On Peer Reviewed Science

[andreasmaaan]

This is what I meant about old measuring better than new:

https://www.community.genelec.com/forum/-/message_boards/view_message/1004384#_19_message_1004384

I will look up Danley.

EDIT: I looked up Danley on a previous occasion but didn’t understand the concept then. Will try again.

Thanks, this is interesting.

Unfortunately, although the OP in that thread makes a lot of statements about then 8531, there are no actual measurements in the link they give to support them. Anyway, I'll take the OP's word for it.

I don't think this is so much a case of the measurements being inadequate. Rather, the measurements do show a clear difference between the two speakers and show that one is superior in one aspect while the other is superior in other aspects.

It's just a question of whether the superior vertical polar response of the 8351 is a worthwhile trade-off for the superior distortion, phase response etc. of the older speaker.

FWIW, I've heard both these monitors and they do sound quite different from each other to me, although I heard them at different times and in different studios.

Let me know if you'd like me to try to explain how the Danleys work (it's complicated and there's a lot of not-quite-right info on forums about it).

 

[svart-hvitt]

Thanks, this is interesting.

Unfortunately, although the OP in that thread makes a lot of statements about then 8531, there are no actual measurements in the link they give to support them. Anyway, I'll take the OP's word for it.

I don't think this is so much a case of the measurements being inadequate. Rather, the measurements do show a clear difference between the two speakers and show that one is superior in one aspect while the other is superior in other aspects.

It's just a question of whether the superior vertical polar response of the 8351 is a worthwhile trade-off for the superior distortion, phase response etc. of the older speaker.

FWIW, I've heard both these monitors and they do sound quite different from each other to me, although I heard them at different times and in different studios.

Let me know if you'd like me to try to explain how the Danleys work (it's complicated and there's a lot of not-quite-right info on forums about it).

I glanced at the Danleys. They are B I G...and they are powerful.

Seems to be an interesting option for the dedicated audio room or home theatre. Noisy amps?

I wonder how their on-off axis response is.

 

[andreasmaaan]

I glanced at the Danleys. They are B I G...and they are powerful.

Seems to be an interesting option for the dedicated audio room or home theatre. Noisy amps?

I wonder how their on-off axis response is.

I don't know of any comprehensive measurements of the commercial designs. Here are some measurements of a DIY version of them that give a good indication of the expected off-axis response of the commercial models. Also interesting to note that they are more or less linear phase.

Wouldn't necessarily recommend them for HT, being PA speakers they are big and heavy and not necessarily as finessed as many monitors and other speakers designed for home use.

If you've been to an IMAX cinema, chances are you've heard some version of them already.

Re: amp noise, most of the models are passive so you could choose your own amps for them.

 

[svart-hvitt]

I don't know of any comprehensive measurements of the commercial designs. Here are some measurements of a DIY version of them that give a good indication of the expected off-axis response of the commercial models. Also interesting to note that they are more or less linear phase.

Wouldn't necessarily recommend them for HT, being PA speakers they are big and heavy and not necessarily as finessed as many monitors and other speakers designed for home use.

If you've been to an IMAX cinema, chances are you've heard some version of them already.

Re: amp noise, most of the models are passive so you could choose your own amps for them.

What do you mean by «finessed»? If a speaker is good, it is «finessed» as well, isn’t it? Or did you talk about the exterior design?

In that case, a dedicated audio room of size should have inbuilt speakers and amps, don’t you think? When a speaker is built into the wall, who cares about the exterior design?

 

[andreasmaaan]

What do you mean by «finessed»? If a speaker is good, it is «finessed» as well, isn’t it? Or did you talk about the exterior design?

In that case, a dedicated audio room of size should have inbuilt speakers and amps, don’t you think? When a speaker is built into the wall, who cares about the exterior design?

Well by finessed I’m talking about design, not aesthetics. I’m just saying I’m not sure they measure as accurately across the whole frequency range, being a passive design that’s very complex. I don’t know though, I can’t find any detailed measurements. Also there are a lot of different models that I’m sure all measure somewhat differently.

I brought this design up not to recommend a specific model, but to draw you attention to a different and very interesting approach to point source speakers.

In any case, I believe the measurements can fully describe the performance of all these speakers

 

[oivavoi]

That thing about the Genelec models is actually very interesting. It's a good example of what I regard as the limits of peer reviewed (audio) science: How does one weigh different aspects of speaker design? It seems undeniable that the 8350 measures better in quite a lot of areas. What it doesn't have, though, is the point source dispersion of the 8351. So how much importance should one give to better phase behavior, ruler flat frequency response and lower THD etc, vs the dispersion pattern of the 8351? Thomas Lund from Genelec actually seems to acknowledge that this is a question without a definitive "scientific" answer:

Given enough time, I rate ears above measurements, but ideas even higher.

This is fascinating, coming from one of the main engineers at a company which is very much in the pro-science camp.

So revising my earlier comment re: the D&D 8Cs vs AVI Dm10: What I may have meant is that above a certain level of engineering competence, I don't think it's obvious what the "correct" way of doing a speaker is. The 8Cs are obviously extremely impressive in myriads of ways. When I move them into the new and slightly bigger listening room, I might perceive them as significantly above the DM10s. But are there trade-offs? The 8Cs achieve extremely even dispersion with frequency, deep bass extension, low THD (I think), and more. But due to the size of the drivers and the wave guide, the drivers are spaced farther apart than with the DM10s, which are small monitors. So less point source behavior. Does the wave guide have an audible impact, beyond controlling dispersion? Does it change the sound waves ever so subtly when it steers them into the room? Perhaps. I don't know. The 8Cs have great power and dynamics - but is efficient class D amplification a trade-off for the tweeter? (some manufacturers of active speakers have used class D for the woofer and class AB for the tweeters, for reasons which I don't think can be completely dismissed). But I assume that for quite a lot of listeners, a comparison between the D&D 8Cs and the AVI DM10s would be a slam-dunk victory for the 8Cs.

So I don't think we are the point yet where we know how the perfect speaker would look. The Morrison omnis, which I'm still patiently waiting for, probably come closest to my "ideas" (to borrow the phrase from Thomas Lund) about what an ideal speaker should do. But even they probably sacrifice some dynamics and bass extension for the price of achieving point source dispersion. There is a legitimate argument to be made that dynamics may trump spatiality and pont source radiation.

 

[andreasmaaan]

@oivavoi I agree in many ways with what you’re saying, but have 2 x 2 = 4 cents to add.

Firstly, not all class D amps perform the same. Many perform worse at high frequencies, but this is not the case for all class D amps by any means. Each designer should look at their budget and determine the most suitable amp for their requirements. Not using class D amps for tweeters as a blanket rule goes against the evidence and good sense. The best class D amps are capable of transparency at all audible frequencies.

Secondly, IMO it’s not that we don’t know what the perfect speaker should look like yet, but rather that there can’t be a single perfect speaker, not even in theory. There’s just no agreed upon dispersion profile at any level of the industry. If we could all agree on this one point, it would be relatively easy to describe the perfect speaker (although not to actually make it with the current state of technology). It’s just that we’re unlikely to ever all agree on it, and in any case it’s likely to remain largely dependent on the playback room, and partly dependent on taste and recording.

However, we can say with some confidence, based on research on reflection audibility and perception, that the more like a point source a speaker behaves, the better it will tend to sound, all else being equal.

Anyway, very interested to hear your impressions of the Morrison omnis.

 

[andreasmaaan]

Oh yeh last point about waveguides.. It does have an audible impact other than directing the sound wave: it reduces nonlinear distortion, while at the same time creating some diffraction effects at high frequencies. The extent of each effect is very dependent on the waveguide and tweeter. I have a strong view on which of these tends to be the lesser evil, however

 

[oivavoi]

@oivavoi I agree in many ways with what you’re saying, but have 2 x 2 = 4 cents to add.

Firstly, not all class D amps perform the same. Many perform worse at high frequencies, but this is not the case for all class D amps by any means. Each designer should look at their budget and determine the most suitable amp for their requirements. Not using class D amps for tweeters as a blanket rule goes against the evidence and good sense. The best class D amps are capable of transparency at all audible frequencies.

Secondly, IMO it’s not that we don’t know what the perfect speaker should look like yet, but rather that there can’t be a single perfect speaker, not even in theory. There’s just no agreed upon dispersion profile at any level of the industry. If we could all agree on this one point, it would be relatively easy to describe the perfect speaker (although not to actually make it with the current state of technology). It’s just that we’re unlikely to ever all agree on it, and in any case it’s likely to remain largely dependent on the playback room, and partly dependent on taste and recording.

However, we can say with some confidence, based on research on reflection audibility and perception, that the more like a point source a speaker behaves, the better it will tend to sound, all else being equal.

Anyway, very interested to hear your impressions of the Morrison omnis.

Excellent points! I agree. Yes, when I think of it, some of the setups I've like the most so far in my life have been running entirely on Class D, for example the wonder Grimm LS1s.

 

[oivavoi]

Oh yeh last point about waveguides.. It does have an audible impact other than directing the sound wave: it reduces nonlinear distortion, while at the same time creating some diffraction effects at high frequencies. The extent of each effect is very dependent on the waveguide and tweeter. I have a strong view on which of these tends to be the lesser evil, however

Interesting. Why does it reduce nonlinear distortion - because the drivers need to move less to move the same amount of air?
When you talk about diffraction - do you have the same thing in mind as Geddes when he talks of HOMs?

 

[Cosmik]

So I don't think we are the point yet where we know how the perfect speaker would look.

We've only had 100 years or so to think about it!

 

[svart-hvitt]

That thing about the Genelec models is actually very interesting. It's a good example of what I regard as the limits of peer reviewed (audio) science: How does one weigh different aspects of speaker design? It seems undeniable that the 8350 measures better in quite a lot of areas. What it doesn't have, though, is the point source dispersion of the 8351. So how much importance should one give to better phase behavior, ruler flat frequency response and lower THD etc, vs the dispersion pattern of the 8351? Thomas Lund from Genelec actually seems to acknowledge that this is a question without a definitive "scientific" answer:




This is fascinating, coming from one of the main engineers at a company which is very much in the pro-science camp.

So revising my earlier comment re: the D&D 8Cs vs AVI Dm10: What I may have meant is that above a certain level of engineering competence, I don't think it's obvious what the "correct" way of doing a speaker is. The 8Cs are obviously extremely impressive in myriads of ways. When I move them into the new and slightly bigger listening room, I might perceive them as significantly above the DM10s. But are there trade-offs? The 8Cs achieve extremely even dispersion with frequency, deep bass extension, low THD (I think), and more. But due to the size of the drivers and the wave guide, the drivers are spaced farther apart than with the DM10s, which are small monitors. So less point source behavior. Does the wave guide have an audible impact, beyond controlling dispersion? Does it change the sound waves ever so subtly when it steers them into the room? Perhaps. I don't know. The 8Cs have great power and dynamics - but is efficient class D amplification a trade-off for the tweeter? (some manufacturers of active speakers have used class D for the woofer and class AB for the tweeters, for reasons which I don't think can be completely dismissed). But I assume that for quite a lot of listeners, a comparison between the D&D 8Cs and the AVI DM10s would be a slam-dunk victory for the 8Cs.

So I don't think we are the point yet where we know how the perfect speaker would look. The Morrison omnis, which I'm still patiently waiting for, probably come closest to my "ideas" (to borrow the phrase from Thomas Lund) about what an ideal speaker should do. But even they probably sacrifice some dynamics and bass extension for the price of achieving point source dispersion. There is a legitimate argument to be made that dynamics may trump spatiality and pont source radiation.

Yes, the quote of @Thomas Lund is interesting. I read it as testament to the fact that measurements are status quo preserving, possibly penalizing innovations that render some of the old measurements obsolete or too narrow to capture the value of an innovation.

I like Lund’s mini experiments that show how a point source has better summation of driver information, and how he demonstrated the power of room DSP in another simple but effective experiment (that I witnessed recently at a Genelec demonstration). He seems to be an Ockham’s razor guy, who is willing to see past measurements if he eyes an idea he believes (yes, I think «belief» is an important driver behind most innovation, including science based ones) is superior.

Having said that, I heard just recently the 8351, 8341 and 8331 in quick ABC shifts. Totally unscientific and prone to errors. Even though the speakers are the «same» apart from size, the sound was a bit different...

It could be due to the importance of the lower frequencies, where size matters, or other variables.... Very interesting in any case. And I would have liked to repeat that little experiment to see if it was just the full moon effect that made all the difference

 

[andreasmaaan]

Interesting. Why does it reduce nonlinear distortion - because the drivers need to move less to move the same amount of air?
When you talk about diffraction - do you have the same thing in mind as Geddes when he talks of HOMs?

Yeh, that’s basically why distortion is reduced.

Re HOMs, yes they are one example, but there are others too depending on the type of driver and the waveguide and how they interact.

 

[andreasmaaan]

interesting. I read it as testament to the fact that measurements are status quo preserving, possibly penalizing innovations that render some of the old measurements obsolete or too narrow to capture the value of an innovation.

Measurements are not inherently status quo preserving. Measurements are the most neutral and objective way to describe a speaker.

Human choices about which measurements to take and/or how to interpret measurements may tend to preserve the status quo.

But please, let’s not attack measurements in and of themselves.

 

[oivavoi]

We've only had 100 years or so to think about it!

Hehe... Well, the audio industry has mostly been a sad state of affairs, methinks. I recently read through some of the old writings of Paul Klipsch, from the 50s and 60s. Very smart guy. Turns out that some of the very first speakers Klipsch made after WWII had active crossovers! At the time, he thought it was a superior technical solution. Then, in the 70s, when they made quite a lot of money on passive speakers, the Klipsch engineers suddenly started arguing against active crossover.

And this is how it's been, I think. Close to everybody who has written about and thought about audio has had financial interests in it. Not everybody, but a sizable number of people. This has hindered innovation and thinking outside the box, and has made manufacturers choose convention time and time again.

EDIT: Here's one of Paul Klipsch's essays, btw. Some of the ideas put forward seem to me to be sound, while others seem more dated. http://www.itishifi.com/2010/05/eight-cardinal-points-in-loudspeakers.html

 

[Fitzcaraldo215]

Interesting discussion. However, I am not convinced of the primacy of point source, myself. It seems beautiful theory, and it is the design tendency for most speakers, even 2, 3-way, etc. albeit with the minor elevation shift among drivers. But, I have liked line source dipolar panels, myself, for many years, including even the delayed, diffused back wave.

But, it there any research anywhere, tough to do, that confirms that humans actually prefer point source as opposed to any other speaker launch pattern? Coaxial speakers, which were the rage in the '50s, don't seem all that popular these days.

So, is point source just a mindset or does it in any meaniful way actually provide higher fidelity?

 

[Cosmik]

So I don't think we are the point yet where we know how the perfect speaker would look.

Probably something like this:

I've always loved the look of them, and I think that converted to active with new drivers, DSP, etc. they might be my ideal speaker. The main reason is that they are not 'tortured' in their design. No horns, waveguides, electrostatic panels, not trying to be compact. Basic monopole 'throw' into the room. Three-way, so giving the drivers an easy life and utilising their optimal ranges. Increasing 'point source-ness' with frequency. Uniformity of dispersion through narrowing baffles with frequency. These are going to be a relaxed, straight speaker. And will sound fantastic.

I can draw these conclusions without any measurements or listening tests; it is all perfectly predictable.

 

[svart-hvitt]

Measurements are not inherently status quo preserving. Measurements are the most neutral and objective way to describe a speaker.

Human choices about which measurements to take and/or how to interpret measurements may tend to preserve the status quo.

But please, let’s not attack measurements in and of themselves.

I am not against measurements per se. Just trying to make the same argument as @amirm when he says standards are too inflexble, hindering innovation etc.

The difference between @amirm and me being that I argued against conventional measurements if one faces real innovation, (which is rare!), while @amirm believes he needs the slack and absence of standards every time he makes measurements of commodity gear.

 

[oivavoi]

Probably something like this:

I've always loved the look of them, and I think that converted to active with new drivers, DSP, etc. they might be my ideal speaker. The main reason is that they are not 'tortured' in their design. No horns, waveguides, electrostatic panels, not trying to be compact. Basic monopole 'throw' into the room. Three-way, so giving the drivers an easy life and utilising their optimal ranges. Increasing 'point source-ness' with frequency. Uniformity of dispersion through narrowing baffles with frequency. These are going to be a relaxed, straight speaker. And will sound fantastic.

I can draw these conclusions without any measurements or listening tests; it is all perfectly predictable.

Yeah, I've had a secret fancy for these babies as well for some time. Did we discuss them on this forum already? They seem to do so many things right at the same time. The white paper on them is amazing. KEF had it figured out (except for the cumbersome passive crossovers, which wasn't a necessity at the time), and then it all went backwards from there instead.

There was a pair for sale in Norway some time ago, and I was very close to purchasing it, also thinking of activating it through DSP. But ultimately I realized that I didn't have the time or the skills needed. Perhaps in the future!

 

[svart-hvitt]

Interesting discussion. However, I am not convinced of the primacy of point source, myself. It seems beautiful theory, and it is the design tendency for most speakers, even 2, 3-way, etc. albeit with the minor elevation shift among drivers. But, I have liked line source dipolar panels, myself, for many years, including even the delayed, diffused back wave.

But, it there any research anywhere, tough to do, that confirms that humans actually prefer point source as opposed to any other speaker launch pattern? Coaxial speakers, which were the rage in the '50s, don't seem all that popular these days.

So, is point source just a mindset or does it in any meaniful way actually provide higher fidelity?

Genelec have never claimed coaxials are a new idea. On the contrary, they underline the heritage of the design and lay out where their innovation is. See this paper for details:

https://www.genelec.com/documents/other/Genelec 8260A Technical Paper.pdf

This way of arguing - where you describe the «literature» and status quo, as you would do in a scientific article in any field - is often lacking in commercial audio where for example the cardioid bass response of Kii and DD has been put forward by many as an innovation despite the fact that for example Gradient pioneered this design in commercially available speakers in the mid 1990s.

So I agree: Innovations are rare. The best design ideas in audio are probably old. What has been lacking is implementation and the tools to overcome some of the distortion byproducts of otherwise superior ideas.