Software sharpening for lens tests

[Kent Shafer]

<p>I want to do some tests comparing the sharpness of various lenses on a D700. The purpose isn't to convince anyone else of anything—just to help me better understand the strengths and weaknesses of the various lenses I happen to own.</p>

<p>When evaluating the results, what's the best practice regarding software sharpening? (I use CS4 and ACR and sometimes PK Sharpener.) It seems to me the three likely possibilities are:</p>

 

<ol>

<li>No sharpening at all.</li>

<li>Default capture sharpening in ACR and nothing else.</li>

<li>Regular sharpening workflow (capture and output), the same as I would use in actual photography.</li>

</ol>

<p>Thanks for your thoughts.</p>

[phototransformations]

<p>If I were conducting such a set of tests, I'd be most interested in (1) and (3); that is, I'd want to know in absolute terms how sharp my lenses were at various apertures and focal lengths, and then I'd want to know how much I could compensate for unsharpness in software or by stopping down. That way I would know how to trade off convenience for raw IQ. Often, using a less-sharp lens that is lighter or more flexible is worthwhile if in the end the images you create (with a little more stopping down, or some post processing, or both) are virtually identical to those you'd get in the same shoot from a couple of sharper but heavier lenses.</p>

[walterh]

<p>But where is the problem? You can change the settings (and try all three alternatives) any time assuming that you shoot raw.<br>

You can do this even for a batch of all files in one go in ACR.<br>

I personally first look at all files with sharpening set at zero.<br>

In a second go I see how much sharpening I need to get similar results with the various lenses.</p>

[elliot1]

<p>I believe as long as you apply the same settings to each comparison shot, the results should be consistent regardless of the method you use.</p>

<p>To complicate things a bit, there are programs that offer lens specific corrections like Bibble and DXO that can have a dramatic effect on the results, narrowing differences from one lens to another..</p>

<p> </p>

[ellis_vener_photography]

<p>ACR's sharpener is a capture sharpening settings based on the camera / sensor / ISO setting combination. I would not use ACR's defaults as they are based on the Canon EOS-1Ds Mark III. I would find a formula that works for your camera and stick to it for that ISO. You might want to start with 75 / 0.8 / 25 / 0. (Amount/Radius/Detail/Masking) and tweak it from there but my suggestion is based on the latest versions of ACR (5.6 and Lightroom 2.6; Lightoom 3 beta is different).</p>

<p>ACR 5.x (the version of ACR used in Photoshop Cs4 is essentially the PhotoKit Capture Sharpener module (which was built around the EOS-1Ds Mark 2) but with user more control.</p>

<p> </p>

[bo_kral]

<p>I just played with 20-35mm not too sharp lens. I was comparing it to 14-24. I used portrait on D700 that has low sharpening. Once I found out I can match the prints (prints - not LCD) with about +2 - +3 NX2 sharpening increase I was done with sharpness testing. I would say start low to see the problems but then step into real world and find difference in needed sharpening compared to lens you like.</p>

[Kent Shafer]

<p>Thank you all for the good suggestions so far. Ellis, that's very interesting about ACR's defaults and PhotoKit Sharpener being built around Canon cameras. I hadn't heard that before. I don't have Lightroom. The latest version of ACR for CS4 is 5.5, but I will try your suggested starting point anyway. (The ACR 5.5 default is 25/1.0/25/0.)</p>

<p>I don't feel I have the experience and judgment yet to adjust sharpening by eye so I like the PhotoKit Sharpener cookbook approach. Jeff Schewe's new "Real World Image Sharpening" book is on my Christmas list. So much to learn!</p>

[oskar_ojala]

<p>It's important to have exactly the same settings for each image, otherwise it's not a too big issue.</p>