The Filters We Need With Digital...

Started Oct 5, 2008 | Discussions thread
OP Joseph S Wisniewski Forum Pro • Posts: 34,130
The "negative SA" theory gathers strength...

Iliah Borg wrote:

Joseph S Wisniewski wrote:

Iliah Borg wrote:

I have just bought a
Tamron SP 300mm f2.8 LD IF and while searching for sites with more
info about it I came across a lens test with a remarkable
finding...As you probably know, Tamron used to offer an optional
112mm "normal" filter for the front of the lens. I dont know what a
"normal" filter is but apparently the tester has found his Tamron
300/2.8 to be considerably sharper with this filter on the lens than
without it!!!

I had a similar experience with Canon 1D MkII and long lenses. A UV
filter improved AF performance of the camera.

Iliah, that is just plain weird.

Can't agree more.

But I have a possible explanation. See my "negative SA and astigmatism" reply to ALf a little earlier in this thread.

Not that I'm doubting you, I've seen
enough weird stuff in this world to know that "weird happens".

But I can't imagine how, unless the AF system was unusually sensitive
to UV, so that whatever was making it through the lens was enough to
drift the AF system off a bit (since the lens isn't corrected for
near UV, or even for violet). Even then, a long lens should block
pretty much as much UV as a filter...

Were you using a stronger UV filter, a UV(1) or something that cuts
off near (or even above) 400nm, like those 415nm UV filters I use for
blacklight sessions...

I use regular slim MRC Schneider 010 type. The experiment was quite
simple: a resolution target lit by sun, 50 shots AF on with filter,
50 shots AF on without filter. Lens was focused to infinity manually
between shots. Resolution was consistently higher on the shots with
the filter.

That ties into the SA theory. The AF sensors tend to look down either 10 degree (f2.8) or 5 degree (f5.6) areas on the lens's exit pupil. Now, the 1D II should have multiple aperture sensors (both f2.8 and 5.6), and for a lens f2.8 and faster, it will use the 10 degree sensors. So, you will typically see a lot of focus shift at smaller apertures.

Do you remember what aperture you shot your test at?

The big problem with trying to compensate for SA with front filters is that you need to add an amount of compensation that would differ depending on how undercompensated the lens already is for SA (overcompensated lenses would be made worse by a front filter) and you would also need to balance this by how much additional SA is being added by the camera's sensor cover glass, IR filter, and AA filter.

Shots without filter exhibited slightly erratic focus
behaviour, with 20% of shots being 8% lower resolution and 40% being
4% less resolution then the shots with the filter on. Only 10% of
shots without the filter were of the same resolution then the peak
resolution obtained with the filter. 70% of shots with the filter
achieved maximum resolution; 20% of shots with the filter were 4%
less then maximum resolution, 10% of shots with the filter were 8%
less them maximum resolution. On the contrary, 10% of the shots
without the filter resolved 30% less then maximum.

That's quite reasonable. You have the same amount of AF error with and without the filter, the filter is just moving the center of the distribution so that there is less disagreement between AF sensor and final image.

I hesitate to attribute this to the effect of UV on the AF system. I
started with a hot mirror because the reports were that AF misses
were mostly under hot outdoor conditions. Getting better AF behaviour
with hot mirror I decided to recheck with a simple UV blocking
filter, and it turned out AF improvement with UV filter was nearly
the same as with hot mirror filter. My colleague used a Tiffen clear
glass filter on his Canon camera with the same AF improvement results.

If you want to eliminate any variability in that part of the test, I'd suggest a green filter.

p.s. when you made your monochrome D2X, did you recycle the original cover glass? I used quartz (well, OK, fused silica) but my new cover was thick, about 2mm, probably as thick as the total of the original cover glass, LiNbO3 AA filters, and IR blocker, so it didn't really alter the SA characteristics of the camera. Next camera, I have to find some 0.5mm quartz. With UV to IR broadband anti-reflective coatings.

And I wonder if lens makers have been designing lenses for proper compensation with a filter in front of the lens and nothing behind (that's the way we usually shot B&W film and early color film), and if they're now adding any compensation for all the stuff in front of the sensors with this "designed for digital" movement.

-- hide signature --

Rahon Klavanian 1912-2008.

Armenian genocide survivor, amazing cook, scrabble master, and loving grandmother. You will be missed.

Ciao! Joseph

 Joseph S Wisniewski's gear list:Joseph S Wisniewski's gear list
Canon EOS 5D Mark II Nikon D90 Nikon D2X Nikon D3 Nikon D100 +43 more
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow