does however clash with the fact that these speakers tend to sound good even as you walk around, and don't suffer from the head-in-a-vice issue that some narrow dispersing monitors or electrostats fx. do.
But maybe that's in fact still down to that rolled off >10k meaning there is less directional info to place the sound source based on, giving the illusion of the stage remaining detached from the speakers...

I've noticed with my speakers in my room, that there is certain distance, or size of stereo listening triangle, that changes perception of stereo image. In short, listening close enough the speakers the local room largely disappears in a way, that the phantom sound gets clarity and focus and envelopment happens. When bit further away from speakers, the sound is more fuzzy blog in front of me and the local room seems to have strong effect on perceived sound. Anyway, there seems to be rather small distance where this shift in perception happens; it is about one step closer and I kind of step inside the recorded sound, or one step further and I'm back in my room. This seems to relate Griesinger work.

Well, what this means? This means, that the two distinct perceptions of stereo seems real to me at least. Everyone likely prefers to listen either one of them, perhaps for all records or perhaps for record to record basis. I think you likely refer the close sound as "head-in-vice" and seem to refer you like the far sound, to walk around. Looking at pictures of peoples listening setups or from hifi show rooms most people seem to like the far sound, but for example me want to hear the close sound.

Now, which ever presentation you want to listen to, and optimize for, likely benefits from different things. If you like the close sound like me, then you likely want to suppress certain early reflections to make it happen, locate and point the speakers so that stereo width is what you want and the stage is natural and the sound doesn't get in your head but nice 3D space. It's big and deep and tall and all that, not head-in-vice if you tune it so but you have to know what you are dealing with.

On the other hand if you like the far sound, a blurry kind of bigness, likely benefits of strong early sidewall reflections that "widen" the stage beyond speakers and so on. Likely quite opposite what how one would adjust on the near field, less toe-in and all that.

But what's even more important is that there is these two perspectives on the stereo sound and sound of your speakers, which you can change at will if you find where the transition happens, and use the two perspectives to learn how to listen how your room affects perceived sound, how speaker directivity and positioning affects sound. You'll learn that some records sound good on either, some times you just want to sit at the transition and zoom in or out as will, or find exact sweet spot. In general you just learn listening a lot, and about various aspects of stereo field.

Anyway, I hope this subjective perspective on this stuff gives you some hints to look for papers. If you didn't know any of this yet I hope it gives you a tool to learn to listen about the stuff at home yourself.

I'm not sure all people have this with various kinds of setups, is it normal or just property of my DIY speakers and room or what. Anyway, it would be interesting to have your comment on it and if you have found the transition at your place and utilize it in listening. I'll follow how the thread develops to get more info about it myself

 

[Lord Victor]

It's 90% in the recording process and has little to with speakers that test good and a lot to do with the room they are in. Look into recording techniques and you can learn much about soundstage.

Oh I'm well aware - thats exactly why I'm trying to figure this out, since as mentioned these speakers, regardless of room (of the 7-10 rooms I've tried them in) tend to present the sound in a certain way, generally bigger and more detached, than others.

I find its more so in headphones, where there are much fewer variables, that the recording becomes 90% dominant - the remainder being down to FR and what few other variables there are. In speakers, in regular (not heavily acoustically treated) rooms, tend to vary much more in their presentation of stage.
Its of course still hugely dependent on the recording - but assuming the same audio is used, there can be a fair bit of variation none the less imo. Which I think is also what the aforementioned research/experiment paper found.

 

[Lord Victor]

I've noticed with my speakers in my room, that there is certain distance, or size of stereo listening triangle, that changes perception of stereo image. In short, listening close enough the speakers the local room largely disappears in a way, that the phantom sound gets clarity and focus and envelopment happens. When bit further away from speakers, the sound is more fuzzy blog in front of me and the local room seems to have strong effect on perceived sound. Anyway, there seems to be rather small distance where this shift in perception happens; it is about one step closer and I kind of step inside the recorded sound, or one step further and I'm back in my room. This seems to relate Griesinger work.

Well, what this means? This means, that the two distinct perceptions of stereo seems real to me at least. Everyone likely prefers to listen either one of them, perhaps for all records or perhaps for record to record basis. I think you likely refer the close sound as "head-in-vice" and seem to refer you like the far sound, to walk around. Looking at pictures of peoples listening setups or from hifi show rooms most people seem to like the far sound, but for example me want to hear the close sound.

Now, which ever presentation you want to listen to, and optimize for, likely benefits from different things. If you like the close sound like me, then you likely want to suppress certain early reflections to make it happen, locate and point the speakers so that stereo width is what you want and the stage is natural and the sound doesn't get in your head but nice 3D space. It's big and deep and tall and all that, not head-in-vice if you tune it so but you have to know what you are dealing with.

On the other hand if you like the far sound, a blurry kind of bigness, likely benefits of strong early sidewall reflections that "widen" the stage beyond speakers and so on. Likely quite opposite what how one would adjust on the near field, less toe-in and all that.

But what's even more important is that there is these two perspectives on the stereo sound and sound of your speakers, which you can change at will if you find where the transition happens, and use the two perspectives to learn how to listen how your room affects perceived sound, how speaker directivity and positioning affects sound. You'll learn that some records sound good on either, some times you just want to sit at the transition and zoom in or out as will, or find exact sweet spot. In general you just learn listening a lot, and about various aspects of stereo field.

Anyway, I hope this subjective perspective on this stuff gives you some hints to look for papers. If you didn't know any of this yet I hope it gives you a tool to learn to listen about the stuff at home yourself.

I'm not sure all people have this with various kinds of setups, is it normal or just property of my DIY speakers and room or what. Anyway, it would be interesting to have your comment on it and if you have found the transition at your place and utilize it in listening. I'll follow how the thread develops to get more info about it myself

I know to an extent what you mean - essentially reaching the nearfield effect where the delay between direct sound and reflection becomes big enough to "eliminate the room" to some minor extent. Which can be fun with some speakers if they suit it. I remember laying on the floor right between my Infinity Kappa 6.2i series ii once listening to field recordings, which was a mesmerizing experience. Like an AKG K1000 on steroids.

But that isn't what I refer to, no. By head in a vice, I mean some speakers simply have an off axis response so jarringly different from the on axis, that in only sounds "right" when sitting in the center - regardless of distance. The paper from University of Surrey did also find listening position to have a fairly minor effect comparatively to axis response (Which I'll admit I do find odd, for the same reason as the experience you mention).

All that said, whilst I don't necessarily think distance has much to do with it in this case; i do like placing the Dynaudio Audience 10 quite wide, rather than equilaterally, as it seems to make them stage even bigger - where as with fx. my monitor audio PL100 the soundstage completely falls apart if I do that.

 

[tmuikku]

Yeah its likely multiple things, attenuating and delaying certain early reflections, and in general increasing direct to reflected sound ratio, which can be simplified as "small enough listening triangle".

Try listening if you hear the transition with your current setup? Walk along center normal between your speakers, put mono white/pink noise on for tight phantom center. Focus on listening size/clarity of the phantom center. Beyond the transition its big and fuzzy, closer the transition it's clarity increases and size shrinks, speakers should seem muted even if you look at them.

If you find it, spend few evenings listening at the transition distance, in general utilize the transition to shift perspective on sound to be able to listen/focus on various aspects of it, like width, or naturalness in various ways, hole in the middle, listen envelopment, play with toe-in, and so on.

I'm curious if you find it useful or not Thanks!

ps. I have rather typical living room with normal furnishing, no special acoustic treatment. Speakers have smooth directivity, cardioid midrange and 90deg nominal waveguide, DI 6-10db from 300Hz to 20kHz. The audible critical distance is a out 2.2m away from speakers. I have no other speakers to compare, or friends to listen on theirs. Going to first hifishow next weekend, curious about if this effect is audible there, and how the various systems sound like

 

[Timcognito]

but assuming the same audio is used, there can be a fair bit of variation none the less imo

That's what DSP is for, to mitigate those effects that obscure the information on recording that has to do with speaker FR inadequacy, distortion, placement and affect of the room. The sound stage is or isn't in the recording depending on mic'ing, mixing and room it's recorded in

 

[Lord Victor]

That's what DSP is for, to mitigate those effects that obscure the information on recording that has to do with speaker FR inadequacy, distortion, placement and affect of the room. The sound stage is or isn't in the recording depending on mic'ing, mixing and room it's recorded in

What you're describing at this point is a semantic issue with the use of the word soundstage. I've had that debate before coming at it from the angle you are - so whilst I agree, largely, DSP or not, the reality is that how the sound actually presents in room, and is perceived, varies a fair amount between speakers, which is well documented afaik in litterature/studies/papers. Regardless of whether we might need better terminology to distinguish the recorded 'stage' and the reproduced 'stage'. Like soundstage vs Apparent Source Width (ASW) - though I think its more than just width, making that term slightly misleading.

I'm trying to figure out which parameters lead to which results; such that, per your point, one might use DSP to achieve it, assuming that is possible - which it doesn't appear it is actually, if it is indeed an effect of directionality of the drivers, which DSP can't to my knowledge, do anything about.

 

[steve59]

How do you measure something's that an artifact of the mind?

 

[NTK]

I'm trying to figure out which parameters lead to which results; such that, per your point, one might use DSP to achieve it, assuming that is possible - which it doesn't appear it is actually, if it is indeed an effect of directionality of the drivers, which DSP can't to my knowledge, do anything about.

Only if you limit the application of DSP to frequency response EQ. BACCH certainly can enhance the spatial sensation of the reproduced sound using DSP. Here is a paper on its crosstalk cancelation process.

https://www.researchgate.net/publication/264886721_Optimal_Crosstalk_Cancellation_for_Binaural_Audio_with_Two_Loudspeakers

 

[Lord Victor]

How do you measure something's that an artifact of the mind?

With a microphone..

 

[Timcognito]

What you're describing at this point is a semantic issue with the use of the word soundstage. I've had that debate before coming at it from the angle you are - so whilst I agree, largely, DSP or not, the reality is that how the sound actually presents in room, and is perceived, varies a fair amount between speakers, which is well documented afaik in litterature/studies/papers. Regardless of whether we might need better terminology to distinguish the recorded 'stage' and the reproduced 'stage'. Like soundstage vs Apparent Source Width (ASW) - though I think its more than just width, making that term slightly misleading.

I'm trying to figure out which parameters lead to which results; such that, per your point, one might use DSP to achieve it, assuming that is possible - which it doesn't appear it is actually, if it is indeed an effect of directionality of the drivers, which DSP can't to my knowledge, do anything about.

Okay but if its not on the recording it can never be there. Optimising your speaker performance in the room is not soundstage although it may enhance hearing it.

 

[Lord Victor]

Okay but if its not on the recording it can never be there. Optimising your speaker performance in the room is not soundstage although it may enhance hearing it.

Like I said, that’s semantics. The reason I gave further description clarification.

I’m talking about the speakers inherent staging/spaciousness, disregarding playback material. Assume pink noise, if that makes it easier/clearer.