Trust me, they all the same speaker, but measured using different methods, and/or a different design axis.
Picture 1 (the 1/3 octave smoothed one)
This was measured later in the design process with the mic about 10 degrees off orthogonal with respect to the baffle, and centered between the tweeter and the lower woofer. I don't recall if I used a sweep or MLS for this plot.
Early in the design process where I was verifying the crossover frequencies and power response. The difference is due to a slightly different crossover, and the fact it was measured on the tweeter axis orthogonal to the baffle. The key piece of information here is not the absolute response, but the relatively small variation in response even off 60 degrees. If you look at a Stereophile measurement, they show the on axis response as 0, and show the off axis responses as changes from the nominal value. This then shows only the difference in response, and results in a much nicer looking plot. Yes, the little Vifa tweeter has a bump on axis, but as you can see, it behaves much better when measured off axis. Sure, from a purest sense, it is not an audiophile design. -But then it was never intended to be.
This was measured using stepped sine waves at 1/12 octave spacing. Due to the difference in measurements format, it will not necessarily look like an MLS or swept sine measurement, as the excitation frequencies may, on may not coincide with the response variations of the driver under test.
Measurements are not, and never will be the panacea we would like them to be. The results are all too often dependent on the measurement conditions and methods under which they were obtained.
Originally Posted by tuxedocivic
It is strange, but Curt says right on the page he had a hard time measuring this speaker. Still funny that most of the anomolies between measurements are in the upper frequencies, where getting a reliable measurement shouldn't be hard.
I'm confused by some of the things said on his write up. He says the speaker is 92 to 94db/2.83V sens, and that 20watts yields 109db. That would take at least 32 watts??? I would think Curt can do math better than that. It's also a 4ohm speaker, so that further adds to 20 watts NOT netting 109db. Hmm.
He also says the sensitivity is approaching pro audio... I'm not so sure. Approaching I suppose, but still a ways off the mark.
Cool design anyways, and cheap too! Look forward to watching the build!
The USPL of the Aura driver is 90 dB/2.83volts/meter. Three sets in parallel will increase that sensitivity by approximately 9.5 dB. Depending on how close they are placed to the front wall determines their ultimate sensitivity. I intentionally conservatively rated their basic sensitivity, but modeled the drivers in CAD to determine their max SPL.
-I never was any good at math