Nearfield Bass Shootout

[Zilch]

http://www.audiokarma.org/forums/showthrea...345#post2972345

All with mic ~1/4" in front of in-place grilles:

[genek]

All with mic ~1/4" in front of in-place grilles:

Do you still have that pair of AR-3a's you were measuring a while back to throw into this?

[Zilch]

Do you still have that pair of AR-3a's you were measuring a while back to throw into this?

Well, yah, but how 'bout some guesses, first....

[Steve F]

http://www.audiokarma.org/forums/showthrea...345#post2972345

All with mic ~1/4" in front of in-place grilles:

Looks like more of a rise in the bass FR just before rolloff than I would have suspected from AS theory and manufacturer's published curves.

Are these high-Q systems? (>1.0?)

Are these just the normal variations we might expect in comparing lit/test report FR curves to 'homegrown' FR curves made under different conditions?

Perhaps it is the scale of the graph that makes the visual seem unusual?

Not criticizing or implying anything; just asking about the apparent variations.

Steve F.

[Zilch]

Are these just the normal variations we might expect in comparing lit/test report FR curves to 'homegrown' FR curves made under different conditions?

Nearfield measurements reflect 2-Pi performance, i.e., with one boundary of bass reinforcement, which is typical of how vintage East Coast speakers were intended to be used.

Published frequency response curves are often anechoic, i.e., 4-Pi freespace, no boundaries. To the best of my knowledge, however, there are and never have been any anechoic chambers large enough to measure accurately down to, say, 20 Hz. Thus, the published results may be "corrected" as required for publication. A speaker buried in the ground is 2-Pi.

I am measuring at the grille, not the woofer dome, the reason being that the particular pair of KLH-6 which launched the compariion are the sealed version, with the grilles not removable. The several that I did check both ways did not vary significantly, though a valid argument may be made that, as consequence, these measurements are progressively less and less accurate above 200 Hz.

Another issue is how to normalize the measurements. You can see that I have tried to make them all equal in the 300 - 500 Hz range, and noted the SPL level adjustment for each relative to the subject KLH-6s, and these are reasonable given the known sensitivities of the various models.

Still, I believe there are valid conclusions to be drawn here, which, for now, I leave to readers and students of "classic" loudspeaker design....

[genek]

Well, yah, but how 'bout some guesses, first....

If these weren't so consistent I might guess leaking surrounds, except that AR-4's, big KLH's and Advents always sounded to me as if there was too much bass even when there were no surround leaks.

I never heard the CSWs, but if Kloss was trying for a "family resemblance" to his earlier works at AR, KLH and Advent, I guess it means he really liked bass a lot. Certainly a lot more than I do.

[Zilch]

I mean guess where the AR3a and Smaller Advent are gonna fall measured under these same conditions....

[genek]

I mean guess where the AR3a and Smaller Advent are gonna fall measured under these same conditions....

Oh. Well, based on AR's old published curves of the 3a woofer, I'd expect to see the same rise, with a lower amplitude and extending about 5Hz lower. No idea if or how that translates to the actual sound in my living room.

I've only heard small Advents a few times and mostly in dealer showrooms and don't recall anything from test reports of the time, so I have no guess on them.

[Carlspeak]

Looks like more of a rise in the bass FR just before rolloff than I would have suspected from AS theory and manufacturer's published curves.

Are these high-Q systems? (>1.0?)

Are these just the normal variations we might expect in comparing lit/test report FR curves to 'homegrown' FR curves made under different conditions?

Perhaps it is the scale of the graph that makes the visual seem unusual?

Not criticizing or implying anything; just asking about the apparent variations.

Steve F.

I have gotten results similar to Z's which have been published both here and at AK. The conclusion I have drawn is Q's in the 1 to 1.5 range were probably the 'modus operandi' of many classic AS speaker producers. After all, clean, rich bass from a small cabinet was one of the main selling points of these speakers in the early days of AS speaker development.

However, if one backs away from NF type measurements, that hump flattens out.

[Zilch]

I'm out of colors and also space in the legend. Original Large Advent is now Violet, The Smaller Advent #853 +2.5 dB from KLH-6 = Red, and AR3a #368 also +2.5 dB = Blue:

[genek]

I'm out of colors and also space in the legend. Original Large Advent is now Violet, The Smaller Advent #853 +2.5 dB from KLH-6 = Red, and AR3a #368 also +2.5 dB = Blue:

Interesting. So is Carl right about the hump smoothing as you move away, how much does it smooth and how far back do you have to go?

I've never listened to any of my speakers with my ear 1/4" from the grills. I'll have to remember to continue that habit.

[Zilch]

Interesting. So is Carl right about the hump smoothing as you move away, how much does it smooth and how far back do you have to go?

If they are not built into a wall or casework, as you move back, baffle step comes into play, but if they're against the wall, it comes back, because the space is once again a hemisphere.

In a corner, or near the floor or ceiling, even more boundary reinforcement occurs. Move them out from boundaries, and the response flattens out, enough so, in some circumstances, that augmentation may be required, hence, PeteB's "BSC" circuit for use with Advents (and others, presumably). Generally, vintage acoustic suspension speakers were intended for use with boundary reinforcement; it says so right on the backs of many of them.

Baffle step is ~6 dB max. Note that the vertical scale here is 5 dB/major division, and also that the peak is not all that is in evidence here; Advents and AR3a are reaching substantially lower frequencies than the others, which are also peaking at higher frequencies. Call it wherever you like, but at -6 dB, I'm seeing them reaching 30 Hz, and KLH-6, 40 Hz, while the others aren't getting down to 50 Hz, even....

[kkantor]

Well, "shoot out" implies winners and losers.

From a designer's perspective, or at least, this designer's perspective, it is difficult to interpret the data without knowing what the pass band sensitivity and directivity looks like, and what the crossover is doing. Any thoughts?

First question: are the crossovers operational during your measurements? (If so, can you repeat, but with a direct voltage drive to the terminals?)

-k

http://www.audiokarma.org/forums/showthrea...345#post2972345

All with mic ~1/4" in front of in-place grilles:

[Zilch]

Well, "shoot out" implies winners and losers.

From a designer's perspective, or at least, this designer's perspective, it is difficult to interpret the data without knowing what the pass band sensitivity and directivity looks like, and what the crossover is doing. Any thoughts?

First question: are the crossovers operational during your measurements? (If so, can you repeat, but with a direct voltage drive to the terminals?)

Hi Ken!

These are all done with the crossovers operating, the stock speakers measured with the grilles in place. Some of the designs have no lowpass filters. Those that required it have new surrounds.

Someone on another forum asked if the curves tell us how the speakers sound, and the answer is "No," as they don't describe how the bass response is blended with the rest of the particular systems.

I'm just starting on KLH-6, but the others are more comprehensively analyzed in individual threads over on AudioKarma. In some instances, I went on to "Zilchify" them, and the drivers themselves are measured separately. I do have intact samples of all of these, and additional spares, if you want to study any of them....

Edit: An AKer wanted to see how Karma Indignia compared so I added that, as well, with the port closed, black, and also bumped AR3a up 0.25 dB to separate it from Large Advent:

[soundminded]

If they are not built into a wall or casework, as you move back, baffle step comes into play, but if they're against the wall, it comes back, because the space is once again a hemisphere.

In a corner, or near the floor or ceiling, even more boundary reinforcement occurs. Move them out from boundaries, and the response flattens out, enough so, in some circumstances, that augmentation may be required, hence, PeteB's "BSC" circuit for use with Advents (and others, presumably). Generally, vintage acoustic suspension speakers were intended for use with boundary reinforcement; it says so right on the backs of many of them.

Baffle step is ~6 dB max. Note that the vertical scale here is 5 dB/major division, and also that the peak is not all that is in evidence here; Advents and AR3a are reaching substantially lower frequencies than the others, which are also peaking at higher frequencies. Call it wherever you like, but at -6 dB, I'm seeing them reaching 30 Hz, and KLH-6, 40 Hz, while the others aren't getting down to 50 Hz, even....

Except for KLH model 17, all of the speakers are underdamped according to the graphs having a system Q above critical damping of .707. This is hardly surprising for some of them and may explain why KLH Model 6 had a tendency for chesty male voices especially when the effect was enhanced by proximity effect by close miking of radio announcers on FM broadcasts. But we know that AR3a was designed for critical damping as were all AR speakers of the day and that quality control was fairly tight even before the existance of ISO 9000. Therefore it is reasonable to assume that some process is at work in "detuning" the systems over time, perhaps a change in the distribution of the damping material within the enclosure or air leaks due to porousity of the surrounds or other factors. At the time they were marketed, equalization was not available to the general public due to cost. But when these devices became consumer items instead of exclusively professional studio equipment things changed. The FR anomolies are correctable within the power handling capabilities of the drivers. Interesting how all of them follow the theoretical 12 db per octave falloff almost exactly regardless of their system Q or Fs. This suggests that in sufficient numbers and with sufficient power, any of them could reach any frequency desired and at any loudness level. (This is why the acoustic suspension version of Bose 901, the original and series II could reach 23 hz with 10% THD even though the system Fs with the equalizer in my room at least, is around 250 hz and 5 db above the 1 khz output level. However, the 6 db per octave boost from the Bose 901 equalizer is not sufficient for flat response. Additional boost makes power requirements astronomical suggesting multiple systems are needed even in smaller rooms.)

At least one other factor besides speaker placement must be taken into consideration and that is the low frequency cutoff of the rooms they are installed in. Unlike live venues where the principal room dimensions are greater than the wavelength of the lowest frequencies of musical interest, rooms in most homes are smaller and therefore have a tendency to introduce their own low frequency rolloff. This must also be compensated for. Also in such room, the bass response at any location in the room will be different for the same reasons as speaker placement. Bass level can vary by an enormous amount even moving a short distance. Some say by as much as 40 db. Also, variations in program material bass levels have in my experience required further equalization as this is also a critical variable that must be dealt with. Sound systems that can reach these depths have the potential for far more enjoyable listening...and new and unexpected problems if not corrected.

[genek]

Someone on another forum asked if the curves tell us how the speakers sound, and the answer is "No," as they don't describe how the bass response is blended with the rest of the particular systems.

So what is it that this measurement does tell us...?

[Carlspeak]

So what is it that this measurement does tell us...?

It tells us how the speaker's woofer would test in an anechoic chamber (without haveing to use one). D.B. Keele, JR. showed in his 1973 AES paper that for low frequencies, the nearfield sound pressure is directly proportional to the farfield sound pressure by the ratio of the mic meas. distance to the woofer's DC divided by the woofer's radius. He concluded that valid NF measurements may be taken in any reasonable environment without the use of an anechoic chamber or large outdoor test site. I'd post a scan of Fig. 11a & 11b from Keele's paper showing comparison's of NF and FF measurements on the same speaker system, but I might violate AES's copyright. The measurments show the drop in woofer's response hump at 36 inches vs NF.

I want to see a correlation scatterplot of Z's system Q measurements with the dB hump levels of each speaker system. It should be pretty tight if theory holds.

[genek]

It tells us how the speaker's woofer would test in an anechoic chamber (without haveing to use one). D.B. Keele, JR. showed in his 1973 AES paper that for low frequencies, the nearfield sound pressure is directly proportional to the farfield sound pressure by the ratio of the mic meas. distance to the woofer's DC divided by the woofer's radius. He concluded that valid NF measurements may be taken in any reasonable environment without the use of an anechoic chamber or large outdoor test site. I'd post a scan of Fig. 11a & 11b from Keele's paper showing comparison's of NF and FF measurements on the same speaker system, but I might violate AES's copyright. The measurments show the drop in woofer's response hump at 36 inches vs NF.

I want to see a correlation scatterplot of Z's system Q measurements with the dB hump levels of each speaker system. It should be pretty tight if theory holds.

How does this method compare to Villchur's old scheme of burying the cabinet face up in the ground outdoors? From the photo in his article about measurements, it appears that the mic was mounted on a curved rail about 3 ft from the woofer (I don't see anything in the article about whether the crossover is in or out), and no grill. The curve of a 3a woofer from that period also shows a bass rise, but the image defies my failing vision; is this the same thing we're seeing here?

[Zilch]

So what is it that this measurement does tell us...?

That Kloss succeeded in packing AR3a bass into the Large Advent, and very nearly also in the Smaller Advent.

I want to see a correlation scatterplot of Z's system Q measurements with the dB hump levels of each speaker system. It should be pretty tight if theory holds.

Thus far, there is a clear correlation between efficiency and LF extension among the vintage drivers.

How does this method compare to Villchur's old scheme of burying the cabinet face up in the ground outdoors? From the photo in his article about measurements, it appears that the mic was mounted on a curved rail about 3 ft from the woofer (I don't see anything in the article about whether the crossover is in or out), and no grill. The curve of a 3a woofer from that period also shows a bass rise, but the image defies my failing vision; is this the same thing we're seeing here?

"Hole in the ground" is 2-Pi, as is nearfield, radiating into a hemisphere.

By definition, anechoic is 4-Pi, though AR and others recessed the speakers in one wall or used large baffles, rendering it 2-Pi, and may have applied correction factors to convert LF response into a 4-Pi representation. You must dig deeper into the methodology used to understand what is shown.

Perhaps Ken or Dave will clarify this issue. In the meantime, I'm going to look in the Villchur measurements paper for more detail regarding the curve you post:

http://www.classicspeakerpages.net/library...esting_and_mea/

Edit: Villchur says it's 2-Pi, and shows a 1 dB peak at 60 Hz. I show a 5 dB peak at 80 Hz. We have a discrepancy....

Re-Edit: Speaker Dave shows more like 10 dB in AR2ax:

http://www.classicspeakerpages.net/IP.Boar...ost&p=78390

[Carlspeak]

How does this method compare to Villchur's old scheme of burying the cabinet face up in the ground outdoors? From the photo in his article about measurements, it appears that the mic was mounted on a curved rail about 3 ft from the woofer (I don't see anything in the article about whether the crossover is in or out), and no grill. The curve of a 3a woofer from that period also shows a bass rise, but the image defies my failing vision; is this the same thing we're seeing here?

Villchur's method, I suspect, was done for lack of a true anechoic chamber. However, just as effective. He was measuring in true 2 pi conditions with the speaker's face flush with ground level. The curved rail set up was to get the off axis measurements. Attached is one of my close mic'd AR3a woofer measurements in cabinet shown as the red line. Disregard the blue line. This is pretty typical of what I've measured in 3a's. My results were not as flat as Villchur's out to 500 hz but similar. Also my -3 dB point is about 55 hz. His curve is pretty flat down to 50. I didn't measure the system Q on this speaker unfortunately.

I should add that the 3a's I have done system Q's on run in the .6 to 1 range and thus don't exhibit the severe humps like those shown in Z's measurements of other speakers. These Q's are the result of my re-stuffing so don't represent 'stock' performance.

[genek]

In the meantime, I'm going to look in the Villchur measurements paper for more detail regarding the curve you post:

FYI, the curve I posted is not the one from the article, though it does have a similar shape. The one I posted is from another AR brochure that dates to about the same period that I have in a bigger and easier to read scan. I don't think this is in the library, so let me know if you want me to post it. The curve I put up is the only one in it with 3a woofer info.

[Zilch]

FYI, the curve I posted is not the one from the article, though it does have a similar shape. The one I posted is from another AR brochure that dates to about the same period that I have in a bigger and easier to read scan. I don't think this is in the library, so let me know if you want me to post it. The curve I put up is the only one in it with 3a woofer info.

It shows about the same, +1.5 dB at 50 - 60 Hz.

Edit: Here's my measurement at the dome in March:

[genek]

Edit: Villchur says it's 2-Pi, and shows a 1 dB peak at 60 Hz. I show a 5 dB peak at 80 Hz. We have a discrepancy....

Only one thing to do now, dig a hole in your back yard and remeasure duplicating Vilchur's method to see if the discrepancy is between test subjects or methodologies...

[Pete B]

Edit: Villchur says it's 2-Pi, and shows a 1 dB peak at 60 Hz. I show a 5 dB peak at 80 Hz. We have a discrepancy....

I would suggest that you check your gate time, and windowing, even for swept sine type

measurements. I have commented about this rise in the bass issue in the past.

Keele gets good agreement with theory, as can be seen in his paper:

http://www.xlrtechs.com/dbkeele.com/PDF/Ke...eld%20Paper.pdf

I get excellent agreement here in Figs 13-15, note that he crossover is bypassed.

Also note that 100 to 200 Hz is roughly the passband, and the tilt down with

increasing frequency is probably due to Lvc as would be expected:

http://baselaudiolabs.googlepages.com/MR-TC50-REB.html

The amplitude response is in excellent agreement with the measured Q at the

measured in box resonance Fc.

Something to consider is that the baffle has little effect below about 150 or 200 Hz for

these large bookshelf speakers and below this frequency there is only a flat amplitude

difference between 2pi and 4pi. Thus nearfield also represents 4pi at very low freq

using a different correction factor to the far field.

[Pete B]

There is also a curve in Villchur's 1957 AES article for the AR Woofer, however

a damping factor of 1 is used which raises Qtc as compared to use with a modern

high damping factor amp. Remember a damping factor of 1 means that the

driving amp has an output impedance equal to the speaker load. The Qtc looks to

be about .9, and I believe that it is about .7 with low impedance drive. The curve

shows little if any peaking, perhaps .5 to 1dB at most even with a damping factor

of 1. There should be no peaking at all with a Qtc of .707 or less since by definition

.707 is a maximally flat response with a monotonic rolloff.