did a bunch of tests this afternoon with a 1kHz sine wave into an 8ohm celestion neo creamback 1x12.
the sound meter i used was a Draper, with +/- 1.5dB accuracy.
one thing i'd like to point out is that reflections from flat surfaces and the resulting interference patterns make a huge difference to the results. moving around the room while testing i could hear myself the various standing waves that were setup. the difference in my case could be up to 10dB which is a massive difference. this was particular noticeable close to the cab. a guitar signal, with a range of frequencies, would minimise these interference effects vs a fixed frequency sinewave.
distance test, measured from front of grill cloth:
1w at 1m 95.3dB(A)
1w at 2m 94.0dB(A)
1w at 3m 92.7dB(A) <- this was the furthest I could get back
reflections in the room prevented the theoretical 6dB drop off when doubling distance from sound source. this would also happen in a gig venue of course. So sound levels will greatly depend on the architecture of the space.
power test, measured using a Fluke true RMS meter into a 7.4ohm load:
0.5w at 1m 97.4dB(A)
1w at 1m 100.1dB(A) <- this is a around 5dB different to the previous test, which I'm putting down to microphone position
2w at 1m 102.5dB(A)
4w at 1m 104.8db(A)
As you can see there is nearly the theoretical 3dB difference when doubling the power. Its actually 2.5dB. Perhaps due to losses due to heating in the speaker coil.
Celestion claim 97dB sensitivity for this speaker, which I guess it could be. Reflections meant I couldn't get anything like an accurate measurement. I had a look into how sensitivity is measured, and its done using a huge baffle board, much larger than used in a cabinet. They are also done in an anechoic chamber to remove reflections. As a result, I would expect the real world sensitivity of a cab would be a bit less than the official sensitivity.
