Nomad Audio: Power Response Technical Paper

[PHilgeman]

http://www.nomad-audio.com/index_fil...r_response.htm

-Paul Hilgeman

[David Gatti]

Informative, concise & clear paper Paul.

2 points:

Provided the listener's ear is at the right height, does it really matter what the vertical off axis response is, assuming the power response is OK. The floor (carpet) will have a different reflection signature than a hard ceiling anyway.

Would you agree that the horizontal axis response of a low (eg 1st) order crossover will suffer far less of a transition "flare" between bass-mid & tweeter than a high (eg 4th) order and may even be equal to the coaxial design in this regard?

[PHilgeman]

Originally posted by David Gatti
Informative, concise & clear paper Paul.

2 points:

Provided the listener's ear is at the right height, does it really matter what the vertical off axis response is, assuming the power response is OK. The floor (carpet) will have a different reflection signature than a hard ceiling anyway.

Would you agree that the horizontal axis response of a low (eg 1st) order crossover will suffer far less of a transition "flare" between bass-mid & tweeter than a high (eg 4th) order and may even be equal to the coaxial design in this regard?

Hey David,

1) I think that it does matter. The only thing that has that precise on-axis sound is what arrives at your ear. EVERYTHING else is in some way or another made up of the off-axis sound times the transfer function of what it reflects off of. Yes, different floors and ceilings can have far different transfer functions as will walls etc, and all will arrive at the listener at different times. However, would it not be best if regardless of these transfer functions, which we don’t know anyways, to have the same tonal balance and same impulse response prior to the reflection? The areas where the off-axis response of the Ronin is most similar are in the areas where there will be the least differences in transfer function anyway. The difference between carpet, drywall and hardwood will mostly be in the 1K plus range. Maybe my next paper.

To sum this up, you say "assuming the power response is OK", my point is that the power response is not OK, and most people haven’t heard a speaker that does have proper power response, unless they listen in an anechoic chamber where you only hear what comes directly at you. With standard closed box, non-coincident, non-dispersion controlled tweeter speakers, not only is the frequency response that is reflected off any nearby surface vastly different, the impulse response is horrendous. Imagine a speaker with the midrange mounted 10 cm behind the tweeter, that’s what you are seeing in the "Test Sample 50 Degrees" plot. Off the ceiling reflection, you hear the tweeter first, then the midrange output. Off the floor you hear the midrange output, then the tweeter. Just because a crossover type has theoretical even power response does not mean that is what you hear. The things you hear most are the on axis, the first reflection off the floor, the ceiling and each wall. That’s why it is critical to take a look at the 30-45 degree range above and below the speaker. The horizontal is hard to determine what is the important range to closely inspect, but with toed in speakers it can be as far out as 60 or 70 degrees depending on room geometry etc.

2) In certain cases, I would agree. I can perform these tests again, possibly with a 1st order system and a smaller midrange. Let me point out however, that even in this case, using the ScanSpeak 15W/8531, dispersion starts to narrow really low. This was the most shocking part to me. Tymphany who tests on IEC baffles always shows 0, 30 and 60 as one line up to about 1500Hz, Seas plots however show data similar to mine. It has to be due to mounting in a real box as opposed to an IEC baffle, either way, it is real and it is there, so we must deal with it. Using a smaller diameter driver more suitable for 1st order crossovers may or may not get rid of this. I have an older 13cm vifa driver that is ruler flat on axis through about 8500Hz. It has a bobble at 900Hz, but so does the Scan. Given what I found in this test, I would believe that it's dispersion would be similar but the effect would have a lower Q centered at a higher frequency due to the fact that the crossover is higher. Keep in mind that on the lower side of the crossover frequency in the off-axis areas, it is easy for the tweeter to boost the response of the woofer up due to it's wide dispersion, but above the crossover frequency, the woofer's narrow dispersion wont 'bring down' the tweeter's wide dispersion up into the 5K range.

It is a mixed bag like most things. I won't have any down time until august, but I will probably do an addendum to this paper. Any recommendations for drivers? To keep things different enough, I will probably get out my big 1.5" round over bit too and do a something like the Seas MCA12 or MCA 15. No, wait, I have the perfect driver for this test, perfect. I will put the details in the write-up, but maybe use the small faceplate version of the XT25 for its extended LF response as well as ultra close driver spacing. This should give nearly the ideal situation for a 1st order crossover.

Ok, maybe that response lacked the conciseness of the paper, but this are big topics that people never really get into around here.

-Paul

[David Gatti]

Thanks for clarifying Paul. I'm thinking about a new project at the moment, and I do like the elegance of your design. A monopole version would be very interesting too, for those of us without the space for a dipole.

[PHilgeman]

It is actually pretty forgiving, I have had them set-up with only about 18 inches from the rear of the boxed section to the rear wall with excelent results, perfectly acceptable. This puts the front at about 30 inches.

It is only better with them out away from the wall though. However, they are fairly light weigt at about 75lbs apiece and are pretty easy to lift, so moving them is not as bad as many speakers with these capabilities. A monopole version would probably tip the scales at 125 lbs or so.

-Paul