When is a lens sharp?

[Abitconfused]

Below is an image (just a snapshot) taken with my Sigma 24-35mm on a Nikon D810. Some reviews give the lens a 4 of 5 for sharpness but I think it deserves more than that. Not only is it versatile and fast at f/2 but the len's colour transmission is to die for.



I tweaked it in Photoshop, true...



But hay. it was RAW and any digital photo can use a little white point/black point tweak and some added brightness and sharpening.

So anyway I love this lens. Love, love,love! Whatcha think?

[AlwaysOnAuto]

I think it looks pretty good Ed.
Could you have gotten that sharpness with better focusing?
I don't have PP skills or sophisticated software so I always try for my best straight out of the camera.

[Abitconfused]

Probably could have done a better job focusing. I am too casual about point of focus. I use one focus point but, perhaps, a larger number toward the center, would be a better option. I like the diffused background but a little smaller aperture may have been better.

[IzzieK]

Interesting shot. Are you trying to focus on the flower or the fence? Since you are using a D810 you could have zoomed in a bit closer after the snapshot then show both so we can see the difference. When testing my lens/camera, same as yours, I used live view to get a closer look using the + button several times. Be very careful doing this as I fried something inside my camera when it got too hot.

[DanK]

Probably could have done a better job focusing. I am too casual about point of focus. I use one focus point but, perhaps, a larger number toward the center, would be a better option. I like the diffused background but a little smaller aperture may have been better.

Looks pretty sharp, but it isn't easy to tell at this resolution.

Re focusing: if the subject is static and you know where you want the focus point to be, single-point is the best option because it forces the camera to use the location you want. Multi-point of any sort allows the camera leeway to choose. that's very useful under other circumstances, but it wouldn't be an improvement in this case.

[DanK]

I think it looks pretty good Ed.
Could you have gotten that sharpness with better focusing?
I don't have PP skills or sophisticated software so I always try for my best straight out of the camera.

I agree, getting focus right in camera is really essential. However, just to be clear, if you are shooting jpeg, you are also using software to sharpen the image. You're just using the preset sharpening algorithm in the camera's firmware rather than doing it yourself with software. I don't know what you shoot, but I shoot Canon cameras, and the lcd will tell you the amount of sharpening (on an arbitrary scale that isn't really interpretable) for every picture style.

[joebranko]

I have always used single point focusing, but have recently started using multi point. For birds and animals I think single point is better; but for landscapes I don't know, so am experimenting...

[george013]

I have always used single point focusing, but have recently started using multi point. For birds and animals I think single point is better; but for landscapes I don't know, so am experimenting...

There's only ONE plane that's in focus. And you as photographer choices which plane.

I think Ed's post was meant to show how satisfied he is with his lens. It looks good to me as far I can see.

George

[chauncey]

Evaluating the image sharpness based on an internet display is surely an exercise in futility,
unless, that display is a 1000 pixel crop.

[xpatUSA]

Evaluating the image sharpness based on an internet display is surely an exercise in futility,
unless, that display is a 1000 pixel crop.

True indeed, William.

Expanding a little, I evaluate lens sharpness based on the raw data (extracted by RawDigger) transformed to a raw composite raster image with no conversion or processing at all. I shoot a slant-edge target and use QuickMTF to determine the edge response and the MTF.

Any other method is really not analyzing what the lens and sensor saw.

In another forum on another site far far away, all you usually get to illustrate sharpness are images of shrubbery or rolling hills, often processed to the max and sometimes even up-sampled! . . . . . followed by "look at the resolution of my xxx" . .

Rant over, sorry.

[AlwaysOnAuto]

I just did a very non-scientific study of a new-to-me old lens to get an idea of where it is most effective for the type of shooting I do (static cars mostly). I won't post the shots I took, but it looks like f5.6 is pretty nice re: DOF vs what is out of focus.
I think the original post is similar to what I'd look for in a lens.
Thanks for posting this thread Ed.

[Shadowman]

True indeed, William.

Expanding a little, I evaluate lens sharpness based on the raw data (extracted by RawDigger) transformed to a raw composite raster image with no conversion or processing at all. I shoot a slant-edge target and use QuickMTF to determine the edge response and the MTF.

Any other method is really not analyzing what the lens and sensor saw.

In another forum on another site far far away, all you usually get to illustrate sharpness are images of shrubbery or rolling hills, often processed to the max and sometimes even up-sampled! . . . . . followed by "look at the resolution of my xxx" . .

Rant over, sorry.

So after utilizing RawDigger what follows? Are you evaluating the lens, the camera, the camera settings? What follows next with the image captured, do you chuck it if your analysis fails? I'm always searching for the best in my gear and for me the final criteria would be a large print, wondering what is your criteria.

[Manfred M]

A sharp lens is a wonderful thing, but unless you are extremely careful in nailing your focus as well as shooting from a solid, sand-bagged, heavy duty tripod, the sharpness of the lens is going to be over-ridden by other technical issues.

Then of course, you are going to have to make a very large print that we can pixel peep, because the downsampling associated with posting an image on the internet means you cannot control the output device we are viewing your image on, so showing us an image is going to be meaningless. Your average computer screen is only capable of displaying a 2 MB image (roughly 2000 x 1000 pixels).

[Dave Humphries]

On the topic of using a number of focus points ...

I use one focus point but, perhaps, a larger number toward the center, would be a better option.

As George says, using more focus points will not, in itself, bring any more of the image in to focus.

In a complex scene such as this, the photographer should be choosing exactly what in the scene the focus is set on, doing this with a knowledge of Depth of Field (DoF) required and then choosing an aperture that achieves this (if possible).
The alternate; giving the camera the option to use more focus points largely (on a scene like this) takes the photographer out of the decision making process, which might work (by sheer luck), but most probably won't, certainly not with any consistency, leading to frustration.

I have always used single point focusing, but have recently started using multi point. For birds and animals I think single point is better; but for landscapes I don't know, so am experimenting...

There is one situation where I think multi point would be better for a single bird - when it is in flight against a featureless blue sky. That's because there's nothing else in frame to confuse the other focus points, they become useful to retain focus if your panning and framing wobbles a bit.

At all other times, I use a single focus point.

As for landscapes, as per my reply to Ed's comment (above), using multipoint just hands the decision on where to focus to the camera - which I really don't see being a good thing. Arguably it may be a moot point if the subject distances, focal length and aperture mean that everything they are on is in the range where hyperfocal distance wins out though.

HTH, Dave

[devicephoto]

Nice picture: colours, lights and shades.
But... there's only one "but". You're absolutely lucky getting this particular lens in good working quality. It's a SIGMA and you're gambling when buy it.
When was looking up for a wide angle lens first I tried Sigma 24mm (due to budget). And it had the worst AF I ever met: huge progressive front focus both on my S5Pro and on D800 from store (in-camera focus adjustment did not affect at all). Then tried Nikon AF-S 24mm 1.4G, of course it was significantly better but the price is almost twice bigger. Then just looked up to 2/28 Distagon from Zeiss. No regret of choosing it.
So you're lucky photographer enjoying good and proper working Sigma ) Congrats and more beautiful pictures!

[Tronhard]

For what it is worth... Regarding the comments on when a lens is sharp, the following web site, WHEN USED CAREFULLY is quite handy in showing lens performance on specific bodies. As they comment, and I have experienced, the same lens behaves a bit differently when placed on different bodies (and not limited to crop vs. full frame).

https://www.dxomark.com/

The following YouTube video is an interesting commentary on the use of the above site.

https://www.youtube.com/watch?v=MRzONYmjH3M

[rtbaum]

Nice picture: colours, lights and shades.
But... there's only one "but". You're absolutely lucky getting this particular lens in good working quality. It's a SIGMA and you're gambling when buy it.
When was looking up for a wide angle lens first I tried Sigma 24mm (due to budget). And it had the worst AF I ever met: huge progressive front focus both on my S5Pro and on D800 from store (in-camera focus adjustment did not affect at all). Then tried Nikon AF-S 24mm 1.4G, of course it was significantly better but the price is almost twice bigger. Then just looked up to 2/28 Distagon from Zeiss. No regret of choosing it.
So you're lucky photographer enjoying good and proper working Sigma ) Congrats and more beautiful pictures!

I will put my Sigma 50mm Art 1.4 up against your Zeiss Distagon any day of the week

[mknittle]

I have always used single point focusing, but have recently started using multi point. For birds and animals I think single point is better; but for landscapes I don't know, so am experimenting...

I seem to always revert back to single point.

[rpcrowe]

IMO shooting foliage is not a great way to test a lens or a great way to illustrate a lens sharpness. There are too many variables involved when doing this type of shot and too many places on which the camera can focus.

[xpatUSA]

Nice picture: colours, lights and shades.
But... there's only one "but". You're absolutely lucky getting this particular lens in good working quality. It's a SIGMA and you're gambling when buy it.

That was said quite often of their older models but their current Global Vision series is getting "rave reviews" for their Art, Contemporary and Sports models. Improved materials, improved performance and improved QA but still at competitive pricing.

I have a 17-70mm Contemporary model and, from experience, it is indeed a step up from their previous 17-70mm models.