About the "sweet spot"

[Krunok]

Briefly on dispersion pattern: My experience so far is that there might be a trade-off between laser sharp imaging and wide dispersion/ambience/envelopment.

I tried to toe in my speakers so the lines are crossing in front of my LP but I wasn't satisfied with the results. I didn't notice any improvements but the soundstage shrinked considerably so I returned it to as it was - crossed app 1m after my LP.

 

[svart-hvitt]

Briefly on dispersion pattern: My experience so far is that there might be a trade-off between laser sharp imaging and wide dispersion/ambience/envelopment. I say "might" because there is very little research on this, and the research that exists points in different directions (some studies claim that early reflections degrade imaging, while others don't find any such effect). I believe that Geddes is about to publish a study on this though.

The trade-off then becomes this, according to my anecdotal experiences: With wide imaging, you can move around without the tonality changing too much, there's a nice and enveloping ambiance to the music, and with true omnis you can get a stereo image between the speakers almost from anywhere in the room. When sitting exactly in the sweet spot though, the stereo image becomes more "airy" and a bit less sharp. With narrow dispersion and/or heavy acoustic treatment, the stereo imaging and often the tonality collapses when you move outside the sweetspot. Music is also more dry and less enveloping. But when sitting exactly in the sweet spot, imaging is more precise.

So what to choose? I think it partly depends on personal preference, and partly on the kind of music one mainly listens to. I listen 80 percent to classical and jazz, and on that kind of acoustic music I find that narrow dispersion and/or no reflections feels sterile and wrong. On electronic and processed studio recordings though, which 99 percent of audiophiles mostly listen to these days, I find that more narrow dispersion feels more correct.

The obvious choice is therefore to have two different systems, one wide and one narrow! Or to choose a boring compromise with speakers that are neither true omnis nor narrow-dispersion speakers. Or multichannel, as has already been pointed out.

PS: I'm also still waiting patiently for the Morrison omnis which I ordered a long long time ago... but which have become heavily delayed due to health issues for the two guys in charge. Everybody who have heard them seems to think that they image just as well - or even better - than narrow-dispersion speakers. We'll see what I think when I finally get them

I am not convinced that your conviction about omni holds true in mathematical terms as you move around in the room. Summation of information will change as your position changes; physics or mathematics 101, isn´t it?

If the omni is a point source like John Watkinson´s design, I believe, however, that the sound will stay enjoyable across the room. But for stereo to work, one needs correct information summation, wouldn´t you think?

My own experience with point source is that you start to hear more from the nearest speaker as I move around. And this is logical, isn´t it?

One more thing: As you start to get more indirect sound at the same time as direct sound, hifi as an ideal breaks down, doesn´t it? We will never be able (in a domestic setting) to eliminate indirect sound, but we can arrange so that the ratio direct/indirect sound that first enters the ear is maximized in order to approach the hifi ideal. But hifi and preferences are two different things and preferences will always win. I think that´s why I convince myself that my preference is hifi - just to keep things as simple as possible

 

[andreasmaaan]

I tried to toe in my speakers so the lines are crossing in front of my LP but I wasn't satisfied with the results. I didn't notice any improvements but the soundstage shrinked considerably so returned it to as it was - crossed app 1m after my LP.

This makes sense. The reflections off the side walls would tend to widen the soundstage. Toeing in the speakers would reduce the level of these reflections.

 

[oivavoi]

I am not convinced that your conviction about omni holds true in mathematical terms as you move around in the room. Summation of information will change as your position changes; physics or mathematics 101, isn´t it?

If the omni is a point source like John Watkinson´s design, I believe, however, that the sound will stay enjoyable across the room. But for stereo to work, one needs correct information summation, wouldn´t you think?

My own experience with point source is that you start to hear more from the nearest speaker as I move around. And this is logical, isn´t it?

One more thing: As you start to get more indirect sound at the same time as direct sound, hifi as an ideal breaks down, doesn´t it? We will never be able (in a domestic setting) to eliminate indirect sound, but we can arrange so that the ratio direct/indirect sound that first enters the ear is maximized in order to approach the hifi ideal. But hifi and preferences are two different things and preferences will always win. I think that´s why I convince myself that my preference is hifi - just to keep things as simple as possible

Concerning omnis and stereo imaging outside the sweetspot, I can at least relate what my own ears have told me very clearly: On two occasions I've listened to proper omni speakers - the MBL 101 and one of the German Physiks models. While neither of these speakers are perfect, I had the exact same experience on both occasions: The stereo image seemed locked between the speakers wherever I moved in the room. Yes, it could move a bit, but it was still there. There will surely be an explanation for this, but I'm 100 percent sure that my ears didn't fool me THAT much.

On indirect sound and high fidelity: That's a big debater!

 

[svart-hvitt]

Concerning omnis and stereo imaging outside the sweetspot, I can at least relate what my own ears have told me very clearly: On two occasions I've listened to proper omni speakers - the MBL 101 and one of the German Physiks models. While neither of these speakers are perfect, I had the exact same experience on both occasions: The stereo image seemed locked between the speakers wherever I moved in the room. Yes, it could move a bit, but it was still there. There will surely be an explanation for this, but I'm 100 percent sure that my ears didn't fool me THAT much.

On indirect sound and high fidelity: That's a big debater!

Why doesn’t one (1)* mastering engineer use omnis?


*Just a wild guess Do you know of any MEs using omnis?

 

[andreasmaaan]

Why doesn’t one (1)* mastering engineer use omnis?


*Just a wild guess Do you know of any MEs using omnis?

More room interaction = more opportunity for the tonal balance to be distorted.

 

[oivavoi]

Why doesn’t one (1)* mastering engineer use omnis?


*Just a wild guess Do you know of any MEs using omnis?

Probably partly because omnis are very uncommon? So they don’t want to master on speakers which are extremely different from the speakers most people listen on? I read a study which indicated that mastering engineers added different levels of reverb, depending on whether they were mastering in a reverberant or dry environment. So it’s probably smart to have speakers and acoustics which are not too different from how people usually listen.

 

[oivavoi]

Also, the sound stage of omnis is quite room dependent, as they engage the room maximally. Might make it more difficult to achieve translation to a variety of speakers and rooms, compared to speakers with more narrow dispersion.

 

[Theo]

Why doesn’t one (1)* mastering engineer use omnis?

I guess the answer is in this excellent post :

With narrow dispersion and/or heavy acoustic treatment, the stereo imaging and often the tonality collapses when you move outside the sweetspot. Music is also more dry and less enveloping. But when sitting exactly in the sweet spot, imaging is more precise.

 

[oivavoi]

I guess the answer is in this excellent post :

True! But mastering engineers are also known to listen in more reverberant environments than mixing engineers, and generally try to listen in environments that are closer to a "normal" listening environment than what is the case in many studios. But it might be that omnis become too much of a good thing, even for them.

 

[Wombat]

True! But mastering engineers are also known to listen in more reverberant environments than mixing engineers, and generally try to listen in environments that are closer to a "normal" listening environment than what is the case in many studios. But it might be that omnis become too much of a good thing, even for them.

What is a more normal listening environment? How do they know? Industry surveys or their own guesses? Shifting sands stuff.

I say make the recording as true as possible to the performance or as desired for release and leave the listening environment stuff to others.

 

[Theo]

Anyway, I have the feeling that Mastering (ie compression...) is degrading the recording fidelity (or what's left of it...) more than the contrary. A normal listening environment? Let's try a Homepod...

 

[Wombat]

Anyway, I have the feeling that Mastering (ie compression...) is degrading the recording fidelity (or what's left of it...) more than the contrary. A normal listening environment? Let's try a Homepod...

I have no control over the quality of releases so don't get my undies in a twist over it. If I don't like it I don't buy/listen to it. I am not the centre of the universe and realise it. Nothing worth getting anal about - just buy all of that good stuff that is out there. Too much time spent on 'how it could be'.

There are those so up the path that they provide DR information. Not so important to me. I enjoy music rather than analyse it.

 

[svart-hvitt]

Genelec go through the history and drawbacks of coaxial designs in the link below. Are there any drawbacks you had in mind that were not addressed in this technical paper from 2009?

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

EDIT: I added an AES paper on the Genelec coaxial driver design: https://www.genelec.com/sites/default/files/media/About Us/Academic_Papers/aes_142_makivirta_et_al_paper_95_2017-04-03_v_1.0.2.pdf

@andreasmaaan , you never came back to this post. Did you read the technical and AES paper? Would you still say distortion and SPL are problems compared to conventional designs?

 

[andreasmaaan]

@andreasmaaan , you never came back to this post. Did you read the technical and AES paper? Would you still say distortion and SPL are problems compared to conventional designs?

Thanks @svart-hvitt, didn't mean to let this slip.

The technical paper was interesting. The eradication of physical discontinuities between the tweeter and midrange must be helpful compared with earlier coaxial designs. The use of spaced outlets for the low-frequency transducer is also a clever way to extend directivity control further down into the lower midrange without spacing the woofers apart from the coaxial arrangement in the centre.

OTOT, this current generation of Genelec coaxials are higher distortion and lower SPL than their spaced designs (IIRC), so it seems these advances in their coaxial designs are improvements, but not complete solutions.

In the technical paper, I was also a bit sceptical of the multitone distortion test and the subjective evaluation that they chose to use. For the distortion test, it isn't specified which other speakers the new coaxials are compared against (presumably not Genelec's best non-coaxial designs), while in the subjective evaluation, they chose a test that was designed to emphasise the issue with spaced drivers and that in no way resembled normal use.

So I think this line of R&D is useful but am still not 100% sold on it's benefits in the widest possible range of applications.

Finally, I'm also not 100% sure I agree that the problems with vertically spaced drivers are as significant as the paper claims. This is what the paper says:

For vertically stacked transducers, movement in the up-down direction changes the timing between the transducers, and this can generate audible tonal changes at the crossover frequencies. Such crossover coloration makes subjective evaluation unreliable in directions where the geometry between the transducers changes with listener location or head movement. Head movements are a stabilizing factor used by listeners when evaluating and controlling sound stage and imaging [11].

The first part of this I agree with, although this will only be a problem when the listener is not positioned on the correct vertical axis.

The second part is is absolutely correct when it comes to horizontal head movements (i.e. we move our heads from side to side to help establish ITD), but AFAIK it is incorrect when it comes to the vertical plane. Our ears are horizontally spaced, so moving our heads up and down just doesn't help us in terms of localisation. So as far as I know, we just don't do it. I may be wrong on this point, however.

I seem to recall you posted an interesting lecture or presentation on this topic at one point which I never got round to watching and then lost. Can you recall which one I'm talking about?

 

[Cosmik]

True! But mastering engineers are also known to listen in more reverberant environments than mixing engineers, and generally try to listen in environments that are closer to a "normal" listening environment than what is the case in many studios. But it might be that omnis become too much of a good thing, even for them.

Do you know what mastering engineers might actually do to the recording in the case of listening to it in a reverberant vs. dry environment? It's not obvious to me how their mastering decisions would change - except for the frequency response discrepancies of real speakers due to their dispersion characteristics. They presumably might change the EQ in an attempt to partially correct the discrepancy. But what would they choose as their 'typical' speaker..?

 

[andreasmaaan]

But what would they choose as their 'typical' speaker..?

For the past 40 years, Yamaha NS10, possibly with a piece of tissue over the tweeter

 

[svart-hvitt]

Thanks @svart-hvitt, didn't mean to let this slip.

The technical paper was interesting. The eradication of physical discontinuities between the tweeter and midrange must be helpful compared with earlier coaxial designs. The use of spaced outlets for the low-frequency transducer is also a clever way to extend directivity control further down into the lower midrange without spacing the woofers apart from the coaxial arrangement in the centre.

OTOT, this current generation of Genelec coaxials are higher distortion and lower SPL than their spaced designs (IIRC), so it seems these advances in their coaxial designs are improvements, but not complete solutions.

In the technical paper, I was also a bit sceptical of the multitone distortion test and the subjective evaluation that they chose to use. For the distortion test, it isn't specified which other speakers the new coaxials are compared against (presumably not Genelec's best non-coaxial designs), while in the subjective evaluation, they chose a test that was designed to emphasise the issue with spaced drivers and that in no way resembled normal use.

So I think this line of R&D is useful but am still not 100% sold on it's benefits in the widest possible range of applications.

Finally, I'm also not 100% sure I agree that the problems with vertically spaced drivers are as significant as the paper claims. This is what the paper says:



The first part of this I agree with, although this will only be a problem when the listener is not positioned on the correct vertical axis.

The second part is is absolutely correct when it comes to horizontal head movements (i.e. we move our heads from side to side to help establish ITD), but AFAIK it is incorrect when it comes to the vertical plane. Our ears are horizontally spaced, so moving our heads up and down just doesn't help us in terms of localisation. So as far as I know, we just don't do it. I may be wrong on this point, however.

I seem to recall you posted an interesting lecture or presentation on this topic at one point which I never got round to watching and then lost. Can you recall which one I'm talking about?

@andreasmaaan , you wrote:

«OTOT, this current generation of Genelec coaxials are higher distortion and lower SPL than their spaced designs (IIRC), so it seems these advances in their coaxial designs are improvements, but not complete solutions».

You need to elaborate on that (with data). I don’t understand what you mean when you claim - without support - that the coaxials yield lower SPL and higher distortions than their conventional(er) designs. You make it sound as if Genelec’s new generation monitors are a step back in terms of distortion products that are - based on psychoacoustic research - audible. That makes me curious! (However, I do know that this claim on distortion is around on different forums, however test methodology, process and psychoacoustic relevance are always lacking).

You also write:

«Finally, I'm also not 100% sure I agree that the problems with vertically spaced drivers are as significant as the paper claims».

Remember, Genelec make both point source and conventional stacked designs. So they have an interest in both camps, so to speak. Their flagship 1236 is heavily (pun intended) in the stacked camp with 1 tweeter, 2 mids and 2 woofers (horizontal). In my experience, researchers who have interests aligned across different camps, are better positioned to talk straight instead of defending one position. Take a look at the directivity plots of the 1236 flagship again. It’s horrible - according to Genelec’s own data - as soon as you wander off-axis (and probably as you approach closer to the monitor as well). If you ever heard the sound of a conventional monitor recorded by a microphone vs a point source recorded by the same microphone - you would see this point clearer. Sure, in a dead control room where one mastering engineer sits tight in sweet spot, the point source advantage is not as big (though it’s always advantegeous) as in a domestic situation of multiple listeners and heavy contribution from the room. Because the point source is rare, we may not have developed measures yet to gauge its advantages in objective, data based terms. In the meantime, practical listening may be a pragmatic «test».

The Genelec «lecture» you mentioned, it may be this video with @Thomas Lund ?

 

[andreasmaaan]

The Genelec «lecture» you mentioned, it may be this video with @Thomas Lund ?

I think that's it, thanks. Will watch it this evening when I get home.

You need to elaborate on that (with data). I don’t understand what you mean when you claim - without support - that the coaxials yield lower SPL and higher distortions than their conventional(er) designs. You make it sound as if Genelec’s new generation monitors are a step back in terms of distortion products that are - based on psychoacoustic research - audible. That makes me curious! (However, I do know that this claim on distortion is around on different forums, however test methodology, process and psychoacoustic relevance are always lacking).

Ironically, I actually thought I recalled you making this claim, along with a link to some DIY measurements Perhaps you were merely pointing to the claim to criticise it? Actually, when posting earlier I thought, "Well if @svart-hvitt says the coaxials have higher distortion then the coaxials must have higher distortion."

Anyway, yes you're right, I've no definitive data on this so I have to qualify my comment by saying that more data is needed. I have access to one of the new generation of coaxials (8351 IIRC) actually, so I'll measure it after the xmas/new year crazies die down. Then I just need to find one of the older Genelecs with similar specs to measure and compare it to.

As I tried to clarify in my original post, I'm not certain there is a major performance disadvantage, so I can only go on what I read online and in the published specs.

I also do recall subjectively preferring the sound of the older non-coaxial Genelec monitors, but of course trying to draw anything from sighted comparisons in different listening rooms spaced years apart is a mug's game.

Remember, Genelec make both point source and conventional stacked designs. So they have an interest in both camps, so to speak.

I know, and I hope I'm making it clear that I'm in both camps too! I.e. I think there are some situations where a coaxial's advantages might outweigh those of a non-coaxial, and vice-versa. I'd basically like more data on the Genelecs to see just how good they've managed to get it...

Their flagship 1236 is heavily (pun intended) in the stacked camp with 1 tweeter, 2 mids and 2 woofers (horizontal).

Yeh, I've always looked at that speaker and wondered what on earth they were thinking

If you ever heard the sound of a conventional monitor recorded by a microphone vs a point source recorded by the same microphone - you would see this point clearer.

Well this was one of my criticisms of the original claims in the paper, really. Although I've never listened to this difference in isolation, I've certainly measured it and used sims to model it, so I'm aware of it. I just think the idea of recording it in such a way and listening to the recording is of limited value, given this is not what happens when one listens to a speaker at the correct angle in a real room.

EDIT: well, here are the measurements, and I most admit, any relative deficiencies of the 8351 seem to be down to the woofers and not the coaxial drivers (see distortion and CSD plots).

We also see a clear advantage to the 8351 in vertical polar response, as expected.

https://www.soundandrecording.de/equipment/studiomonitore-im-test-genelec-8350a/

https://www.soundandrecording.de/equipment/studiomonitor-genelec-8351a-im-test/

In other words, I’m pretty satisfied that Genelec’s waveguide mounted coaxial driver seems to work as least as well in terms of linear and nonlinear distortion as the arrangement it superseded

 

[oivavoi]

Do you know what mastering engineers might actually do to the recording in the case of listening to it in a reverberant vs. dry environment? It's not obvious to me how their mastering decisions would change - except for the frequency response discrepancies of real speakers due to their dispersion characteristics. They presumably might change the EQ in an attempt to partially correct the discrepancy. But what would they choose as their 'typical' speaker..?

I read one study which manipulated this condition. As far as I can remember, the mastering or mixing engineers in the more reverberant condition chose to finish their mix in a more dry condition, whereas the engineers who mixed or mastered in the more reverberant condition chose to add more reverb. That makes sense to me!