Another excellent suggestion there JA. Certainly a much more scientific way to carry out the tests. This however may be where the problem lies. A test on real DSOs is a real world test WITH the uncontrollable variables in place. Agreed that these variables will affect the quality if the data. If all 3 cameras are being tested under the same conditions ( as much as is practical as obviously time of the night can change sky glow etc ) then the results should have a greater applicability to the reason for testing which is creating images of DSOs. The other small consideration is the need for more equipment ( cheap as it may be )to carry out the test.
Thank you for your input JA. It's good to get other perspectives to validate the quality of the test.