nospam wrote:
In article , Chris Malcolm
wrote:

If as was claimed more than half Sigma's lenses are bad, then a random
sample (e.g. me and friends) of 8 Sigma lenses which are ok is
probabilistically better than 8 heads in a row when spinning a
coin. That's enough to make one wonder whether there is some trickery
going on, enough to make one question whether the original claim about
Sigma unreliability and inconsistency was correct. Not enough to
refute it of course. It's possible to spin 8 heads in a row by lucky
chance. Just pretty unlikely.

how did you test them?

Wide angle lenses, so I did my usual specific tests of gravel beaches,
grass lawns with tape measures laid out on them, gravel beaches, brick
walls, and examined the results carefully with side to side, top to
bottom, edge to centre pixel peeping comparisons. I was surprised how
good they were at f8, I'd have been happy at the price if they'd been
worse. Especially the 8-16mm, which was noticeably sharper at f5.6
than at f8, one of the signs of a rather good lens.

One thing I quickly discovered is that wide angle lenses for various
technical reasons I don't fully understand autofocus poorly with at
least some cameras, including mine, which has given some of these
lenses an undeservedly poor reputation. I always manually focus
them. Which for most shots simply means carefully finding a good
hyperfocal setting and leaving it there.

And I always examine the results of any good shot carefully at pixel
level in difficult or critical areas. That sometimes gives me clues
about new aspects of lens or camera performance. It also sometimes
alerts me to unsuspected damage such as slight decentering.

--
Chris Malcolm