|
|
| Re: 4 Pi Design Questions [message #72887 is a reply to message #72885] |
Fri, 01 June 2012 21:35   |
 |
Wayne Parham
Messages: 18786
Registered: January 2001
|
Illuminati (33rd Degree) |
|
|
Yes, now you're getting good measurements. You still see room modes down low, like the one around 120Hz. This is a vertical mode, and is part of what flanking subs fix. Below that, the response is dominated by room modes - both speakers look the same in the modal region - proving that what you measure down low is not the speaker, but the room. But the response up high shows up pretty clearly. This is a useful setup when you just want to see the crossover region, for example.
I don't use any smoothing at all. The difference is mostly due to my measurement being outdoors. Inside, on its back, my measurements look very close to what you have. I have an eight foot ceiling, and I'm guessing you do too.
You can kind of see the similarity in my indoor measurements and yours in the "Vertical Nulls" video in the "Crossover Optimization" thread. See the wiggles down in the ~100Hz region? Those are room modes.
I think the Omnimic measurements made by BigmouthinDC do have some smoothing applied. I'd say probably 1/12 octave. It actually looks less like smoothing and more like the sampling resolution is a little bit lower than my LMS system.
As an aside, there is a difference between reduced resolution and smoothing, but the end result is similar: Detail is lost.
The biggest difference is smoothing will actually reduce the amplitude of peaks and increase the amplitude of troughs. Reduced resolution simply misses some data points, so a sharp spike may be missed. But where a point is recorded, its amplitude is left intact, and the system just connects the dots to make the SPL chart.
Smoothing really takes the edge off a graph and can make a peaky response curve look pretty smooth. Huge spikes just get knocked off. Then again, it isn't necessarily a bad thing, since most published curves are smoothed, sometimes as much as 1/3 octave. That was pretty common in the 1970s and 1980s - if you could find a chart at all, it was likely smoothed to 1/3 octave resolution.
|
|
|
|
|
|
| Re: 4 Pi Design Questions [message #72895 is a reply to message #72892] |
Sat, 02 June 2012 10:30 |
|
Wayne Parham
Messages: 18786
Registered: January 2001
|
Illuminati (33rd Degree) |
|
|
The measurement above 200Hz is pretty much the same as a standard four π loudspeaker when measured in this manner. And yes, frankly, that's a very good chart. But below 200Hz, it's ambiguous because only room modes can be seen. That's where you would expect most changes due to the box modification to manifest themselves.
This brings us to the hypothetical question, which is, "How much does it matter?" My concern about box mods is always that they might make ripples in the midrange due to internal standing waves. But as long as they're all below the modal range, one might question how much they matter. If we use flanking subs to mitigate midbass and lower midrange anomalies, they'll tend to smooth that range, whether caused by standing waves inside the box or reflections outside, in the room. Still, my take is the less anomalies we can introduce, the better. Less problems to mitigate.
And of course, if we're using subs, we don't need the extra extension the larger box gives, and leaving it stock eliminates the possibility of introducing midrange ripple. The room is dominant below 200Hz, and the biggest peaks and valleys are from standing waves in the room, not in the loudspeaker box. So truly, the best approach is to use multisubs (flanking and distributed) to smooth this range, which will provide both extension and modal smoothing.
My conclusion is this: Since the mods have not introduced any additional ripple above 200Hz, I'd call it "verified". That's the main thing we're looking for here. We don't want any midrange ripple from internal standing waves. Of course, we can't see what it might be doing below 200Hz, but indoors, that almost doesn't matter.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Posting Rules
Members
Search
Help
Register
Login
Get a T-Shirt!
