measuring in-room response of speaker system
I was so excited about my new DIY FT17H supertweeters, I decided to go to Radio Shack and pick up an analog SPL meter to make some measurements with the stereophile test CD 2 warble tones and see if I could figure out why things sounded so much better. Up to about 8kHz, all looks really good; no peaks/dips of more than 3-4db across the whole range of 20Hz to 8kHz, so my T-line subs have been well integrated with my Alpha's. However, much to my surprise/confusion, when I try to measure the treble response, things are really weird. Everything appears linear to about 8kHz, but then things appear to drop off sharply at 10kHz and above, with all readings being -8 to -10db (this is with the supertweeters attached crossed at somewhere around 18kHz with a 1uF cap, only a 6db/octave slope, so there should be significant output below that, output I can easily hear). The strange thing is that I can hear all of the test tones very clearly (even the 20kHz one) at my listening position, which is where I set the meter to make the measurements. Also, I can hear all of the tones clearly from my main tweeters, but the 20kHz I can only hear from the supertweeters. Does anyone know if the radio shack meter is really inaccurate at high frequencies, or do I need to start figuring out what's "wrong" with my system to cause this tremendous roll-off? The system sounds better than ever, so I'm thinking perhaps the meter is very poor. I know for low bass it's no good, so I had to use some "fudge factors" I found at numerous sites on the net from people who sat down and figured out with much more expensive meters what the deficiencies of the radio shack one were, but most of these people seemed to think that it worked pretty well for higher frequencies.
Thanks for any input