Nearfield measurements have no baffle losses, right?

[Dirk]

That is all.

[bmaupin]

Re: Nearfield measurements have no baffle losses, right?

If by "nearfield" you mean with the mic at the baffle surface, then yes. I haven't done any other close proximity measurements to otherwise comment, but it would be interesting to see how the response changes as you move away inch by inch.

[Dirk]

Re: Nearfield measurements have no baffle losses, right?

I mean with the mic an inch from the cone.

I'm trying to figure out the best way to measure these TMWW towers. I had a previous thread here but I don't know if my measurement-fu is strong enough to intelligently go splicing nearfields and farfields and whatnot. I tried Curt's recommendation of measuring the woofers at 1m with a bunch of blankets and pillows on the floor to help damp some of the floor bounce but I still see a lot of ripple up as high as 700hz (see attached, blue = far, red = near). Since I think I need to cross below the point where ripple rears its head I'm looking at all my options.

If nearfield measurements do NOT include baffle losses, then I'm wondering if I can just import the FRD files into Jeff B's Response Modeler to add the losses and then go about my XO design.

[billfitzmaurice]

Re: Nearfield measurements have no baffle losses, right?

I'm trying to figure out the best way to measure these TMWW towers.

There's two schools of thought there. The most popular is to measure them outdoors, then if you want to account for BSC etc. you'll have a speaker that's flat in half space. The only problem is that you don't listen to speakers in half space, and in room they'll be totally different. Option 2 is to measure in room with the mic at the average listening position and go from there, keeping in mind that severe narrow peaks and dips are boundary reflection sourced and shouldn't be considered for passive correction, that's better handled by DSP, if you bother at all.

[curt_c]

Re: Nearfield measurements have no baffle losses, right?

If nearfield measurements do NOT include baffle losses, then I'm wondering if I can just import the FRD files into Jeff B's Response Modeler to add the losses and then go about my XO design.

Nearfield mesaurements, for all practical purposes, do not includes baffle edge diffraction effects. They do have a limited bandwidth determined by the driver cone diameter. For a typical 10" driver, this limitation would be frequencies under approximately 750 Hz.

Modeling the woofer in box low frequency response and diffraction effects using Jeff's nifty program would probably be the easiest, and likely the most accurate choice. -Although you will still want to splice in your far field measuements somewhere above the baffle step.

-Or you can simply smooth everything to 1 octave. :D Not very accurate, but boy does it look good! A process used by design professionals everywhere to get great looking plots....

C

[Dirk]

Re: Nearfield measurements have no baffle losses, right?

There's two schools of thought there. The most popular is to measure them outdoors, then if you want to account for BSC etc. you'll have a speaker that's flat in half space. The only problem is that you don't listen to speakers in half space, and in room they'll be totally different. Option 2 is to measure in room with the mic at the average listening position and go from there, keeping in mind that severe narrow peaks and dips are boundary reflection sourced and shouldn't be considered for passive correction, that's better handled by DSP, if you bother at all.

Yeah, I'm having a hard time "looking past" the hills and valleys caused by the room. I think to get the results I want I'm just going to have to learn the process. The situation is exacerbated by SoundEasy, which really needs a manual just for the peasantry. Don't get me wrong, the documentation is a great technical guide, but it could really do with some footer / ledger notes on the theory and rationale behind each step for the rookie.

If I'm merging nearfield and farfield, I need to at least get my farfield measurements BELOW the BSC corner correct? Because with a 10.5" baffle, that means I'd need to get farfield down to about 400hz or so and I'm seeing ripple effects almost an octave above that.

[Dirk]

Re: Nearfield measurements have no baffle losses, right?

-Or you can simply smooth everything to 1 octave. :D Not very accurate, but boy does it look good! A process used by design professionals everywhere to get great looking plots....

C

Tempting. I'm pretty sure that's what Tang Band does for their spec sheets.

How did you measure the HDS in the Mavericks? Let me guess -- you didn't, and it was accomplished by touching the cone and achieving a zen-like state of true enlightenment. ;)

[curt_c]

Re: Nearfield measurements have no baffle losses, right?

I measured it like I suggested previously: 1 meter on the woofer axis with a stack of foam on the floor. I just mentally attempt to 'remove' the residual floor bounce effects when I design the crossover, (The TLAR method in action!:D) and fine tune any irregularities when I voice it.

Yes, I'll admit that using this method I've had to substitute a different value inductor from time to time, and trim the odd mid and tweeter attenuation a bit. John's ground plane measurement suggestion is a good one, and would double the accurate low frequency response in a given room. Remember, all else being equal, subtract 6 dB from the ground plane measurement compared to the free standing ones.

C

[MSaturn]

Re: Nearfield measurements have no baffle losses, right?

More argument for active electronics, I suppose: if one were to build a stationary "reference system", using DSP and active xovers would make it possible to - after averaging reflections and power response - actually get very near a linear response at the listening position.

I still adovocate speakers that can be moved around, and if my dollars weren't so tight, I WOULD be going active ... oh well, I'll stick with passive for now.

[curt_c]

Re: Nearfield measurements have no baffle losses, right?

More argument for active electronics, I suppose: if one were to build a stationary "reference system", using DSP and active xovers would make it possible to - after averaging reflections and power response - actually get very near a linear response at the listening position.

I still adovocate speakers that can be moved around, and if my dollars weren't so tight, I WOULD be going active ... oh well, I'll stick with passive for now.

Yes, but that opens another measurement can o' worms.

- after averaging reflections and power response -

Seems like we are in the same boat here: How / where to measure to obtain that ideal response -and still only at a single listening postion.

-And of course, you can't equalize out a dip...

C

[Dirk]

Re: Nearfield measurements have no baffle losses, right?

Curt, I tried your "damped floor reflection" method but I don't know how much it helped. Then again, I didn't do the OBVIOUS thing and try it once without to see what happened. I'm usually a pretty methodical person when it comes to problem solving, but I fell down flat this one.

I didn't take anything down last night, so I'll try again once I get the kidlets in bed and find some way to get the dogs to stop running around upstairs... :rolleyes:

[dthomas]

Re: Nearfield measurements have no baffle losses, right?

Curt, I tried your "damped floor reflection" method but I don't know how much it helped. Then again, I didn't do the OBVIOUS thing and try it once without to see what happened. I'm usually a pretty methodical person when it comes to problem solving, but I fell down flat this one.

I didn't take anything down last night, so I'll try again once I get the kidlets in bed and find some way to get the dogs to stop running around upstairs... :rolleyes:

I do similar to what Curt is suggesting. I measure in room at 1M use egg crate foam on the floor between the woofer and the mic. It does not help much but it does some. I then use Jeff's response modeler and model the driver in the desired enclosure and patch that to my measurement at 300Hz. Add Baffle step losses based on what I think is needed depending on where the speaker is intended to be used. It seems to work pretty well.

I use BDS to work out position of drivers on the baffle and attempt to mitigate potenital floor bounce issues.

[Dirk]

Re: Nearfield measurements have no baffle losses, right?

SUCCESS!

I measured with a bare floor and then with a stack of dense pillows and corduroy blankets on the floor (I used a lot more dense stuff tonight than last night). The majority of the ugliness is gone.

The remaining difference between the 6ms gate and the 15ms gate is minimal. The 15ms gate has burps of +/- 2.5dB, but if you squint at the monitor it looks a lot like the 6ms measurement. There's a slight variation of *maybe* a dB at 300hz. Anything wrong with using the 6ms gate? I'll attach a screen shot in a bit.

I can now see why people do 2-ways with the 8" Silver Flute. Veeery smooth midrange.

[critofur]

Re: Nearfield measurements have no baffle losses, right?

...and find some way to get the dogs to stop running around upstairs... :rolleyes:

I find that non-constant extraneous noises don't seem to affect my measurements if I do say, 3 sweeps + average.

[Dirk]

Re: Nearfield measurements have no baffle losses, right?

You're evidently unaware of the kinetic energy of two Corgis stuck in the house all day due to rain...