A question about what is measurable.

[fas42]

If I understand the scenario correctly, it implies that the SS amp was changing its behaviour due to the less than perfect loading of the speaker; the triode was relatively unphased by it, in contrast. The dummy speaker load was not realistic as a replacement for the actual speaker.

I would test that by removing the triode, driving the speakers with SS at the relevant volume, tapping the signal at the speaker connectors, feeding them, appropriately attenuated, to a normal headphone amp, and listening to the sound on headphones, in another room, say, feeding the leads under the door; then repeat the exercise with the dummy speaker resistors replacing the speakers and speaker cable ... does that tell us anything?

 

[Blumlein 88]

If I understand the scenario correctly, it implies that the SS amp was changing its behaviour due to the less than perfect loading of the speaker; the triode was relatively unphased by it, in contrast. The dummy speaker load was not realistic as a replacement for the actual speaker.

I would test that by removing the triode, driving the speakers with SS at the relevant volume, tapping the signal at the speaker connectors, feeding them, appropriately attenuated, to a normal headphone amp, and listening to the sound on headphones, in another room, say, feeding the leads under the door; then repeat the exercise with the dummy speaker resistors replacing the speakers and speaker cable ... does that tell us anything?

I didn't include all the variations in my discussion with Opus 111. Though I earlier posted about it. I also swapped position of the amplifiers. Wadia to Triodes (with power resistor loading stepped down to unity gain) to SS amp. You got the full expansive sound of the triode in that connection. The best conclusion that fits all the way around is the expansive layered soundstage is an artefact of the triode amp, and the SS amp was accurately able to portray that signal when presented with it, while not interfering with it when placed between source and triode amp.

 

[wgscott]

In a discussion of forum's purpose and the kinds of positions that are appropriate here, J Kenny just asked, "Amir, are topics to be avoided which "accepted audio science/engineering" have nothing to offer? Topics like soundstage depth, timbre just to name two.
Are topics only allowed if measurements "proving" their audibility are available?"

And it leads me to another question: In my relative ignorance compared to many of you here, I would assume that "soundstage depth," is a function of the relative volumes of sounds in the mix, and "soundstage" itself is a nontechnical term for a combination of that depth, and the placement of sounds in the stereo mix. I would further assume that playback equipment's ability to reproduce that stage properly is a function of well-defined and quite measurable properties like dynamic range, noise, plus distortions like IMD, jitter or harmonic distortion that can make sounds seem more or less prominent than they are on the recording, channel separation, etc. I would assume that "timbre" is similarly a function of FR, noise and distortion.

All much more clearly defined terms than "soundstage" and "timbre," all measurable. What have I missed?

Tim

Sorry if this is redundant, but I haven't read beyond your post. "Soundstage" to me should be the ability of a stereo system to reconstruct the spatial arrangement of performers, instruments, and so forth. It is what should collapse if you hit the "mono" button, so the difference should be readily quantifiable, and therefore measurable (in terms of phases differences and interference patterns, in analogy with optical image reconstruction). What remains after you hit the mono button should be pure and uncolored; otherwise it falls into the category of audiophile sound effects.

 

[fas42]

I hadn't registered what Tim stated about soundstage until just now - and would add that it is certainly more than loudness. It is exactly equivalent to a teacher in a classroom hearing one of his boys making a sound, and instantly being able to flick his eyes to the position of the wrongdoer, laterally and in distance - he literally heard, in 3D space, where the sound was coming from. And competent playback achieves the same thing, for the listener - you can lock your mental focus onto where an individual sound is coming from, in a space behind the speakers, and "watch" it.

Remarkably, mono largely replicates this! There is no difficulty in looking into the distance to see individual performers, and the brain seems to conjure a pseudo width to match, of sorts. No lateral placement, of course, but still very satisfying.

To me it demonstrates the remarkable ability of the ear/drum to recreate the original event, if it has enough data - it can "fill in the gaps" to an extraordinary degree - I was truly amazed when I first heard it happen, for me.

 

[Blumlein 88]

Sorry if this is redundant, but I haven't read beyond your post. "Soundstage" to me should be the ability of a stereo system to reconstruct the spatial arrangement of performers, instruments, and so forth. It is what should collapse if you hit the "mono" button, so the difference should be readily quantifiable, and therefore measurable (in terms of phases differences and interference patterns, in analogy with optical image reconstruction). What remains after you hit the mono button should be pure and uncolored; otherwise it falls into the category of audiophile sound effects.

I generally agree, but one can alter stage width in ways that collapse perfectly to mono. One way is to break out the L minus R signal. Then boost it 4 dB below 700 hz. Mix it back into the stereo. This will widen, slightly deepen, and soften the soundstage. Collapses to mono fine.

 

[Phelonious Ponk]

I hadn't registered what Tim stated about soundstage until just now - and would add that it is certainly more than loudness. It is exactly equivalent to a teacher in a classroom hearing one of his boys making a sound, and instantly being able to flick his eyes to the position of the wrongdoer, laterally and in distance - he literally heard, in 3D space, where the sound was coming from. And competent playback achieves the same thing, for the listener - you can lock your mental focus onto where an individual sound is coming from, in a space behind the speakers, and "watch" it.

Remarkably, mono largely replicates this! There is no difficulty in looking into the distance to see individual performers, and the brain seems to conjure a pseudo width to match, of sorts. No lateral placement, of course, but still very satisfying.

To me it demonstrates the remarkable ability of the ear/drum to recreate the original event, if it has enough data - it can "fill in the gaps" to an extraordinary degree - I was truly amazed when I first heard it happen, for me.

I would agree that soundstage is more than loudness. The rest would require a remarkable amount of imagination in stereo, even more in mono.

Tim

 

[fas42]

I would agree that soundstage is more than loudness. The rest would require a remarkable amount of imagination in stereo, even more in mono.

Tim

Tim, the remarkable thing is that it happens all on its lonesome - and, you can't make it go away. You don't decide that there is no soundstage, and the illusion vanishes, by mentally poo-pooing it. This is rock solid, untouchable stuff - but, I would say that there are some people who would not pick it happening, simply because their brains were wired somewhat differently.

As is stated over and over again, how to measure it? Apart from normal listening tests, I haven't answers as yet - I would need to be motivated to create interesting audio test files, and I don't see the interest yet to make it worth my while.

 

[Phelonious Ponk]

Tim, the remarkable thing is that it happens all on its lonesome - and, you can't make it go away. You don't decide that there is no soundstage, and the illusion vanishes, by mentally poo-pooing it. This is rock solid, untouchable stuff - but, I would say that there are some people who would not pick it happening, simply because their brains were wired somewhat differently.

As is stated over and over again, how to measure it? Apart from normal listening tests, I haven't answers as yet - I would need to be motivated to create interesting audio test files, and I don't see the interest yet to make it worth my while.

No, I would think not.

Tim

 

[Blumlein 88]

I would imagine you could use a goniometer using it for both left and right sides matrixed with L-R and L+R signals for a formal measure of soundstage. This on your source file. As for what you hear, like so much in audio you are lost without a reference. More sound stage is not automatically better.

With everything prior to the speaker being low distortion, flat response and near infinite channel separation it will not impact soundstage other than in someone's imagination. Or when the above parameters are compromised.

So if I had a metric for soundstage and you get something different, it is mostly going to be down to speakers and room setup.

So this business about soundstage not having a metric is just one more subjective regression from evidence. Much like God in the gaps regression among believers.

 

[John Kenny]

I would love to see some measurements of soundstage from the Jim LeSurf's interferometry technique that AJ mentioned. LeSurf is usually on the ball but I wonder why no posted results on his website?

He did something similar in his "IQ test" which was a quadrature technique for measuring phase (timing) differences of replay chains & published some results from a DACMagic & a Haldie Bridge which showed differences in measurements.

But nothing posted for the Interferometry technique - maybe there were some practical issues with it?

 

[March Audio]

I would imagine you could use a goniometer using it for both left and right sides matrixed with L-R and L+R signals for a formal measure of soundstage. This on your source file. As for what you hear, like so much in audio you are lost without a reference. More sound stage is not automatically better.

With everything prior to the speaker being low distortion, flat response and near infinite channel separation it will not impact soundstage other than in someone's imagination. Or when the above parameters are compromised.

So if I had a metric for soundstage and you get something different, it is mostly going to be down to speakers and room setup.

So this business about soundstage not having a metric is just one more subjective regression from evidence. Much like God in the gaps regression among believers.

Spot on, people talk about soundstage without any reference and presume more is better or more accurate. In the 80's every ghetto blaster had a stereo wide switch.

You echo comments I made elsewhere, that if significant differences in SS are noted between amps or dacs, then one the items technical performance is deficient in some area.

God in the gaps . Nice one

 

[Fitzcaraldo215]

Spot on, people talk about soundstage without any reference and presume more is better or more accurate. In the 80's every ghetto blaster had a stereo wide switch.

You echo comments I made elsewhere, that if significant differences in SS are noted between amps or dacs, then one the items technical performance is deficient in some area.

God in the gaps . Nice one

I agree. Soundstage is an ill-defined term, therefore almost meaningless. I think it is just a metaphor, really, perhaps imperfectly alluding to certain aspects of sonic perception. And, those may vary subjectively depending on the listener. It might be about perceived width and depth, but it also might get into questions of image focus vs. spatial smearing of sounds within the overall sonic image. Measurements would seem hopeless unless it is better defined in the first place.

So, since soundstage is meaningless and unmeasurable, I will take off my objectivist hat and wallow in some anecdotal subjective observations. I have heard differences in "soundstage", however defined, and they tended not to be about soundstage width with the same speaker/room setup. They have tended to be more in the image depth and focus dimensions I cited. Is this more accurate to what is on the recording? I have generally tended to think so, perhaps delusionally, and I generally preferred equipment that offered a greater sense of plausible depth and better image focus. This probably does relate to equipment technical performance somehow, but we might never know exactly which performance measures. The time domain performance might be something to look at, but maybe it is much more than that.

The ghetto blasters typically increased apparent soundstage width but at the expense of depth and image smearing. I try to listen for a plausible sense of apparent instrument size within the image, myself: does the saxophone sound more or less focused at a single point in space or does its apparent size become more blurred and less distinct?

Not to belabor my love for Mch sound, but done right, it seems to flesh out the instruments more plausibly for me. The cello in a string quartet sounds like a more coherent 3D body of an instrument, less like the 2D cardboard cutout I hear in stereo by comparison. But, that is me. YMMV.

The discussion of L+R/L-R reminded me of the old Hafler Dynaquad circuit from the early 70's. That simply routed L-R with a separate level control to one or two speakers, normally situated behind the listener - early simulated Mch. I found it very interesting, but I never installed it permanently.

 

[John Kenny]

I agree. Soundstage is an ill-defined term, therefore almost meaningless. I think it is just a metaphor, really, perhaps imperfectly alluding to certain aspects of sonic perception. And, those may vary subjectively depending on the listener. It might be about perceived width and depth, but it also might get into questions of image focus vs. spatial smearing of sounds within the overall sonic image. Measurements would seem hopeless unless it is better defined in the first place.

So, since soundstage is meaningless and unmeasurable, I will take off my objectivist hat and wallow in some anecdotal subjective observations. I have heard differences in "soundstage", however defined, and they tended not to be about soundstage width with the same speaker/room setup. They have tended to be more in the image depth and focus dimensions I cited. Is this more accurate to what is on the recording? I have generally tended to think so, perhaps delusionally, and I generally preferred equipment that offered a greater sense of plausible depth and better image focus. This probably does relate to equipment technical performance somehow, but we might never know exactly which performance measures. The time domain performance might be something to look at, but maybe it is much more than that.

The ghetto blasters typically increased apparent soundstage width but at the expense of depth and image smearing. I try to listen for a plausible sense of apparent instrument size within the image, myself: does the saxophone sound more or less focused at a single point in space or does its apparent size become more blurred and less distinct?

Not to belabor my love for Mch sound, but done right, it seems to flesh out the instruments more plausibly for me. The cello in a string quartet sounds like a more coherent 3D body of an instrument, less like the 2D cardboard cutout I hear in stereo by comparison. But, that is me. YMMV.

The discussion of L+R/L-R reminded me of the old Hafler Dynaquad circuit from the early 70's. That simply routed L-R with a separate level control to one or two speakers, normally situated behind the listener - early simulated Mch. I found it very interesting, but I never installed it permanently.

Yep, you got it, Fitz!

 

[Karl-Heinz Fink]

Hi,

during our work on cabinet vibrations, we found a very clear relation between radiating walls of a speaker cabinet and soundstage width, height and depth. I think this is down to some psychoacoustic effect happening in our brain. In case of or cabinet vibrations, a lively side wall with resonances in a certain midrange area, would make the soundstage wide, but no very deep or focused. So you get a second sound source with it's own directivity and a resonance pattern, depending on the construction. The top panel of a cabinet also does something with the perceived height of the sound stage. You can make a simple experiment to test it. Take a speaker with no special treatment for lowest cabinet vibrations and listen to it. Now take some magazines in the right size (internally, the test is called HiFi+ test, because of the magazines format), place it on top and listen again. In case your setup and room is half way decent, you will notice a change in soundstage height. Did that many times and most people spotted it without knowing why I was trying it.
So that does not mean we can measure sound stage, but we can measure output from cabinet panels and hear the effect of them. A well made cabinet, always sounds a more narrow (BTW, some people prefer the wider stage of the radiation cabinets), but with a much better focus in the middle. A lot of research went into "Head Related Transfer Functions" and if you look into the curves you see some narrow band peaks our brain uses to identify directions. However, reducing unwanted cabinet radiation, cannot be wrong.

Best KH

 

[h.g.]

Now take some magazines in the right size (internally, the test is called HiFi+ test, because of the magazines format), place it on top and listen again. In case your setup and room is half way decent, you will notice a change in soundstage height. Did that many times and most people spotted it without knowing why I was trying it.

Are we changing the radiation from the panel or the diffraction from the top edge? Perhaps both? Do large cabinets suffer significantly more than small cabinets? Does it help explain the baffling popularity of expensive small 2 way speakers on stands? It is an interesting topic.

So that does not mean we can measure sound stage,

I suspect it may be straightforward to measure if people would first agree what it meant. Or is only a vague notion of what it means very much part of it's definition so that it can be used as part of the "subjective audiophile" process?

However, reducing unwanted cabinet radiation, cannot be wrong.

Even if it leads to fewer sales and audiophiles getting less pleasure from the product? Of course I wholly agree for the subset of people interested in high fidelity for the home but how large a group is that?

 

[Karl-Heinz Fink]

Are we changing the radiation from the panel or the diffraction from the top edge? Perhaps both? Do large cabinets suffer significantly more than small cabinets? Does it help explain the baffling popularity of expensive small 2 way speakers on stands? It is an interesting topic.


I suspect it may be straightforward to measure if people would first agree what it meant. Or is only a vague notion of what it means very much part of it's definition so that it can be used as part of the "subjective audiophile" process?


Even if it leads to fewer sales and audiophiles getting less pleasure from the product? Of course I wholly agree for the subset of people interested in high fidelity for the home but how large a group is that?

1. I'm talking about panel vibration. Yes, large cabinets suffer more and are more difficult to get right. And yes, that makes small speakers "better"

2. Not sure what is so complicated with defining "soundstage". But OK, this is something we deal with every day.....

3. No, no, that's not the case, as it gives you a lot more in the center between the speakers. However, I personally don't like the pseudo image that creates large "room"s from a dead dry studio recording.

Best KH

 

[amirm]

It is an interesting point of measuring the cause for a complicated effect. The former can be much easier as you mention KH.

I am always bothers by excessive cabinet movement in high-end speakers. I like to see an inert enclosure or close to it, regardless of audible effect.

 

[Karl-Heinz Fink]

Hi Amir,

I think there are a few nicely made cabinet like some Wilsons or Magico, but they use the Sledgehammer method....fine, but not what I like to do. I don't think there is a cabinet that is completely quiet - you just have to shift the problems to the right area. The BBC made that already in Studio monitors, shifting the problems down in frequency and kept the midband clean. We also found that it is a lot better to have a cabinet with some sort of cabinet "noise", but without huge peaks sticking out of the noise instead of a super, duper low average "noise" level, but also some prominent spikes. We are lucky that we can do a lot with simulation before we cut the fist piece of wood .. and that's a big help. The technology itself is not rocket science. Sort of constrained layer damping with modern damping material inside and the right bracing on the right spot. That's it

Best KH

 

[amirm]

Thanks KH. What do you use for measurements of prototypes? Piezo strips or something else?

 

[NorthSky]

In a discussion of forum's purpose and the kinds of positions that are appropriate here, J Kenny just asked, "Amir, are topics to be avoided which "accepted audio science/engineering" have nothing to offer? Topics like soundstage depth, timbre just to name two.
Are topics only allowed if measurements "proving" their audibility are available?"

And it leads me to another question: In my relative ignorance compared to many of you here, I would assume that "soundstage depth," is a function of the relative volumes of sounds in the mix, and "soundstage" itself is a nontechnical term for a combination of that depth, and the placement of sounds in the stereo mix. I would further assume that playback equipment's ability to reproduce that stage properly is a function of well-defined and quite measurable properties like dynamic range, noise, plus distortions like IMD, jitter or harmonic distortion that can make sounds seem more or less prominent than they are on the recording, channel separation, etc. I would assume that "timbre" is similarly a function of FR, noise and distortion.

All much more clearly defined terms than "soundstage" and "timbre," all measurable. What have I missed?

Tim

If ASR was strictly for objective members, the science would be omitting one of the main fundamental rules; exploration.

On your second question; not everything can be measured, was, or will be. And you are no more ignorant than the smartest one.
Hey, you can measure an audio product, then another exact same one and they'll measure differently.
Example: A professional Audio Mag is reviewing an audio product, but the measurements are below audio standards. The reviewer likes how it sounds (we cannot blame his ears, they're his), but the editor decides to check another same product from the same manufacturer just to make sure. They get the new product but it measures even worst. They don't publish it. Then a third one is sent to the mag's headquarters, and that one measures better (take 2).
Ok, it's not common, but how can we measure if it is not common or not?

Anyway, without subjectivity the world of science would be in greater chaos...IMO.
In particular in this audio hobby.

There are guides to help us; and @ the end it's us who climb Everest with our two feet and lungs.
After we summit it, if we die in our descent could be related to our ascent.
I think audio is like that; and soundstage is a personal view, with width, depth, frequency response, lows, mids and highs.
Some music recordings and loudspeakers are better matches to our personal hearing aptitudes (age related , +++) and to our soul.
If you want to measure the things you just mentioned you better first measure each room and each set of ears from each person.

Balance...