Is There a Standard for Lens "Sharpness"?

[mrgtrx850r44]

I just bought a new lens (Canon 100-400 L IS USM). Before my deadline to exchange it expires, I wanted to confirm that it is a sharp copy (as I have seen mentioned many times on this forum). Is there some test I can perform for myself. In my opinion it is sharp, but when I select the "view full sized image" command (Canon ZoomBrowser EX) I see that the image is fuzzy. Is this normal?

I'm not an expert in lenses so I would appreciate any guidance you could provide. I don't want to keep this lens if it is not a sharp focus. I have a 24-70L and don't see any issued with that lens on this same camera.

Is there anything else I should photograph to make the anaylysis better? I'm far from a professional photographer so any guidance would be appreciated.

Here is a sample photo taken with the lens on a tripod. ISO 100; Canon 40D; no flash.
...Thanks in advance

[johnastovall]

MTF charts are the only objective standard by which sharpness between different types of lenses can be compared. Here is a tutorial on how to read an MTF chart.

If it's a Canon and underwarranty, you can send it in an they will make sure it meets their standards.

[Richard DeYoung]

It does not appear to be as sharp as it could be. However I am used to prime lens and they seem to be always sharper.

[crazeazn]

is that a 100% crop?>

[Jim Victory]

Honestly there is no way you can tell if this lens is sharp by the example you have provided.

You need to post an unsharpened 100% crop shot wide open. I would suggest taking a shot of a Dollar bill if you don't have a resolution chart to shoot.

I personally rather test my lenses on what I will be shooting. I just got a 100-400 the other day and went out to a local park and shot ducks with it wide open at 400mm and then viewed them in PS. Thats all I needed to convince me it was a good copy.

If I'm comparing various lenses or bodies I usually use charts with normal testing methods.

Here is a 100% crop unsharpened @ f/5.6 and 400mm. This is what you need to evaluate the sharpness.

[canoflan]

I would argue that sharpness first of all is relative. I think you will find an average consensus out there as to what looks sharp, but take someone that uses Canon L prime lenses most of the time; they will think the 24-70 is blurry!

It sounds like you have a program where you can zoom in to 100% so you have that covered.

I was testing a lens last night for sharpness and here is what I did. I put on my 580EX flash and set the shutter to 1/125 to ensure no issue with shutter speed. I set the camera file mode to RAW (jpeg usually has sharpening settings built in and this distorts what the lens will provide natively in terms of sharpness). I usually will use a white/off white piece of paper with writing on it since the grain of the paper, the fibers, the ink of the pen, etc... provides a good way to tell if you have a sharp lens since a sharp lens typically in RAW mode will be able to show you the paper fiber, ink absorption by the paper (pen ink on paper isn't perfectly smooth and the lens should show the jagged edges of the pen marks), and other details. You want to ensure you get a good exposure so that that isn't an issue while you are viewing. I always place the camera on a tripod, or table to ensure absolutely no camera shake. I always use the center focus point (you have to selectively choose the middle focus point, else you cannot guarantee exactly where the camera chose to focus).

If your lens has IS, turn it off (I know some say it doesn't make a difference, but I believe if you don't need it, turn if off).

Take a few shots (always take several). Sometimes I manually adjust focus to make the image blurry, then let the autofocus refocus since sometimes if the focus is barely off (and this will make your photo look blurry), it will not refocus if you haven't moved your camera.

Evalute these photos at 100%. If you cannot see some semblance of the paper fibers and the jagged edges of the letters written with the pen, then you probably don't have a very sharp copy.

Also, you have to usually stop down 1 or 2 stops to get to the "sweet spot" of your lens's sharpness abilities; it is just the way they are made.

Try this out and you will find a way that works well for you to test your lenses for sharpness.

Let us know if you need more help or PM me if you have questions for me.

Pat

[Lghtning4u]

Maybe this will help out!

http://photo.net/learn/focustest/

[johnastovall]

The presception of sharpness may be relative but the actual 'sharpness' of a given lens and/or design can be measured and quantified and it has nothing to do with pixel peeping at 100%.


DigitUnderstanding image sharpness part 1:

[Jim Victory]

Come on John who generally has access to equipment to test MTF of a given lens?

Sure you can find this information for a type of lens but it really doesn't have any validity on how sharp your individual copy maybe. In fact I have seen different MTF ratings for this lens from different pro testers.

There is copy variation in many lenses and especially the 100-400 the OP is interested in gaining info about. The only way to determine if it is sharp enough for your purposes is to test it yourself. The final output is what is important not some MTF rating. That information is only good for determining the relative sharpmess of a given lens design, ie. 100-400, or a reference of performance capabilities and not a particular copy which may have a completely different performance level than a given MTF figure.

[Ruben Dario Cruz]

my suggestion is to send it to Canon for a bench test on a Collimator. While MTF charts are well regarded there is no standard benchmark in the optical industry for MTF curves. I know this working for an optical manufacturer.

[zeroendless]

Jim is right on the money, i'm quite familiar with 1-4 copy variation in sharpness as well. Post a 100% crop with no sharpening applied, preferably feather with good light.

When convert with photoshop (raw file) remember to turn sharpening all the way to 0, crop 100% about 800 pixel in width, post it again without down-sizing.

Also, 1-4 IS mechanism knows to turn bad in time. When that happens, you'll see the focused image with IS engaged make a sudden shift via viewfinder, resulted in rather soft image. I have Canon replaced the IS before sold mine. Time to let go of that dirt-sucker :)

[canoflan]

Originally Posted by johnastovall

The presception of sharpness may be relative but the actual 'sharpness' of a given lens and/or design can be measured and quantified and it has nothing to do with pixel peeping at 100%.


DigitUnderstanding image sharpness part 1:

I would argue that real sharpness of a lens can only be understood when comparing to a sharper lens of the same kind. Due to the fact that we cannot ever see too much detail, no lens is too sharp unless you are seeing too much detail for your purposes. Again, back to perception and relative use.

There is no disputing this.

Man, I love this stuff.
Pat

P.S. In conclusion, posting pictures on the forum and asking whether they are sharp will only generate more opinions and then result in a bell curve of sorts with the top part of the bell ending up being where most people "perceive" the image lies in the relative sharpness continuum.

[Pascal]

If the lens is new and under warranty, you really don't have much to worry about. Canon does a good job with lenses under warranty.

Keep in mind, it takes a while to get used to the 100-400L. The lens is heavy and the 2 stage IS, while effective, does require effort on your part to get used to the lens.

Place the lens on a sturdy tripod. Make sure IS is turned off. Focus on a contrasty subject and shoot with 10 or 12 second timer with mirror lockup. Take the fisrt shot at wide open and another at f5.6 and maybe f8.0. The only thing that should change is DOF (area in focus). Your focus point should remain the same. If you see that in your out of camera JPG's, chances are you have a good lens.

Just an FYI: All 3 Canon push-pull zooms can fall out of whack easily - a little bump is all it takes. It's just the way they are I guess. I have owned all 3 and loved the 35-350L, 28-300L, and 100-400L for their color and exceptional image quality (35-350L wasn't as sharp) - especially for a zoom. Also, best to send it in before the warranty expires for a CLA as these zooms cost a bit mroe than your usual zooms to calibrate or repair if something goes wrong.

[mrgtrx850r44]

I was not expecting such a large response. I am truly grateful. You have given me several options to try. I realize now that there may not be any specific threshold between what is sharp vs what is not. However, at least I have some form of anlaysis to try and then evaluate...it is certainly much better than what I started with (just a photo).

To answer some the question below:
The photo below is not 100% crop. It is right out of the camera. I didn't now better at the time.

FYI...I did shoot some of the photos using manual focus with 40D at full optical zoom...no difference noted from auto focus versions.

I will try one or more of the suggestions below and hopefully be able to form an opinion. I'll revert back here if questions.

Appreciate everyone's efforts to educate me on lense sharpness.

Thanks...

[mrgtrx850r44]

I see that Jim Victory below posted a photo with a 100-400....I'll assume this is a Canon lens similar to the one I bought. To be perfectly honest with you, this is the type of sharpness I am looking for. I'll shoot a few more shots during the afternoon of squirrel or birds. If they come out as crisp as your photo, then I'm really happy and will keep the lens.

Request a little guidance...When I crop 100% is it the same thing as clicking on the link "show image actual size" in ZoomBrowser EX? I realize that saving the image this way for uploading requires other action...I'm just trying to ensure I will be making valid comparisons since of the 10 or so photos I took already with this lens, when I click "show image actual size"...the images don't look "crisp" like the duck shown below (appreciate the image by the way).

What prompted all this analysis was that when I received the lens, it appeared to have been "rewrapped"...all the plastic bags were twisted at the ends like you would do to a loaf of bread (my other 2 Canon lenses were simply wrapped closed on the ends). It is possible the vendor simply wanted to inspect the lens before shipping or perhaps Canon changed how they wrap the ends. However, when I took the first few photos with it...I was not happy when I clicked "show image actual size". Hence the question below regarding whether this method "show image actual size) was valid or not.