Some samples from the Sigma 17-70

[996GT2]

I got my Sigma 17-70 Macro today, and so far am very impressed with the image quality. Even at close to minimum focusing distance, the lens is still quite sharp. Overall, the build quality is very slightly worse than my previous Sigma 18-50 f/2.8 EX, but the image quality is very comparable and I have noticed that the AF in the 17-70 seems more accurate (although a little more "scratchy" sounding).

The lens I got came with a 72mm Tiffen UV filter, which is pretty low-end as far as filters go (however, it is worth noting that Canon branded filters are in fact made by Tiffen). So, to show the effects with/without a filter, I put my 40D in manual exposure mode, turned on the live view, set focus manually, and took a series of 4 shots to show the differences in IQ with/without a filter.

None of these shots were touched up in any way. The shots of the watch were slightly cropped to fit under Photobucket's 1 MB file limit. These pictures were saved as JPEG quality 9 in Photoshop (12 is highest, 1 is lowest).

The second shot of the quarter was slightly cropped to fit in under 1 MB (but not resized, so the quarter itself is the same size in both)
Quarter, with UV filter
Quarter, No UV Filter

Watch, with UV Filter
Watch, No UV Filter

There is clearly some loss of detail, sharpness, and contrast with the filter on.

 

[PurdueRy]

Wouldn't a UV filter only have significant affect if you significant distance between you and the subject? I see very little difference between the two shots of which the differences could be cause by minor differences in your camera/lighting.

 

[996GT2]

Originally posted by: PurdueRy
Wouldn't a UV filter only have significant affect if you significant distance between you and the subject? I see very little difference between the two shots of which the differences could be cause by minor differences in your camera/lighting.

No, the camera was held in the exact same spot (tripod) with the exact same lighting and aperture/shutter/ISO. Basically, all conditions were held constant except for the use of the filter.

If you look closely, you can see the loss of detail and contrast when the filter is on.

 

[andylawcc]

Originally posted by: PurdueRy
I see very little difference between the two shots of which the differences could be cause by minor differences in your camera/lighting.

if you look at it at 100%, you will notice there is a loss of detail on the face on the quarter. However, at my "IE-preshrunk-screen" size, I really can't tell the difference.

 

[corkyg]

The UV filter really is not involved with those shots. Frankly I don't see much difference. UV on or off should be done outdoors on a landscape scene where there actually is ambient UV radiation.

Many pros use UV or ND filters simly to protect their objective lenses.

Aside from framing problems, those are excellent closeups.

 

[angry hampster]

Originally posted by: corkyg

Many pros use UV or ND filters simly to protect their objective lenses.

.

This. However, single-coated filters are junk. All of them. As cheap as multicoated UV filters are, there's no reason not to get that over a single coat. Anyway, all of our lenses at work have multicoated UV filters, and we've got ND filters for the wide-aperture (f/1.2-1.4) lenses so they're usable wide-open during the day.

 

[996GT2]

Originally posted by: corkyg
The UV filter really is not involved with those shots. Frankly I don't see much difference. UV on or off should be done outdoors on a landscape scene where there actually is ambient UV radiation.

Many pros use UV or ND filters simly to protect their objective lenses.

Aside from framing problems, those are excellent closeups.

I was trying to test the effects of putting an extra piece of glass in front of a lens. A perfect UV filter, in this case, should have zero effect, meaning the 2 pictures should be the same. However, if you look closely at the 100% images (not the IE/firefox resized ones) of the quarter and watch, you can notice loss of detail, especially in the quarter. Loss of contrast is also noticeable.

Basically, this kinda backs up my point of always using lens hoods and not UV filters to protect lenses.

 

[fuzzybabybunny]

I've done a similar comparison before using Hoya Pro 1 (thin 1mm glass), Hoya ND, and Hoya CPL multicoated filters, and I wasn't able to see much loss of sharpness or contrast from the Pro1 UV. The worse was the CPL, which kinda makes sense because you're shooting through two layers of glass. The ND filter performed quite well. I would have to conclude that UV multicoated filters do produce noticeably better results than single coated UV filters, and that UV multicoated filters do not degrade image quality all that much. But once you start stacking them... well, noticeable loss of sharpness and contrast result.

http://fuzzybabybunny.smugmug...._8Nt4s#124187930_j4iES

 

[rivan]

Originally posted by: 996GT2
(however, it is worth noting that Canon branded filters are in fact made by Tiffen).

A single manufacturer of two separate items means what, exactly? Purina makes several brands and tiers of dog food, including brands they don't have their own name on. Are all their dog foods the same?

Originally posted by: 996GT2
I was trying to test the effects of putting an extra piece of glass in front of a lens. A perfect UV filter, in this case, should have zero effect, meaning the 2 pictures should be the same. However, if you look closely at the 100% images (not the IE/firefox resized ones) of the quarter and watch, you can notice loss of detail, especially in the quarter. Loss of contrast is also noticeable.

Basically, this kinda backs up my point of always using lens hoods and not UV filters to protect lenses.

I see a miniscule loss of detail with a UV filter on, but you yourself mention it's a low end filter. The kicker is that I also see other significant differences, at least in the watch: the reflections in the bezel, the added details in the shadowed portion of the dial in the filtered shot, etc.

From my point of view, if a .05% loss in sharpness is killing your shots, you've likely got larger issues.

You can draw the conclusion of not using filters if you like. I'll continue using a ND filter to protect my lenses.

Edit: grats on the new lens... looks great

 

[ghostman]

From my point of view, if a .05% loss in sharpness is killing your shots, you've likely got larger issues.

I have to agree with this. Even with the less-than-top-of-the-line filters that I use, the difference between a good and bad photo I've taken has never come down to such small differences.

Things look pretty sharp with this lens. Not sure how much research you do to keep digging up decent, overlooked lenses, but good work. I do notice some purple fringing going on on the top right quarter of your photos. Not sure if that's an effect caused by your lens or just a sensor thing reacting to the bright light.

 

[magomago]

Originally posted by: ghostman

From my point of view, if a .05% loss in sharpness is killing your shots, you've likely got larger issues.

I have to agree with this. Even with the less-than-top-of-the-line filters that I use, the difference between a good and bad photo I've taken has never come down to such small differences.

Things look pretty sharp with this lens. Not sure how much research you do to keep digging up decent, overlooked lenses, but good work. I do notice some purple fringing going on on the top right quarter of your photos. Not sure if that's an effect caused by your lens or just a sensor thing reacting to the bright light.

I agree....much like with the poster above this one, I saw some detail increase elsewhere and actually couldn't figure out exactly what the OP was complaining about.

But anytime you have to pick up at 100% and look closely, then I personally wouldn't care too much For such a 'low end' filter, it, in this situation, did surprisingly well.

I suppose I went overkill slapping a UV filter and my hood over my 16-45

edit:

A canon shooter, eh? Now this thread makes more sense....


.....j/k....maybe

 

[GrJohnso]

Lens looks nice... Not a bad range to work with... As for your test shots, putting the filter on changed your focus point the tiniest amount. This contributed to the slightly softer look on the quarter, and seems to have affected some of the details on the watch. A 1-2mm shift makes a difference... However, the watch shot has some glare and lighting differences between the two shots that seem to affect the contrast as well.

Anyway, as stated, filters do have an affect on final image quality. Better filters have less of an affect. Sometimes it's worth it, sometimes it's not... UV indoors, not generally a necessary move. Outdoors, sometimes. Polarizers and various ND filters, worth it when used correctly. Benefit almost always outweighs the cost/risk...