>From: "Dario Bonazza" >Subject: Re: Amazon buys dpreview.com > > > > > OK, that looks like a good method for removing variations. > > Unfortunately, it will also be completely useless for customers, as none > > of > > them will ever use that camera that way. > >Very few use cameras the way they are tested anyway. >A good test removes as many variables as possible between the articles >being >compared. > >William Robb >
You both of course are correct, which highlights the difficulty in building and conducting tests that are universally meaningful to all persons. The best tests would include RAW output with the same sample of the same lens on each camera body (accepting the variables induced by the in-camera raw processing and the post-processing .jpg conversion). They would also then demonstrate the same, with a common lens(es) of that camera system at RAW, and same with .jpgs at the default setting. I don't see much point in conducting actual comparitive quantitative tests with the camera at other than the default settings as those adjustments are highly variable based on camera brand and personal preference enters in quite a bit. I think dpreview has done a decent, if not not totally complete job, in providing test results that can be compared. I consider a test image made with a same brand camera system lens that an individual is likely to acquire, to probably be more meaningful to in the hand camera system performance, than an image produced with a lens one is unlikely to own. Tom C. -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net
