Does tweeter dispersion affect it's SPL compared to a woofer at different distances?

[fjhuerta]

Hope my question was clear enough.

Here's what happened. I'm taking a second look at a project I finished some time ago (The Monoliths). They are a 2.5 speaker, using the Peerless 830884 (apparently discontinued) and the SEAS 27TFC.

Specifically, I was taking farfield measurements at a greater distance (1.8 meters vs. 1.0 meters) trying to find out whether driver integration farther away could help me make the design better.

Much to my surprise, the tweeter level is quite a bit lower at that distance than at 1.0 meters - relative to the woofers' SPL. That is, at 1.8 meters, the tweeter has less relative (to the woofers) output. The speakers' FR is shelved downwards at that point.

Measured at 1 meter, the speaker measures flat on axis.

Could this really be happening, or should I measure a bit more?

[alias2]

If a bass/mid becomes directional towards the top end of its passband one
would then expect the inverse square law not to apply, i.e. you get beaming.

The power response, or set of responses over a range of axis, is the most
important, not the single 1m axial response, thought ideally they are all
relatively flat, with the wider angles having the inevetable losses but
no off axis flaring. Sterephile tests are worth perusuing, see RAR's.

Not sure this is enough to explain your results though.....

/alias2

[fjhuerta]

I'd hope I wasn't using the Peerless too high for an 8" woofer (I'm crossing over at 1.4 KHz). I could be wrong - it could be the woofer beaming, though.

Also, the Peerless has a rising response up to around 3 KHz.

Maybe a sharper order filter would help this issue?

[fbov]

Javier,
As I understand it (and this is a very recent understanding), the near/far field transition distance is a function of frequency, and the 1m distance is a calibration distance, not a fundamental requirement.

That is, to do far field measurements, you need to meet the geometric requirement of "far field" as defined by the drivers and frequencies you're measuring. If you're measuring a distance-sensitive metric like sensitivity, you then scale the result to 1m distance, regardless your actual mic location.

Mind you, this is purely theorietical as I'm just starting to learn about frequency response measurement, but it makes sense.

Have fun,
Frank

[HareBrained]

"You can not change the laws of physics."

As far as I know, SPL decay is not frequency dependent. As long as the measurements at 1m & 1.8m are on the exact same axis, it wouldn't matter if a driver was beaming. But I doubt you were far enough off axis to have resulted in the tweeter significantly decaying.

You could calculate a SPL for 1.8m based on the 1m measurements and see which is off. Is the woofers too loud, or the tweeter too soft.

The other consideration is integration of the woofers at 1m. If the mic was positioned on the tweeter's vertical axis, and the lower woofer is significantly lower, it possible that at 1m, you were too far off axis for the lower woofer to contribute to the output and at 1.8m, it does. Based on a 200mm driver, the acoustic center would be at least 350mm off axis, resulting in a 20deg offset. Could be significant.

[fbov]

Actually, SPL decay is frequency dependent when you're straddling the transition region from near to far field. Near field decays at 3dB/double distance while far field decays at 6dB/double distance (inverse square law). Since the distance at which you integrate to far field is frequency dependent, so is SPL until you're fully far field.

You can't change Physics, but you have to interpret the laws properly to get the right answer. I just don't recall the transition region distance rules... And you also have a plausible cause; more for Javier to investigate!

[HareBrained]

Actually, SPL decay is frequency dependent when you're straddling the transition region from near to far field. Near field decays at 3dB/double distance while far field decays at 6dB/double distance (inverse square law). Since the distance at which you integrate to far field is frequency dependent, so is SPL until you're fully far field.

You can't change Physics, but you have to interpret the laws properly to get the right answer. I just don't recall the transition region distance rules... And you also have a plausible cause; more for Javier to investigate!

Thinking about what you say makes sense if you consider the reinforcement provide near-field by the baffle. Once you reach far-field, by definition, you are in 4pi space and the energy is now dissipated through twice the volume.

Thanks for correcting me.

[fjhuerta]

Thinking about what you say makes sense if you consider the reinforcement provide near-field by the baffle. Once you reach far-field, by definition, you are in 4pi space and the energy is now dissipated through twice the volume.

Yes. I began wondering whether this was happening.... that's why I thought of dispersion as the main issue.

I'll probably double check my measurements and post them here later during the week.

[Dennis H]

I think the near-field/far-field thing is separate from the 2pi/4pi thing. One is source size vs. distance and the other is front radiation vs. rear radiation.

Javier, I'd look at phase effects due to differing path lengths. See if your LspCAD models match what you are measuring at the different distances (assuming you have the driver diameters and XYZ locations specified in LspCAD.)

[alias2]

"You can not change the laws of physics."
As far as I know, SPL decay is not frequency dependent.

Your quote is true but then interpreted.

For example consider a 10cm x 1cm line source. At low frequencies it will
be omnidirectional, at high frequencies a line source. One can consider
the high frequencies as beaming. The hf cylindrical radiation falls off
at 3dB as opposed to 6dB for the lf omni/hemispherical radiation.
Depends what you call beaming, though to me its fairly explanatory.

And the raison d'etre of long throw (as opposed to short) PA systems.

/alias2

1.4KHz c/o point is too low for any significant mid/bass driver beaming.

[fbov]

Last night, I found the formula in Toole and ran some numbers. He states it's 3-10x the size of the driver, but gives a formula that makes sense when you tie driver size to its frequency range:
distance^2 = wavelength^2/36

At a 1 meter distance, wavelength is 6m, and each meter farther gets you another 6m of wavelength. Since 6m is 56.7Hz (and at 2m distance, wavelength is 12m or 28.4Hz); it's unlikely that's what Javier's seeing.

His driver is 200mm, so the size multiplier gives 0.6m to 2m as the transition zone, clearly with respect to the low frequency capabilities of an 8" driver at the 2m end of the range.

[fjhuerta]

Your quote is true but then interpreted.

For example consider a 10cm x 1cm line source. At low frequencies it will
be omnidirectional, at high frequencies a line source. One can consider
the high frequencies as beaming. The hf cylindrical radiation falls off
at 3dB as opposed to 6dB for the lf omni/hemispherical radiation.
Depends what you call beaming, though to me its fairly explanatory.

I understand the concept.. but then.. if you have a speaker with, say, the Neo8 and Neo3 drivers, wouldn't they be a line source at their passband, and as such, would be impossible to properly mate with a cone driver woofer, since their radiation falls at different rates?

[ThomasW]

would be impossible to properly mate with a cone driver woofer, since their radiation falls at different rates?

Not impossible, integrating a line and point source means they must be set for a specific distance to the listener.

[fjhuerta]

Not impossible, integrating a line and point source means they must be set for a specific distance to the listener.

8O

Oh, my.

Now I know...

I got this very nice app for the iPhone - an RTA meter. I thought it was a toy, until I found how closely it tracked SynRTA with my PC setup.

Today, using it, I found out that, on my Miniliths (Seas P21 / Neo8 / Neo3), mids and highs were a bit higher in level than the bass at my listening position... but that at less distance they were flatter and flatter.

This, of course, means the speakers are only accurate at a specific distance from them, and have a narrower sweet spot than conventional drivers. Yikes. Not good. Maybe they'd work for HT.

Back to the drawing board! ops:

[alias2]

8O

This, of course, means the speakers are only accurate at a specific distance
from them, and have a narrower sweet spot than conventional drivers. Yikes.
Not good. Maybe they'd work for HT.

Back to the drawing board! ops:

I'd say it means set up for a typical listening distance it they would have an
atypical direct / reflected room sound, that does not mean not accurate.

For some this is the strength of line arrays - for others their downfall.

FWIW there is a thing called the critical distance, ideally this should not vary
too much with frequency and its listener preference for direct / room ratio.

See the JBL SSDRM Section 5 page 14 : http://www.pispeakers.com/ssdm_99.pdf

/alias2