[FredT300B]

Looks like I'm just about the only one initiating threads, but I suppose that's OK to get this forum going.

I'm confused about the issue of measuring a line array's frequency response? If the standard method of measuring with the microphone placed one meter from the front is used, will interference caused by the varying distances of the drivers from the mike yield readings that don't reflect the speaker's true response at typcial listening distances? How do the experts measure line arrays?

[Rick Craig]

Looks like I'm just about the only one initiating threads, but I suppose that's OK to get this forum going.

I'm confused about the issue of measuring a line array's frequency response? If the standard method of measuring with the microphone placed one meter from the front is used, will interference caused by the varying distances of the drivers from the mike yield readings that don't reflect the speaker's true response at typcial listening distances? How do the experts measure line arrays?

Your measuring distance should be equal to or greater than the length of your line. If you have one line of drivers longer than the other (for instance 5 ft. of tweeters and 6 ft. of woofers) use the distance that's the greater of the two. Also measure the left and right horizontal angles so you can see the lobe and adjust your crossover topology as desired. Vertical location of the mic can also affect the response quite a bit.

If you plug in the raw curves to a crossover program the distance of the mic can have a big effect on the simulations (garbage in, garbage out). If you measure too close the response curve will tilt downward and your resulting crossover probably won't have enough baffle step compensation.

[Danny Richie]

I have to agree with Rick (believe it or not).

I also will take measurements at various distances from 8 or 9 feet to 16 feet away. I do this to determine what gains are room related as different distances will vary the low end response. I can then get an average to work with. Doing this will keep you from trying to correct a room related peak that is no fault of the speaker.

[Wayne Parham]

Only a hypothetical omnidirectional point source in an anechoic environment will radiate the same energy in all directions.

Real-world speakers have directional characteristics, most are non-uniform.  They have interference between sound sources, which, when further than 1/4 wavelength apart cause diffraction effects which make lobes appear.  They have astigmatism, a different apparent source location when movement in the listener occurs along one axis verses another.  They have collapsing directivity as a result of radiator diameter.  There are other things too;  Those are just a few right off the top of my head.

That's why off-axis and polar charts are helpful for describing directional characteristics, in addition to a single on-axis measurement made at a fixed distance.  It would also be helpful to have measurements taken at various distances in addition to angles, a sort of 3D plot of energy distribution.

You're right about the problems of measuring at 1 meter.  You can always measure at 10 meters and 100 watts.  That will tend to make the angles between the top and bottom of the speaker less different.  All angles are closer to zero, and this sort of works as a correction.  The SPL figure at 10 meters and 100 watts is also the same number as 1W/1M.  But it does require a bit more space, a bit more power and is probably a far field measurement, depending on loudspeaker characteristics.  Distortion measurements are obviously going to reflect the higher power input too, so it doesn't make sense on a loudspeaker designed to be used at lower power levels.

I like to do ground plane measurements.  The idea is to put all drivers within 1/4 wavelength of the radiating plane, and have no other reflectors present.  This means you don't have to gate out reflections, because there are none.  Pseudo-anechoic (gated) measurements are fine for indoors measurements above the midrange band, to check summing through a crossover band or something, but for best full-range accuracy, I like outdoors ground plane measurements best.  If all drivers are within 1/4 wavelength distance from the ground, then the speaker can simply sit there, facing the microphone, which is placed on the ground a certain distance away.  If it isn't possible for all drivers to be within 1/4 wavelength of the ground, a pit is dug for the speaker to rest in, facing up.  The microphone is then suspended above the speaker.

Ground plane measurements works well for line arrays, because you can lie them on their sides to keep all diaphragms within 1/4 wavelength of the ground.  Put the microphone on the ground also.  Measure at 100 watts and 10 meters and you'll have a very good indication of what the speaker is doing, without any outside interference from reflections.

That's how I measured your ART Arrays.  I put one of them on its side outdoors.  I only measured 3 meters away in this case though.  Since the speaker wasn't designed for high power use, I didn't want to overdrive it and possibly shift the electro-mechanical parameters of its drivers, changing the response curve as a result.  So I used less power, and to keep the signal-to-noise ratio high, I moved the speakers closer.  I did not get closer than line length though;  My goal was to balance competing criteria of low power, low signal-to-noise and measuring distance.

Still, that was just one way of measuring the speakers, and only gives one sample of response.  It will undoubtedly change at different distances and different angles.  If you want the best view of what the speaker is doing, I think you'll need a grid of listening points to really see what's going on.

[Danny Richie]

Actually laying an array on the ground, on its side, and the mic on the ground will really skew the response. It will make it considerably heavier in the ranges below 1000 to 900Hz.

Most of the wavelengths below 800 to 900Hz (depending on baffle width) will loose output as they become more omni directional. This is compensated for with the crossover. Laying it on the ground and on its side will give you half of that loss back.

It also gives you the surface reflections of the ground along with the direct output and could cause ripples all along the way.

[Daryl]

Speakers aren't necessarily measured at 1m and 2.83v.

Measurements usually are converted (referred to) to what you would have if the speaker were a perfect point source, 1m away and driven with 2.83v.

You might measure a fairly large speaker at 2.5m to be in it's far-field and 5.65v to increase your S/N ratio and then subtract 14db to refer it to 1m and 2.83v even though the actual measurement distance and voltage were different and if you did measure at 1m and 2.83v the results wouldn't coincide with your converted data because the drivers and baffle edge reflections wont sum the same at that close distance.

You must measure in a way that provides meaningfull data and then present it in a way that is comparable to data from other speakers and that is where the 1w/1m comes from.

With a very small speaker you might measure at 0.5m to dilute artifacts of your measurement environment and then subtract 6db.

For bass you likely will measure from less than an inch and refer the result to 1m and 2.83v and combine it with far-field measurements if it's not a subwoofer.

The response of a practical line array will be changing constantly with listening distance unless you plan on listening from so far away or the line is so small that you will be in it's far-field.

In theory the line would be so large vertically and the drivers so close together that you will do all your listening in it's near-field and distance would not matter.

This will not be achieved in practice.

You must have an intended listening distance in mind (actually an intended range) at the outset of your design process for line arrays because you probably will not be in either it's nearfield or farfield which means it will be in constant transition.

So no you don't measure line arrays from 1m neither do you measure from a suitably far distance, you measure within your intended range of listening distances.

No other approach will give usefull data.

For the crossover your room should be large enough you can window out room anomalies or you can smooth the response to remove anomalies as shown below.



The dotted line is a composite quasi anechoic response of two midbasses.

The solid line is an unwindowed/ungated inroom measurement of the same that has simply been smoothed to remove room artifacts.

The agree very well above 700hz which is fine for crossovers.

You see the smoothing has erased the anomaly around 4khz which makes no difference since it's too tight a wrinkle to address in a crossover anyway.

For bass you simply will be looking at the average bass level inroom and ignoring room specific anomalies.

You can see from the bass level in this inroom curve that 93db would be where you want to set your sensitivity for your crossover and that this satelite speaker will have an inroom extension of 80hz.

You might want to use several different rooms to get an average unless one room is all you care about.

Even though this is a small speaker and the listner will always be in it's far-field unlike your line arrays it is still important to do this measurement at a likely listening distance with both the speaker and the microphone in the appropriate locations within the room for speaker and listener because these factors effect inroom bass level and you need to gauge inroom bass level to know at what sensitivity you will align your crossover.

[Rick Craig]

Actually laying an array on the ground, on its side, and the mic on the ground will really skew the response. It will make it considerably heavier in the ranges below 1000 to 900Hz.

Most of the wavelengths below 800 to 900Hz (depending on baffle width) will loose output as they become more omni directional. This is compensated for with the crossover. Laying it on the ground and on its side will give you half of that loss back.

It also gives you the surface reflections of the ground along with the direct output and could cause ripples all along the way.

Wayne is correct on this one Danny. If you read Joe D' Appolito's book "Testing Loudspeakers" there's a section devoted to ground plane measurements that contradicts your comments. Mark Gander (well-known in the audio field as an acoustics and measurement expert) did the original AES paper. This method is a very accurate way to measure a line array.

[Daryl]

Actually laying an array on the ground, on its side, and the mic on the ground will really skew the response. It will make it considerably heavier in the ranges below 1000 to 900Hz.

Most of the wavelengths below 800 to 900Hz (depending on baffle width) will loose output as they become more omni directional. This is compensated for with the crossover. Laying it on the ground and on its side will give you half of that loss back.

It also gives you the surface reflections of the ground along with the direct output and could cause ripples all along the way.

Wayne is correct on this one Danny. If you read Joe D' Appolito's book "Testing Loudspeakers" there's a section devoted to ground plane measurements that contradicts your comments. Mark Gander (well-known in the audio field as an acoustics and measurement expert) did the original AES paper. This method is a very accurate way to measure a line array.

You won't get a half-space response devoid of diffraction loss as Danny was suggesting or any ripple due to multipath since the mic being on the ground in groundplane measurements causes the direct and reflected arrivals to coincide but you will effectively have a baffle that is wider which will alter baffle effects.

You should keep the enclosure vertical and lean it forward to get the mic on-axis and at the intended listening distance and of course subtract the 6db for a ground-plane measurement.

This would make the baffle taller but in line arrays the horizontal baffle effects dominate.

[Danny Richie]

Think it through again Rick. Come on, you can do it. 

If you take the baffle of a line array and make in infinitely wider on one side or lay it sideways on the ground then you are adding back half of the baffle step loss that it had when free standing.

Daryl, We may not agree on the sonic differences of wire and caps, but your pretty spot on with how to measure a speaker.

[Rick Craig]

Actually laying an array on the ground, on its side, and the mic on the ground will really skew the response. It will make it considerably heavier in the ranges below 1000 to 900Hz.

Most of the wavelengths below 800 to 900Hz (depending on baffle width) will loose output as they become more omni directional. This is compensated for with the crossover. Laying it on the ground and on its side will give you half of that loss back.

It also gives you the surface reflections of the ground along with the direct output and could cause ripples all along the way.

Wayne is correct on this one Danny. If you read Joe D' Appolito's book "Testing Loudspeakers" there's a section devoted to ground plane measurements that contradicts your comments. Mark Gander (well-known in the audio field as an acoustics and measurement expert) did the original AES paper. This method is a very accurate way to measure a line array.

You won't get a half-space response devoid of diffraction loss as Danny was suggesting but you will effectively have a baffle that is wider which will alter baffle effects.

You should keep the enclosure vertical and lean it forward to get the mic on-axis and at the intended listening distance and of course subtract the 6db for a ground-plane measurement.

This would make the baffle taller but in line arrays the horizontal baffle effects dominate.

I've not used this for measuring my arrays but I recently helped a friend measure an array outdoors with the ground plane method and it works great provided you measure as suggested by the experts. The main problem is finding a place that's quiet and where you won't disturb the neighbors 

[Rick Craig]

Think it through again Rick. Come on, you can do it. 

If you take the baffle of a line array and make in infinitely wider on one side or lay it sideways on the ground then you are adding back half of the baffle step loss that it had when free standing.

Daryl, We may not agree on the sonic differences of wire and caps, but your pretty spot on with how to measure a speaker.

I did think about it and stand by what I said. I'm in pretty good company as well 

You've got plenty of flat open space in Wichita Falls to try it. D'Appolito also shows graphs of measurements using different methods to compare against the ground plane and demonstrates that it can be very accurate. It works and I've seen it myself with a friend's array measurements. It's a no-brainer to me.

[Danny Richie]

I'm in pretty good company as well

Okay, let's get something straight here about who is in who's company. Joe D and I are part of the same global design team, and you are NOT.

Secondly, if you want to take a ground plane measurement of a woofer that's fine. There are valid reasons for the ground measurement. But when you lay a line source on its side you are effectively altering the width of its baffle in what is actually its horizontal plane and you will get an increase in output in the lower frequency ranges because of it.

But hey if you want to measure your arrays that way then knock yourself out.

[Rick Craig]

I'm in pretty good company as well

Okay, let's get something straight here about who is in who's company. Joe D and I are part of the same global design team, and you are NOT.

Secondly, if you want to take a ground plane measurement of a woofer that's fine. There are valid reasons for the ground measurement. But when you lay a line source on its side you are effectively altering the width of its baffle in what is actually its horizontal plane and you will get an increase in output in the lower frequency ranges because of it.

But hey if you want to measure your arrays that way then knock yourself out.

"In the same company" meant Joe D., Mark Gander, Don Keele, the late Dick Heyser, and all of the other people who agree that nearfield and ground plane meausrements work very well. Global design team? Last time I checked Joe only worked for Usher as a consultant and is employed by Snell on a full-time basis.

The ground plane and room curves of the arrays we measured were in agreement so that the step response 200hz-1K looked the same. If the ground plane was altering the apparent baffle width the step response would be different between the two. The ground plane measurement actually has better resolution in the lower octaves because there are no room boundary effects (similar to nearfield or a large anechoic chamber).

[Danny Richie]

...and all of the other people who agree that nearfield and ground plane meausrements work very well.

Hey I agree that they work very well too (for what they are intended for). Does that put me in the same company too?    If you want to lump me in there with those guys, fine, but you and I are NOT in the same "company".

Global design team?

Have you not done your homework? I am not bringing you up to speed. Figure it out on your own.

Good luck Rick.

[RAW]

...and all of the other people who agree that nearfield and ground plane meausrements work very well.

Hey I agree that they work very well too (for what they are intended for). Does that put me in the same company too?    If you want to lump me in there with those guys, fine, but you and I are NOT in the same "company".

Global design team?

Have you not done your homework? I am not bringing you up to speed. Figure it out on your own.

Good luck Rick.

http://audiojunkies.com/page/4

[FredT300B]

I started this thread with a question about the best distance for accurately measuring a line array. Apparently everybody agrees a line array should be measured from distances greater than 1 meter to obtain a result that's more representative of its response at the typical listening distance. There is some disagreement about the advantages and disadvantages of ground plane measurements. This is relevant to the topic, but I suggest we move it to the Lab since it's a technical issue that applies to all speakers and not just line arrays. There's also a third issue which is unrelated to speaker measurement. That issue is the undercurrent of hostility between Danny and Rick, which for me is a disraction from the technical merits of this discussion.

[Rick Craig]

...and all of the other people who agree that nearfield and ground plane meausrements work very well.

Hey I agree that they work very well too (for what they are intended for). Does that put me in the same company too?    If you want to lump me in there with those guys, fine, but you and I are NOT in the same "company".

Global design team?

Have you not done your homework? I am not bringing you up to speed. Figure it out on your own.

Good luck Rick.

http://audiojunkies.com/page/4

Thanks for the link Al. Congrats to Danny as well.

[Rick Craig]

I started this thread with a question about the best distance for accurately measuring a line array. Apparently everybody agrees a line array should be measured from distances greater than 1 meter to obtain a result that's more representative of its response at the typical listening distance. There is some disagreement about the advantages and disadvantages of ground plane measurements. This is relevant to the topic, but I suggest we move it to the Lab since it's a technical issue that applies to all speakers and not just line arrays. There's also a third issue which is unrelated to speaker measurement. That issue is the undercurrent of hostility between Danny and Rick, which for me is a distraction from the technical merits of this discussion.

Fred,

I've tried to keep this discussion focused on the facts and not make it personal. I'm not trying to be argumentative; in fact, my posts would change very little if it was someone other than Danny disputing what I've said. All I'm trying to do is give everyone the best and most reliable information there is on doing your own measurements. I would highly reccommend that anyone interested in learning more purchase Joe D'Appolito's book "Testing Loudspeakers" and the AES papers by Don Keele and Mark Gander.

[Wayne Parham]

Baffle step is actually directionality shift.  Directionality increases, so on-axis SPL does too.  It is frequency sensitive, in that directionality only increases at frequencies where the baffle is acoustically large.  The power response doesn't change, but the on-axis response does.  It occurs whether a speaker operates in freespace or half space.  The only time it doesn't happen, is when the radiating angle is narrower than the baffle, like in the case of horns.

When you perform an outdoor ground plane measurement, what you're doing is to measure the speaker in a pure anechoic half-space, with no reflectors.  The idea is to place the speakers within 1/4 wavelength of the ground, which prevents it from causing reflections that would interfere with the source.  It's a launch point boundary, simply constraining the radiating angle without causing anomalies from reflections.

If you lay a tall, thin array speaker on its side, then two important things are accounted for.  One is the distance between sound sources and the ground are small.  This prevents the ground from causing destructive interference, and makes it act as a pure half-space boundary.  The other is the baffle still defines the radiating pattern, so baffle step is still measured.  At frequencies high enough that the baffle is acoustically large, the radiation pattern is reduced to quarter-space.  At lower frequencies, the radiation reverts to half-space.

Whether the speaker is lying on its side or standing upright, the baffle halves the radiating angle at frequencies where it is acoustically large.  That's key here.  On its side, the baffle step transition is between halfspace and quarter space.  Upright, the baffle step transition is between freespace and halfspace.  But in each case, the baffle step transition halves the radiating angle.  This means you will get a reliable measurement that includes baffle step influence either way.  What you won't get, when measuring on its side, is all the non-minimum phase interactions from all the reflections and the changing directivity from moving between half-space and free-space due to the distance to the ground.

Think about what happens when you stand a tall, thin array speaker upright.  The acoustic environment isn't as consistent because of changing radiation angles and reflections.  At very low frequencies, the ground is within 1/4 wavelength of all drivers.  So the whole speaker is radiating into half-space.  As frequency rises, some drivers are within 1/4 wavelength and others are not.  So some drivers are radiating into half-space, others into free-space.  Further, the ground acts as a reflector for drivers further than 1/4 wavelength, causing non-minimum phase interference for those drivers.  As frequency rises further still, all drivers are further than 1/4 wavelength and are radiating into freespace, with the ground acting as a reflector.  A little higher in frequency, the baffle becomes acoustically large and it begins to influence directivity.  At this point, the speaker begins radiate into half-space again, this time because of the baffle and not because of the ground.  So you see, standing upright causes a lot of peculiar interactions.

If you want a true anechoic measurement of the speaker, free of influence from the environment, a proper ground plane measurement is the only way to do it.  Well, you could suspend the speaker 50 or 60 feet above the ground, and suspend a microphone up there a few meters away.  That would give you an anechoic freespace measurement.  But I think the anechoic half-space measurement is just as good, and much easier.

In spite of all this, I think there is merit in making measurements standing upright too.  That's the way the speaker will be used, so it makes sense to measure it that way.  One of the strengths of arrays, in my opinion, is their "tolerance" for floor bounce.  The reflections from the floor act sort of like virtual array nodes, like extra speakers in a longer line.  It causes dense interference which smooths the response curve.  I would suggest it would be best to make several measurements with the microphone in different positions, to show the energy distribution in the listening area around the array.  This will take into account all the interactions between drivers, reflections and changing directionality with respect to frequency.

But if I were to be asked to make a single measurement of a speaker, and it was to be used to compare with other speakers, I'd do an outdoors ground plane measurement, where all drivers were within 1/4 wavelength of the ground.

[HAL]

An observation would be that the microphone is now a boundary effect mic like a Crown PZM with half sphere coverage.  The only reflections are from other boundaries that the mic is not placed on.

Doubt anyone has a calibrated PZM, but will give it a try to see what happens with CLIO.