Got a question for all you smart people

I just received a 990/7700 combo and am in the process of setting it up. I normalized the speaker level with a SPL and the 990 internally generated test tones.

I assumed that this normalization would hold true for any pink-noise source. But when I checked my work with Avia disk's pink noise I got radically different results (in the order of +/- 6DB for some of the speakers).

Why is this? Which tone set would you recommend I utilize, the Avia DVD or the 990’s internal test tones.

Thanks for your help

-Peter