Listening Window Versus Toe In

[fineMen]

When using the Klippel machinery to evaluate speaker performance, a reference axis has to be chosen. It is always the normal to the front panel, originating in the tweeter's center or alternatively in between tweeter and midrange.
How does this align with the quite regular set-up of not toeing in the speakers? The listening axis lies then, in the stereo triangle, some 30° off the reference axis. The calculated data for "listening window", directivity and all should no longer be applicable due to the shifted listening position.

Is it possible to re-calculate the data of the Klippel for an alternative reference axis?

Of course, the shift doesn't affect the grand total of sound, reverberation included, too much. But the relation of direct to indirect sound is altered. Regarding that I've got another question.

Is the following conclusion from Toole's book, correct?
The direct sound is, by human hearing, only used to give directional cues. The true colour of sound, the tone is evaluated from the grand total, reverberation field included.

That would implicate, that one better equalizes for a regularly tilted flat grand total. Small deviations in the direct sound would for the worst become detrimental to stereo imaging.

 

[lherrm]

Not sure I understand you, the NFS measures at many different angles around the speaker, so you have the data for these :

 

[Elk]

I think that 'the quite regular set up' is wrong, and that speakers typically should be toed in, so that one is listening on axis. I can only understand their being toed out to tame an unnatural high frequency rise.

 

[sergeauckland]

I think that 'the quite regular set up' is wrong, and that speakers typically should be toed in, so that one is listening on axis. I can only understand their being toed out to tame an unnatural high frequency rise.

Quite so! I've noticed that quite a few modern loudspeakers are deliberately balanced to be bright on-axis. How much of this is so they stand out at Dealer demos, I don't know, but one other (very doubtful) benefit of being bright on-axis is the perceived better WAF from not being having to be toed-in.

In the past I was a fan of Hugh Brittain of GEC's Hurst Research Centre who proposed that loudspeakers should be toed-in so their axes cross well in front of the listener, using a combination of intensity and arrival time to widen the stereo seat. Stick THAT in your WAF pipe and smoke it!

S.

 

[No. 5]

When using the Klippel machinery to evaluate speaker performance, a reference axis has to be chosen. It is always the normal to the front panel, originating in the tweeter's center or alternatively in between tweeter and midrange.
How does this align with the quite regular set-up of not toeing in the speakers? The listening axis lies then, in the stereo triangle, some 30° off the reference axis. The calculated data for "listening window", directivity and all should no longer be applicable due to the shifted listening position.

Correct.

You can make an educated guess of how far off everything is by looking at the separation between the listening window and on-axis curves, as well as the slope and shape of the DI curve. Typically, the "correct" listening axis is whatever the manufacturer specifies and that's what should be used for the measurements, since that's the sound the designer intended.

I'd like to make a minor clarification in that a reference axis must be chosen no matter what measurement system is being used, not just Klippel.

Is it possible to re-calculate the data of the Klippel for an alternative reference axis?

Not having a Near Field Scanner myself, I can't really answer that. But if the software can't do it, if it can export the relevant data, it can be done in other software.

Is the following conclusion from Toole's book, correct?
The direct sound is, by human hearing, only used to give directional cues. The true colour of sound, the tone is evaluated from the grand total, reverberation field included.

That would implicate, that one better equalizes for a regularly tilted flat grand total. Small deviations in the direct sound would for the worst become detrimental to stereo imaging.

Do you have a reference to a page number on that? I don't believe that's a correct conclusion about the direct sound, it provides a lot more than just directional cues.

 

[fineMen]

I'd like to make a minor clarification in that a reference axis must be chosen no matter what measurement system is being used, not just Klippel.
...
Do you have a reference to a page number on that? I don't believe that's a correct conclusion about the direct sound, it provides a lot more than just directional cues.

Thanks a lot for asking back!

(1) sure, that's well understood; only that the Klippel, or better to say the "new standard" asks for additional data that is heavily related to the reference axis
(2) alas, I can't give a page number; what of a "more" do You think of other than its contribution to the total sound field?

 

[Plcamp]

As one example, I found that an unequalized Tangband W8-1808 fullrange (which has a hf rise on axis) is best listened to about 8-10 degrees off axis, because the response is flatter there, and the change in response as you shift left/right at listening position is minimized.

 

[fineMen]

As one example, I found that an unequalized Tangband W8-1808 fullrange (which has a hf rise on axis) is best listened to about 8-10 degrees off axis, because the response is flatter there, and the change in response as you shift left/right at listening position is minimized.

What exactly makes You feel that the rise on reference axis could be equalized by selectivly toeing in? Do You have any measurement that would document the change in the grand total sound field, reverberation included--at Your ears?

More or less my current situation resembles Yours. I have to decide if I add a tweeter to an otherwise grand wideband driver. The complete speaker will be situated in a real bookshelf, carrying 400kg of human wisdom ... no toe in available. The grand total at ear position differs only little with redirecting the reference axis, but the direct sound does +/- 2dB.

Reiterated: for the tone, timbre the total sound, direct and reverberant combined count, only for the stereo impression the direct sound is evaluated individually. Correct?

Does the selection of a particular reference axis affect the "preference estimate"?

 

[Plcamp]

What exactly makes You feel that the rise on reference axis could be equalized by selectivly toeing in?

I actually had it toed more down and just slightly in, and measurements at LP showed what I mentioned. Preferred that to equalization on axis.

 

[ahofer]

Quite so! I've noticed that quite a few modern loudspeakers are deliberately balanced to be bright on-axis. How much of this is so they stand out at Dealer demos, I don't know, but one other (very doubtful) benefit of being bright on-axis is the perceived better WAF from not being having to be toed-in.

In the past I was a fan of Hugh Brittain of GEC's Hurst Research Centre who proposed that loudspeakers should be toed-in so their axes cross well in front of the listener, using a combination of intensity and arrival time to widen the stereo seat. Stick THAT in your WAF pipe and smoke it!

S.

In many shops the speakers are not toed in, apart from the high end listening room.

 

[aarons915]

I'm sure it can be done on a Klippel but as another poster said just look at the on-axis vs listening window to decide toe in. If on-axis looks better you can toe them in, if the listening window looks better it's generally better to have no or minimal toe-in.

 

[No. 5]

Thanks a lot for asking back!

Certainly!

(1) sure, that's well understood; only that the Klippel, or better to say the "new standard" asks for additional data that is heavily related to the reference axis

Yes, CEA-2034 specifies that measurements start from the "reference axis", and that doesn't account for someone not listening on that axis, or a technician/reviewer using something other than what was intended by the designer. In my opinion, that's why it's helpful to be able to do the measurements yourself if a person is technically inclined.

(2) alas, I can't give a page number; what of a "more" do You think of other than its contribution to the total sound field?

My response was to the wording of the original question, so if I misunderstood your intent, consider this as information for someone else happening across this thread (emphasis mine):

The direct sound is, by human hearing, only used to give directional cues. The true colour of sound, the tone is evaluated from the grand total, reverberation field included.

Well I agree that the true color of the sound comes from the grand total of what reaches the ears, the direct sound is the first thing your brain takes notice of, providing much of the information for perception of timbre, tone, and quality of the sound field (in addition to directional cues), and research shows that too much reverberant energy directly correlates with a reduction in sound quality. In the third edition of Toole's book on page 118, he remarks that:

the direct sound has a high priority in perception, establishing a reference to which latter arrivals are compared in determining such important perceptions as... timbre.

Think about what happens when you are conversing with someone in a noisy environment: the qualities of their voice are still clearly apparent even though any reverberation created by their voice is swamped by noise and completely inaudible. Additionally, if you are listening in an environment with a high enough direct to reflect ratio, block the direct sound and note the changes that happen to the timbre, tone, and quality of what you are listening to.

So yes, direct sound has a contribution to the total sound field, but as a major contributor to the tone, not just localization.

 

[kceenav]

Typically, the "correct" listening axis is whatever the manufacturer specifies and that's what should be used for the measurements, since that's the sound the designer intended.

I agree with that. But what if the manufacturer does not specify any listening axis?

Then it makes sense - in my opinion - to look for the angle (within, say, 0° - 30° off the orthogonal on the baffle) at which the direct sound is "reasonably flat" and as similar to the sound power as possible. (Not sure whether this is really the best way to look at it ... Anyway, I'll try to give an example of what I am talking about.)

Let's take a 2-way with the typical 6.5" low-mid driver and a direct radiating tweeter of a typical diameter. The woofer's sound will become more and more directional with rising frequency; at a typical crossover frequency (2 kHz - 3 kHz) it generally is significantly quieter off-axis than the tweeter is at the same frequency (assuming they both are about equally loud at the point of crossover). Because of this it may be a smart move by the designer to let run the woofer a little "hotter" at these midrange frequencies than what is needed for a maximally flat response "on axis" (what usually is considered as on-axis == the orthogonal on the baffle).

Thus the resulting "on axis" FR will NOT be flat - but the sound power will be better behaved (more constant / more steadily declining with frequency). And in order to not hear too much of the "too loud" mid-frequencies (too loud they are only "on axis" though!) it obviously is a good solution to simply listen at an angle of 20° to 30° "off-axis". Which now is the designated listening axis.

By the way, to me it looks like the above approach can regularly be observed in multi-way speakers.

 

[No. 5]

I agree with that. But what if the manufacturer does not specify any listening axis?

Then it makes sense - in my opinion - to look for the angle (within, say, 0° - 30° off the orthogonal on the baffle) at which the direct sound is "reasonably flat" and as similar to the sound power as possible.

That sounds like a perfectly reasonable solution to me, and what I would do too. I would also add looking at a vertical set of angles; I once had a speaker in my collection who's woofer and tweeter only summed properly at around twenty degrees below "on axis".