new lens not sharp

[wommby]

Hi Guys,

I have just purchased a Nikon 17-35mm f2.8mm lens. I have been testing it for sharpness using the hyperfocal focusing and it doesn't appear to be as sharp as my old 24-85mm lens.

Take at look at these photos. There are 4 photos. 2 @ f/16 at 24mm and 2 @ f22 at 24mm. 1 taken with the old lens of each and 1 taken with the new lens. I placed a marker on the ground at 2ft and again at 4ft. I set the hyperfocal distance the same for both f16 shots and the same for both f22 shots.

The second photo and fourth photo are with the new lens and I don't think they are as sharp.

What do you think? Would really appreciate any feedback, even if you think I am doing something wrong - please let me know.

Deb

[Colin Southern]

Hi Deb,

I'd suggest not testing the way you are ...

1. Using the hyperfocal distance doesn't guarantee the sharpest result (I won't go into the technical details of why!)

2. At F16/F22 Diffraction will be entering into the capture

3. I suspect that you're not applying capture sharpening

4. Most lenses are generally sharpest around 2 to 3 stops from wide open, and ...

5. We really need to see 100% crops to properly evaluate sharpness, not down-sampled images (like I suspect these are) (the down-sampling introduces issues of it's own).

What I'd suggest is ...

- Set the new lens to F5.6

- Use the centre AF point to focus on something of interest

- Use a tripod and remote release (or tripod & self-timer)

- Apply capture sharpening of around 300% @ 0.3 pixel

- Post a 100% crop of the image so that we're looking at absolute pixels, not something that's been down-sized.

Does this help?

[wommby]

Colin,

I will do as you suggest. How do I capture sharpen by 300% (don't really know what to do here).

Question: When you are taking landscape shots don't you need to take them at around f/11 or f/16 to ensure everything is in focus? If so, how will a test at f5.6 tell you if it is sharp enough for landscapes?

Sorry to sound stupid but I am really a learner when it comes to topside photography.
Deb

[RockNGoalStar]

Hi Deb,

I will do as you suggest. How do I capture sharpen by 300% (don't really know what to do here).

You need to do this in some sort of post-production software like Photoshop. It is not done in camera. All of Colin's other suggestions are ways to ensure that the image your camera records is as sharp as it can possibly be. After that you may also want to play with the image in PP software afterwards to get the best out of it. Including by sharpening it a touch.

Question: When you are taking landscape shots don't you need to take them at around f/11 or f/16 to ensure everything is in focus? If so, how will a test at f5.6 tell you if it is sharp enough for landscapes?

I think for the purposes of this test then take a photo of anything. It isn't necessary to take a wonderfully composed landscape scene in order to determine if there are any issues with your lens.

I tend to take a lot of my landscape shots at around f/11. As Colin pointed out, if you go smaller than that (f/22 for example) then you will start to get diffraction which will certainly affect the sharpness of your images.

Don't worry about 'sounding stupid', you don't sound stupid at all. And besides, the only stupid question is the unasked question

[wommby]

Hi Deb,



You need to do this in some sort of post-production software like Photoshop. It is not done in camera. All of Colin's other suggestions are ways to ensure that the image your camera records is as sharp as it can possibly be. After that you may also want to play with the image in PP software afterwards to get the best out of it. Including by sharpening it a touch.



I think for the purposes of this test then take a photo of anything. It isn't necessary to take a wonderfully composed landscape scene in order to determine if there are any issues with your lens.

I tend to take a lot of my landscape shots at around f/11. As Colin pointed out, if you go smaller than that (f/22 for example) then you will start to get diffraction which will certainly affect the sharpness of your images.

Don't worry about 'sounding stupid', you don't sound stupid at all. And besides, the only stupid question is the unasked question

Hi,

O.K. have taken note of all of this. Will have to take this photo tomorrow as it is now dark so will post tomorrow.

Thanks for all your help.
Will see what you all think tomorrow when I post another photo. Should I take a photo which includes a large range of DOF?

Deb

[Colin Southern]

Should I take a photo which includes a large range of DOF?

Hi Deb,

Not really - we're only testing sharpness here - so we really want to setup conditions where it'll be as sharp as possible.

Things like AF accuracy - diffraction - DoF are all seperate issues.

[wommby]

Hi Deb,

Not really - we're only testing sharpness here - so we really want to setup conditions where it'll be as sharp as possible.

Things like AF accuracy - diffraction - DoF are all seperate issues.

Colin,

Just another question - Am I taking this test photo in Auto Focus mode or Manual Focus?
Deb

[wommby]

Colin,

How do I do a 100% crop? I tried just changing the size to 700pixels X 477pixels but not sure if this is what you are meaning to do, will wait to here from you

[wommby]

Hi Deb,

I'd suggest not testing the way you are ...

1. Using the hyperfocal distance doesn't guarantee the sharpest result (I won't go into the technical details of why!)

2. At F16/F22 Diffraction will be entering into the capture

3. I suspect that you're not applying capture sharpening

4. Most lenses are generally sharpest around 2 to 3 stops from wide open, and ...

5. We really need to see 100% crops to properly evaluate sharpness, not down-sampled images (like I suspect these are) (the down-sampling introduces issues of it's own).

What I'd suggest is ...

- Set the new lens to F5.6

- Use the centre AF point to focus on something of interest

- Use a tripod and remote release (or tripod & self-timer)

- Apply capture sharpening of around 300% @ 0.3 pixel

- Post a 100% crop of the image so that we're looking at absolute pixels, not something that's been down-sized.

Does this help?

Colin,

here is a photo at f5.6, taken on a tripod, mirror up, remote release, exposure delay on, aperture priority. This is sharpened as suggested above. I think I am posting a 100% crop but please let me know if I have done this wrong.
Deb

[Colin Southern]

Hi Deb,

It looks good, although it's not a 100% crop.

To get a 100% crop you need to stay away from the resizing commands and just use the crop tool.

From what I can see though, the lens looks fine to me.

[wommby]

Hi Deb,

It looks good, although it's not a 100% crop.

To get a 100% crop you need to stay away from the resizing commands and just use the crop tool.

From what I can see though, the lens looks fine to me.

Collin

Have downloaded 4 more. 1 set is at f5.6 and 1 set is at f8. It really appears to me that the 17-35mm lens is not as sharp. These were all taken hand held in aperture priority just moments apart and from the same standing point. What do you think? I forgot to add. The ones taken with the 17-35mm are all taken at a slightly faster shutter speed. Must have been more sun at that particular time.

Sorry also, these are without sharpening. I know you should capture sharpen but there is a distinct difference in sharpness even before the sharpening process.

[Colin Southern]

Collin

Have downloaded 4 more. 1 set is at f5.6 and 1 set is at f8. It really appears to me that the 17-35mm lens is not as sharp. These were all taken hand held in aperture priority just moments apart and from the same standing point. What do you think? I forgot to add. The ones taken with the 17-35mm are all taken at a slightly faster shutter speed. Must have been more sun at that particular time.

Sorry also, these are without sharpening. I know you should capture sharpen but there is a distinct difference in sharpness even before the sharpening process.

Hi Debbbie,

The problem is - when you're doing back-to-back testing like this the results are inconclusive if you introduce more than one variable at once - and in this case you've hand-held them (so I can't say whether or not any camera shake entered in to either shot), and on top of that, they all look like low-quality jpegs which makes it even harder to evaluate them

I'm also still trying to work out if these are 100% crops, or if you've down-sized them image. If they're 100% crops then you should only be seeing a small part of the original scene that you captured -- to give you an example, the first image below is a 100% crop (be sure to click on it to look at it full size), whereas the 2nd image is a down-sampled shot of the whole scene that it was taken from ...

[wommby]

Colin,

Thanks for all your help. I would still like to know how I go about using this lens for hyperfocal focusing for landscapes to get everything sharp from front to back. I have set it up on a tripod, used aperture priority, mirror up, exposure delay on, used f16 and the hyperfocal distance of 2.03 feet for 17mm end of lens and I can't get it sharp. What do I do next? What am I doing wrong?

Here,I have had another attempt at downloading a 100% crop and included the original. Shot in RAW then converted to a jpeg file so it should be a high quality jpeg at 300ppi.

How do you get your images so big. Mine keep going in as thumbnails and when you click on them they are still small???

Here is one that may be better. This is taken with the same set up at above but at 24mm and hyperfocal distance at 4feet for f16.

[Black Pearl]

Stop using f16 and switch the camera to AF then actually point it at the thing you want sharp would be a good start.

[RockNGoalStar]

Debbie, can you take 2 photos, preferrably on a tripod, using the same settings with the two different lenses that you want to compare, and then upload the RAW files to www.mediafire.com and then post the links to them on here.

I'm sure that a few of the members will look at them and will have a better idea to determine whether there are any issues with the lens or not.

[wommby]

Debbie, can you take 2 photos, preferrably on a tripod, using the same settings with the two different lenses that you want to compare, and then upload the RAW files to www.mediafire.com and then post the links to them on here.

I'm sure that a few of the members will look at them and will have a better idea to determine whether there are any issues with the lens or not.

Thank you, I will try that. I will put it on a new thread so others may look. Hopefully I won't have any problems downloading

[wommby]

Robin, thanks for this advice but I am trying to get everything in focus as for landscapes. I am finding it hard to close focus on something and get everything else in focus which I believe you should be able to do with this lens.

[wommby]

Debbie, can you take 2 photos, preferrably on a tripod, using the same settings with the two different lenses that you want to compare, and then upload the RAW files to www.mediafire.com and then post the links to them on here.

I'm sure that a few of the members will look at them and will have a better idea to determine whether there are any issues with the lens or not.

Tommy,

Have taken two photos - one with each lens and have put them on the mediafire.com but have no idea how to get the onto here

[Colin Southern]

Hi Debbie,

I can see what you're trying to achieve here, but the theory doesn't always marry up with the practice. Hyperfocal calculations are all about maximising depth of field at any particular aperture - so one would assume that the narrower the aperture (ie "bigger F-Stop number), the greater the depth of field - which is true - but that doesn't necessarily mean that a greater depth of field will equate to greater image sharpness, because images will start to become diffraction limited past about (on average) F11. Whether the effects of diffraction are important is another story (personally, I don't worry about it, as it's not going to be resolveable in any normal online image that only presents around 3% of what was originally captured anyway).

So "sharpness testing" is really a bit different to "maximising DoF".

Sean touched on this in his Hyperfocal tutorial:

https://www.cambridgeincolour.com/tu...l-distance.htm

The problem with the hyperfocal distance is that objects in the far background (treated as ~infinity) are on the extreme outer edge of the depth of field. These objects therefore barely meet what is defined to be "acceptably sharp." This seriously compromises detail, considering that most people can see features 1/3 the size of those used by most lens manufacturers for their circle of confusion (see "[COLOR=#afbddb]Understanding Depth of Field[/COLOR]"). Sharpness at infinity is particularly important for those landscape images that are very background-heavy.

In terms of testing the "sharpness" in terms of "suitability of your new lens for landscape" - personally - I'd stick it on a tripod - enable MLU - attach a remote release (or use the self times) - use something like F11 - shoot RAW - focus on something about 1/3 into the scene (I know Sean advises against this old "rule of thumb", but I've found it adequate for my purposes) - APPLY CAPTURE SHARPENING - and lets see what you get. If you're after a pure sharpness test then I'd repeat the above with the lens stopped down about 3 stops from wide-open - and see how it resolves something that it's focused on (evaluating after the application of capture sharpening).

In summary,

- If you shoot at an aperture that's smaller that F11 (ie F16 or F22 etc) then you may increase your depth of field, but you'll decrease overall sharpness

- If you don't apply capture sharpening (300% @ 0.3 pixel) then you're not compensating for the softness introduced by the anti-aliasing filter and digitization & demosaicing processes

- If you rely on hyper focal distance, you're likely to have "marginal" sharpness with objects that are far away.

Hope this helps

[wommby]

Colin,

Just spent 30 minutes answering this and lost while trying to upload, so here goes again.

I understand what you are saying with hyper focal distance but I can't get everything into focus within a 20 metre distance let alone infinity.

I have taken so many shots this morning that I am confusing myself and don't even know what I am expecting to get now.

Here are 3 photos with 100% crops of each. All taken at f11, on a tripod, with self timer. The first one is in auto focus, the second one if Hyper Focal focus and the third on is auto focus but from further back. The last appears a lot better focus. How do I get acceptable focus while trying to include something in the foreground which is what I was trying to achieve in the first and second images.

Thanks again for all your help. Maybe I am just expecting too much and should just take landscapes in Auto Focus.
Deb