Focus on New Canon 70-200mm f2.8 L IS seems soft

[jimcrotty]

I just purchased a new (at least I hope it's new) Canon 70-200mm f2.8 L IS lens to replace my older 70-

200mm f4. I've been doing some practice shots with the new lens and my Canon 1D Mark III. It just seems

that the lens is not quite as sharp as I had hoped, especially when I compare to my Canon 300mm f4 IS.

Any help or suggestions ? I will try to attach an image I took this morning with the new 70-200mm, of a

Raven, shot at iso 400 in program mode with the IS on.<div></div>

[shambrick007]

I don't shoot many birds, but it looks a lil softer than what I get outta mine. What f-stop/shutter speed was this shot at? Try some more tests.

[colinsouthern]

The Canon EF 300mm F4 IS is pretty much the bench mark for sharpness. Your image looks OK (as far as can be seen with a 515 pixel down-sample). Biggest issue I can see is it doesn't appear to be capture sharpened - see if you think this is any better?<div></div>

[beauh44]

Looks OK to me - I'm not sure what aperture you used or where the active focus point was. In general, a prime lens like the 300mm will be a tad sharper than most zooms. I love my 70-200 f/2.8L IS though.

[elliot1]

If you can, try another sample to see how it compares to your lens. If you don't have access to another lens, you can send it to Canon to have it checked. If it is brand new, exchange it for another.

[John Crowe]

The 70-200/4 L is said to be sharper than the 70-200/2.8 L and the 70-200/2.8 L IS will be even softer with the extra elements involved. The 300/4 L IS will be considerably sharper than any of the zooms!

 

 

I suggest taking a photo with your 70-200/2.8 and 300/4 to compare them properly. Take a photo of an identical scene adjusting your shooting distance with each lens to match the subject size in the scene as close as possible. ie. frame the scene the same. Shoot the 300 at f4 and f5.6 and shoot the zoom at f2.8, f4, and f5.6. Then view your images again at identical size on your monitor and see what you think. Ideally you will do these in conditions that do not need IS and will be shot at 1/750 at least if possible. A tripod with IS off in both cases would be the best test for lens resolution only. Post some here if you wish another opinion.

 

 

Try the following website for some comparisons, he used to have more with the 70-200/2.8 IS, but at least you will get an idea of what clarity should be acceptable and what is not.

 

 

www.pbase.com/drip/lens_tests

[John Crowe]

I forgot to mention that you must focus on the same subject in the scene, using manual focus if necessary. Also the reason I listed the apertures to use is so that you can compare the lenses at the apertures that they share, f4 and f5.6, and you will see how each lens gets sharper as you stop down. Also you will be able to compare each lens wideopen, f2.8 and f4 respectively.

[bob_osullivan]

Jim:

 

There is some "copy varrience" as we call it with these lenses. One person can get one that's razor sharp and another that's kind of soft and both could be within "factory spec". Canon is good about testing and recalibrating. Buy you may not need that.

 

First, you don't metion what aperture you used. If it was F2.8 then you should expect it to be softer than F4. This was shot in program mode, so did you select an AF point or let the camera decide on focus? It almost looks like the pole and birds feet may have been the point of focus with the rest of him being thrown out of focus a bit by lack of DOF, however it's a small target so maybe not.

 

Do a search on "70-200 soft copy" here and you'll get lot's of good information.

 

I'd spend some time getting to know the lens and checking it out before exhcanging it. Who knows you could get a worse copy. Make sure it's really bad before doing that. Also, Canon will recalibrate as I sated earlier.

 

I don't like test charts for front/back focus testign on zooms, because the starting and ending point of DOF changes as focal length changes. This can make you think a good lens is defective.

 

Intead, I like to do more real world testing but controlled. Put a teddy bear in good window light. Put a medicin bottle or business card on him. Shoot that with IS off, on a tripod, mirror lock up on and cable release. Compare you lenses doing this test. Then PRIINT don't pixel peep. An 8x10 print will tell you more. You'll never print so large that you can see individual pixel flaws anyway.

 

IS can actaully blur the image if it's not needed to get the shot ie; lot's of good light and fast shutter, so turn it off for testing.

 

If you can get just one good sharp shot this way, it indicates it was operator error and you just need to get used to the new lens.

 

Hope that helps.

[jimcrotty]

Thanks so much for the very helpful information. The shot I posted of the raven was done

at f6.3 and 1/1000 at ISO 400. Auto focus was on, so I do think the point of focus shifted

a bit downward.

 

I am going to do another test shot today using the recommendations noted by John and

Bob. I will post the result later today, but I will also do a test print.

 

After reading the comments and suggestions I'm feeling a little bit better about the

purchase. I think there's always a bit of anxiety when making a big purchase such as this,

so I just want to make sure I've invested wisely. I've never had a problem with any of my

other Canon lenses.

[alan_myers]

Hi,

 

If you are new to IS, there's a common mistake people make.

 

Many who are new to IS shoot too fast, not giving the IS enough time to completely engage.

 

Half-press your shutter release button and give IS a moment to wind up and become fully effective. Under ideal conditions, it might take a half second, or it may take 2 or 3 seconds (compare to big steady mounts used with movie cameras, that can take 5 or 10 *minutes* to wind up and become fully effective!).

 

If shooting a series of shots in rapidly changing situations, keep the shutter release button half pressed in between shots to keep the IS active and up to speed. That way you'll be ready for the next shot.

[steve_grant1]

Jim,

 

I'm new to the lens as well. I think the replies here are great and helpful. Definitely give

the tripod a shot to first test the lens. And, if you need more samples, I linked to a few I

shot in this thread I started:

 

http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=00No3V

 

Enjoy

[jimcrotty]

Colin - yes, the sharpening you did on the raven definitely improves the image.

 

I'm starting to feel better about the lens, especially since today I did a test shot using the

recommended settings.

 

With the shot below I used a tripod w/Kirk head, IS off, manual focus, ISO 50, manual at f2.8,

shutter at 1/200, mirror lock-up. My point of focus was the window frame. Let me know

what you think.<div></div>

[colinsouthern]

Hi Jim,

 

Again, I think it's just your post-processing that's letting you down - what do you think of this?

 

Cheers,

 

Colin :)<div></div>

[jimcrotty]

Hi Colin - yes, that does look better. I wanted to send as close as possible of an un-edited

image, which is a JPEG for web export from the raw file in Aperture. I didn't even put it

through Photoshop. What are your preferred settings for sharpening ?

[jimcrotty]

Wanted to see if I could duplicate what Colin did, using the same low-res JPEG version<div></div>

[colinsouthern]

Hi Jim,

 

I think I still win :)

 

Keep in mind that with a down-sampled image like yours the sharpening workflow will be completely different. There's 3 phases to good sharpening (Capture, Content/creative, and Output). For a full resolution RAW image I'd typically use something like 300% / 0.3 / 0 on an USM for capture sharpening, but on a down-sampled image like yours that'll over-sharpen the high frequency components (grass etc). so I applied 300% / 0.2 / 0, (very small change to the numbers, but BIG difference to the result) - but I did this 4 times. Even then, the grass was starting to look a bit "frosty" (please excuse the pun), so I used the history brush to regress just the grass by 2 levels (so grass is only sharpened twice, not 4 times).

 

At that point I changed the mode to LAB - added a curves layer - boosted the midtones (ie "grab the middle of the L curve and drag it out a bit). On the A & B curves I brought the end points in about 10% each - for material with subtle colours (like the timbre of the cabin) it has the effect of raising saturation, but also driving subtle colours apart.

 

With that done I flattened the image - flicked it back to RGB mode - gave it a touch of content sharpening (20% / 20 / 0 USM), and job done :)

 

Again I should mention though that I HATE working on low-res JPEGS - you (can) get big swings with small sharpens - and it's a bit of a lottery when you post the final result back.

 

If you want, send me the original and I'll give it a professional going-over for you :)

 

With regards to lens sharpness, it's all relative - a 300mm/F4 is always going to beat the pants off a 70-200/2.8L - BUT - the difference between sharpened and unsharpened is much bigger again - and yet only 5% photographers can sharpen an image properly (unfortunately, 95% think that they belong to the 5%).

 

If you're keen, pop onto the likes of www.amazon.com and pickup a copy of Bruce Frasers "Real World Image Sharpening with CS2" - it'll make your head spin for a couple of weeks, but you'll pop out the other end with enough knowledge to start learning how to do it properly.

 

Hope this helps,

 

Cheers,

 

Colin

[jimcrotty]

Thanks Colin. Lots of great info. Much I did not know. I'm going to look into the Bruce Fraser

book. I will see if I can email you a psd file of the image. Thanks.

[colinsouthern]

You're welcome Jim :)

 

If you're feeling rich, there are 3 books by Bruce Fraser that did more to lift my game than all the rest put together):

 

* Real World Image Sharpening

 

* Real World Camera RAW (CS2 version - CS3 version with Jeff Schewe and Bruce now out) (Bruce died at the beginning of the year)

 

* Real World Colour Management

 

Cheers,

 

Colin

[doe_koe]

Colin,

 

How did you sharpen the Raven? What steps were taken (I am assuming in CS2/CS3)?

 

Thanks.

[jimcrotty]

I just got my 70-200mm f2.8 L IS back from the Canon factory service center in NJ. I was

right - it needed adjustment. The service detail report states "replaced lens ass'y EMD/IS and

collar part, checked all, adjusted focus, cleaned to factory specs. Hats off to Canon - they

were quick and made the process easy.

 

The following are some of the shots I took here in my studio with the corrected lens:<div></div>

[jimcrotty]

another test shot from today -<div></div>

[jimcrotty]

and one more. This Eagle print is on canvas<div></div>