Mad About Sharpness

[TN Args]

Several things:

• The 24-240 isn't the sharpest lens to begin with, so manage expectations

Among the 32 FE-mount lenses that DxO has tested, the Sony 24-240mm achieved by far the least sharp / lowest resolution, at least by the somewhat simplistic metric of DxO's overall 'sharpness' score. Go to

I would not give their sharpness score much merit they provide very good data summing it up in one number is a bit daft.

No doubt reducing any complex data set to a single number is prone to less useful or even misleading results. But (1) that single sharpness score is not totally meaningless, and that the Sony 24-240mm scored so much lower than any other FE lens they tested is somewhat remarkable. And (2) discussions about DxO data over the years have made me think that a lot of people like such single-number scores and most don't want to do what they really should do, which is drill down a couple of levels into the underlying results. Sometimes dumbing-down is marketing-driven; but at least DxO offers you the choice to look at either.

Other than the gold standard lensrentals blog who have access to skilled technicians, proper testing gear and best of all multiple samples of lenses. Alas they don't do much reviews :-( Most of the "test" sites have their problems . I like Dustin Abbot { blog and YT } and Christoper Frost { YT } who give a decent idea of performance . Though of course based on one sample

Oh there are all sorts of issues with testing, and not having (1) a reasonable sample size (one lens certainly doesn't do it) that have been (2) obtained through normal commercial channels (i.e. not possibly cherry-picked by the manufacturer or regional distributor) are certainly real issues, as are the facilities for proper tests. What can I say other than: the more data of at least decent quality, the better.

Their sharpness test data, field map is I feel a good indicator of performance

Yes, to a degree. The problem is, IIRC, DxO's maximum-green field map coloration represents the maximum that a 12 MP sensor--yes, only a twelve megapixel sensor--can really benefit from. On one hand, with a maximum-quality 12 MP,

I did not know that , I had assumed that selecting a given camera gave results for that sensor MP count . You know what they say about assuming

Well I have looked, and alas the answer is more complicated than I recalled.

Thank you for self-checking and saving me the trouble of explaining the error in your field map comment above.

DxO's 'perceptual megapixels' scale tops out at 12 P-MP, above which is the maximum score:

However, the "Field map" we were discussing seems to be something somewhat different:

I have not yet determined what "Acutance map for use case", expressed as a percent, means. Does that mean it's related to the resolution of the sensor on which it's tested? Apparently so. Or I should say, obviously so putting the same lens in front of an APS-C versus a FF sensor. Beyond that maybe. I checked the Sony 24-240mm, and pulled up 240mm @ f/6.3 for the A7R (original) versus the A7R IV. The plot is clearly more favorable for the A7R IV than for the A7R. I guess that is showing that even a mediocre lens benefits a little from a higher-resolution sensor. But beyond that, I'm not sure precisely what the plot is showing.

The percentage scale is the lens acutance (in perceptual MP) as a percentage of the sensor MP.

Hence, as you say in your correction above, a higher resolution camera of the same sensor size will result in a different and ‘less green' field map for a given lens.

cheers

 

[NAwlins Contrarian]

Their sharpness test data, field map is I feel a good indicator of performance

Yes, to a degree. The problem is, IIRC, DxO's maximum-green field map coloration represents the maximum that a 12 MP sensor--yes, only a twelve megapixel sensor--can really benefit from. On one hand, with a maximum-quality 12 MP,

I did not know that , I had assumed that selecting a given camera gave results for that sensor MP count . You know what they say about assuming

Well I have looked, and alas the answer is more complicated than I recalled.

Thank you for self-checking and saving me the trouble of explaining the error in your field map comment above.

DxO's 'perceptual megapixels' scale tops out at 12 P-MP, above which is the maximum score:

However, the "Field map" we were discussing seems to be something somewhat different:

I have not yet determined what "Acutance map for use case", expressed as a percent, means. Does that mean it's related to the resolution of the sensor on which it's tested? Apparently so. Or I should say, obviously so putting the same lens in front of an APS-C versus a FF sensor. Beyond that maybe. I checked the Sony 24-240mm, and pulled up 240mm @ f/6.3 for the A7R (original) versus the A7R IV. The plot is clearly more favorable for the A7R IV than for the A7R. I guess that is showing that even a mediocre lens benefits a little from a higher-resolution sensor. But beyond that, I'm not sure precisely what the plot is showing.

The percentage scale is the lens acutance (in perceptual MP) as a percentage of the sensor MP.

Hence, as you say in your correction above, a higher resolution camera of the same sensor size will result in a different and ‘less green' field map for a given lens.

An actual DxO field map comparison shows the opposite: the same lens at the same focal length and aperture shows more green with a higher-resolution sensor. Here is the Sony 24-240mm at issue, at 240mm and f/6.3, on the NEX-7 (APS-C, 24 MP), A7R (36 MP, FF), and A7R IV (60 MP, FF):

As you can see the green area is larger and greener with the A7R IV than with the A7R. So my understanding was wrong, and your understanding is wrong.

I have not worked out in my own mind what this means. However, I think that the A7R-to-A7R IV comparison shows that a higher-resolution sensor provides some sharpness benefit even with a lower-performing lens. I think it also shows that DxO sharpness rating values, I guess on a total-captured-detail basis, capturing a full frame work versus capturing just the APS-C size region of it at about the same pixel pitch.

In short, I think the DxO sharpness field map provides useful information, but maybe you have to be careful about apples-to-apples comparisons, or something. I'm still trying to understand it all.

 

[TN Args]

Their sharpness test data, field map is I feel a good indicator of performance

Yes, to a degree. The problem is, IIRC, DxO's maximum-green field map coloration represents the maximum that a 12 MP sensor--yes, only a twelve megapixel sensor--can really benefit from. On one hand, with a maximum-quality 12 MP,

I did not know that , I had assumed that selecting a given camera gave results for that sensor MP count . You know what they say about assuming

Well I have looked, and alas the answer is more complicated than I recalled.

Thank you for self-checking and saving me the trouble of explaining the error in your field map comment above.

DxO's 'perceptual megapixels' scale tops out at 12 P-MP, above which is the maximum score:

However, the "Field map" we were discussing seems to be something somewhat different:

I have not yet determined what "Acutance map for use case", expressed as a percent, means. Does that mean it's related to the resolution of the sensor on which it's tested? Apparently so. Or I should say, obviously so putting the same lens in front of an APS-C versus a FF sensor. Beyond that maybe. I checked the Sony 24-240mm, and pulled up 240mm @ f/6.3 for the A7R (original) versus the A7R IV. The plot is clearly more favorable for the A7R IV than for the A7R. I guess that is showing that even a mediocre lens benefits a little from a higher-resolution sensor. But beyond that, I'm not sure precisely what the plot is showing.

The percentage scale is the lens acutance (in perceptual MP) as a percentage of the sensor MP.

Hence, as you say in your correction above, a higher resolution camera of the same sensor size will result in a different and ‘less green' field map for a given lens.

Sorry, I meant ‘more green’, as should be obvious since I said I was agreeing with your finding.

An actual DxO field map comparison shows the opposite: the same lens at the same focal length and aperture shows more green with a higher-resolution sensor. Here is the Sony 24-240mm at issue, at 240mm and f/6.3, on the NEX-7 (APS-C, 24 MP), A7R (36 MP, FF), and A7R IV (60 MP, FF):

As you can see the green area is larger and greener with the A7R IV than with the A7R. So my understanding was wrong, and your understanding is wrong.

I said same sensor size, so comparing your left-most figure and centre figure is irrelevant as a check of what I wrote.

I have not worked out in my own mind what this means. However, I think that the A7R-to-A7R IV comparison shows that a higher-resolution sensor provides some sharpness benefit even with a lower-performing lens.

I have always been completely clear that total resolution of a camera-lens pairing goes up with a higher resolution sensor of same size. I fully expect (and predicted in my prior post) that the one on the right will have the higher resolution compared to the one on the centre. The only thing that isn’t completely intuitive is that the high-res camera on the right is so much sharper that it is even a higher percentage of 60 MP than the centre pairing is as a percentage of 36 MP.

I think it also shows that DxO sharpness rating values, I guess on a total-captured-detail basis, capturing a full frame work versus capturing just the APS-C size region of it at about the same pixel pitch.

In short, I think the DxO sharpness field map provides useful information, but maybe you have to be careful about apples-to-apples comparisons, or something. I'm still trying to understand it all.

Well, it’s pretty rare for someone to buy a lens and then look at which body to get for it, as per the 3 pairings discussed above.

More usual is to own the camera and compare lenses for use on it….and used that way it is very easy and intuitive to apply the sharpness field map.

 

[NAwlins Contrarian]

Their sharpness test data, field map is I feel a good indicator of performance

Yes, to a degree. The problem is, IIRC, DxO's maximum-green field map coloration represents the maximum that a 12 MP sensor--yes, only a twelve megapixel sensor--can really benefit from. On one hand, with a maximum-quality 12 MP,

I did not know that , I had assumed that selecting a given camera gave results for that sensor MP count . You know what they say about assuming

Well I have looked, and alas the answer is more complicated than I recalled.

Thank you for self-checking and saving me the trouble of explaining the error in your field map comment above.

DxO's 'perceptual megapixels' scale tops out at 12 P-MP, above which is the maximum score:

However, the "Field map" we were discussing seems to be something somewhat different:

I have not yet determined what "Acutance map for use case", expressed as a percent, means. Does that mean it's related to the resolution of the sensor on which it's tested? Apparently so. Or I should say, obviously so putting the same lens in front of an APS-C versus a FF sensor. Beyond that maybe. I checked the Sony 24-240mm, and pulled up 240mm @ f/6.3 for the A7R (original) versus the A7R IV. The plot is clearly more favorable for the A7R IV than for the A7R. I guess that is showing that even a mediocre lens benefits a little from a higher-resolution sensor. But beyond that, I'm not sure precisely what the plot is showing.

The percentage scale is the lens acutance (in perceptual MP) as a percentage of the sensor MP.

Hence, as you say in your correction above, a higher resolution camera of the same sensor size will result in a different and ‘less green' field map for a given lens.

Sorry, I meant ‘more green’, as should be obvious since I said I was agreeing with your finding.

Ah, well that flips the result, doesn't it!

An actual DxO field map comparison shows the opposite: the same lens at the same focal length and aperture shows more green with a higher-resolution sensor. Here is the Sony 24-240mm at issue, at 240mm and f/6.3, on the NEX-7 (APS-C, 24 MP), A7R (36 MP, FF), and A7R IV (60 MP, FF):

As you can see the green area is larger and greener with the A7R IV than with the A7R. So my understanding was wrong, and your understanding is wrong.

I said same sensor size, so comparing your left-most figure and centre figure is irrelevant as a check of what I wrote.

The idea was to see what DxO is or is not telling us. My reason for including the NEX-7 is that, among the APS-C models tested, it has a pixel pitch (266 p/mm) very close to that of the A7R IV (255 p/mm). So if all this DxO plot were showing us is lens resolution relative to the underlying sensor pixels, then the plot for the NEX-7 should be very similar to a crop from the center of the A7R IV plot, i.e., overall quite green.
I'm still trying to get a better sense of what DxO is doing. One day I need to take some time and do a deep dive. Alas, not tonight.

 

[AlesSor]

Hi Everyone,



yrsterday I was able to shoot again at my kid using 1/2000 (1.000 iso) as shutter time and this is one of many pictures I took. They are all similar in sharpness...

This is the pic

And this is the 100% crop

I am quite sure I was the most steady I could be, when shooting.

There was a speak about if it was the lens to blame or if it was a motion blur issue. Afetr this new information, what's the answer? I start thinking that it's a lens sharpness issue...

 

[dazzerfong]

Hi Everyone,

yrsterday I was able to shoot again at my kid using 1/2000 (1.000 iso) as shutter time and this is one of many pictures I took. They are all similar in sharpness...

This is the pic

And this is the 100% crop

I am quite sure I was the most steady I could be, when shooting.
There was a speak about if it was the lens to blame or if it was a motion blur issue. Afetr this new information, what's the answer? I start thinking that it's a lens sharpness issue...

Test your lens indoors on a tripod, that's the only way to know if it's your lens or not. I personally think it's heat miraging from shooting over artificial turf on a humid and sunny day.

 

[DCSteve]

I am quite sure I was the most steady I could be, when shooting.
There was a speak about if it was the lens to blame or if it was a motion blur issue. Afetr this new information, what's the answer? I start thinking that it's a lens sharpness issue...

Sorry, but look at the net behind the boy. Looks like camera shake. Might be heat waves but I don't know much about that.

Steve

 

[4Photos]

Probably someone already mentioned this, but too tired to read through all the responses.

There is a a setting somewhere in the menus about RELEASE PRIORITY. Do you want the camera to take the photo as soon as you press the shutter button? Or do you prefer to let it finish its AF function, which means you are guaranteed sharpness (as much as the lens allows) but may miss the action by a fraction of a second.

Unless your subject is super fast moving, I would change this setting to AF priority, especially before purchasing a better, more expensive lens.

 

[NAwlins Contrarian]

yrsterday I was able to shoot again at my kid using 1/2000 (1.000 iso) as shutter time and this is one of many pictures I took. They are all similar in sharpness...

This is the pic

And this is the 100% crop

I am quite sure I was the most steady I could be, when shooting.
There was a speak about if it was the lens to blame or if it was a motion blur issue. Afetr this new information, what's the answer? I start thinking that it's a lens sharpness issue...

At a shutter speed of 1/2000 s, this is emphatically not camera shake, and with the boy apparently standing still, I seriously doubt it's subject motion.

The other suggestion of heat miraging seems to me unlikely to be a large factor because the distance from the camera to the subject appears to be not that long. I can't rule out it contributing, though.

So IMO the likely causes are lens performance and autofocus performance. The 24-240mm is by most accounts the very worst Sony FE lens, and this is at 212mm, i.e., the longer end where it tends to struggle. The A7 IV does not have the absolute best autofocus system, but this is not a challenging shot, so unless some setting is very wrong or there's something broken, I would not expect autofocus to be a big factor here.

At this point, I would probably test it by tearing a page out of a well-printed magazine or book, taping it to an outdoors wall (preferably at a time where it's neither very hot nor very humid), standing about as far away as you were here from the boy and perpendicular to the wall at the spot of the page, zooming to these settings (212mm here, at f/6.3), and taking several shots with autofocus single, and several shots with manual focus with focus peaking so you can try to get it best by eye. Outdoors during the day you should be able to get shutter speeds where a tripod isn't necessary; if you have a very good tripod, then that would improve the test, but IMO an inexpensive or light tripod won't, and no tripod is necessary to get useful information. After such a test, how share is the printed text?

 

[Mr_Win]

I am quite sure I was the most steady I could be, when shooting.
There was a speak about if it was the lens to blame or if it was a motion blur issue. Afetr this new information, what's the answer? I start thinking that it's a lens sharpness issue...

Sorry, but look at the net behind the boy. Looks like camera shake. Might be heat waves but I don't know much about that.

Steve

naw, the lens is just not that sharp, the image is fine, even the shirt seems have some sharpening.



this is 60mp on arguably the worse lens, and quite frankly, the image is technically fine. If you want super sharpness, the 24-240 + 60mp is not a good idea.



60mp is going to stroke your gear lust, that is by design, sony knows what they are doing.

 

[AlesSor]

Probably someone already mentioned this, but too tired to read through all the responses.

There is a a setting somewhere in the menus about RELEASE PRIORITY. Do you want the camera to take the photo as soon as you press the shutter button? Or do you prefer to let it finish its AF function, which means you are guaranteed sharpness (as much as the lens allows) but may miss the action by a fraction of a second.
Unless your subject is super fast moving, I would change this setting to AF priority, especially before purchasing a better, more expensive lens.

thanks, i had the option set to balanced and now i set to af priority. Anyway I am quite sure that i push the shutter when i hear the confirmation sound of the autofocus so i think it's unlikely that i pushed the button before the focus action was finished

 

[AlesSor]

I am quite sure I was the most steady I could be, when shooting.
There was a speak about if it was the lens to blame or if it was a motion blur issue. Afetr this new information, what's the answer? I start thinking that it's a lens sharpness issue...

Sorry, but look at the net behind the boy. Looks like camera shake. Might be heat waves but I don't know much about that.

Steve

naw, the lens is just not that sharp, the image is fine, even the shirt seems have some sharpening.

this is 60mp on arguably the worse lens, and quite frankly, the image is technically fine. If you want super sharpness, the 24-240 + 60mp is not a good idea.

60mp is going to stroke your gear lust, that is by design, sony knows what they are doing.

my a74 is 33 mp not 60

 

[4Photos]

Probably someone already mentioned this, but too tired to read through all the responses.

There is a a setting somewhere in the menus about RELEASE PRIORITY. Do you want the camera to take the photo as soon as you press the shutter button? Or do you prefer to let it finish its AF function, which means you are guaranteed sharpness (as much as the lens allows) but may miss the action by a fraction of a second.
Unless your subject is super fast moving, I would change this setting to AF priority, especially before purchasing a better, more expensive lens.

thanks, i had the option set to balanced and now i set to AF priority. Anyway I am quite sure that i push the shutter when i hear the confirmation sound of the autofocus so i think it's unlikely that i pushed the button before the focus action was finished

In that case there is another delay: between you hearing the AF confirmation beep and your pressing the button is a enough time for a fast moving object to move out of focus. By setting your camera to AF Priority, you don't wait for the confirmation beep (in fact, I switch the beep off). You push the shutter release through, and the very moment the AF is obtained, the picture is taken. Sharps as it gets. Only limiting factor than the quality of your lens (and you could and should test this with stationary objects).

 

[AlesSor]

Probably someone already mentioned this, but too tired to read through all the responses.

There is a a setting somewhere in the menus about RELEASE PRIORITY. Do you want the camera to take the photo as soon as you press the shutter button? Or do you prefer to let it finish its AF function, which means you are guaranteed sharpness (as much as the lens allows) but may miss the action by a fraction of a second.
Unless your subject is super fast moving, I would change this setting to AF priority, especially before purchasing a better, more expensive lens.

thanks, i had the option set to balanced and now i set to AF priority. Anyway I am quite sure that i push the shutter when i hear the confirmation sound of the autofocus so i think it's unlikely that i pushed the button before the focus action was finished

In that case there is another delay: between you hearing the AF confirmation beep and your pressing the button is a enough time for a fast moving object to move out of focus. By setting your camera to AF Priority, you don't wait for the confirmation beep (in fact, I switch the beep off). You push the shutter release through, and the very moment the AF is obtained, the picture is taken. Sharps as it gets. Only limiting factor than the quality of your lens (and you could and should test this with stationary objects).

That's very interesting information, i din't think about the possibility of the subject moving between the confirmation and the shutter button pressing but, since i shoot in af-c, shouldn't the camera offset the subject movement?

 

[dazzerfong]

Probably someone already mentioned this, but too tired to read through all the responses.

There is a a setting somewhere in the menus about RELEASE PRIORITY. Do you want the camera to take the photo as soon as you press the shutter button? Or do you prefer to let it finish its AF function, which means you are guaranteed sharpness (as much as the lens allows) but may miss the action by a fraction of a second.
Unless your subject is super fast moving, I would change this setting to AF priority, especially before purchasing a better, more expensive lens.

thanks, i had the option set to balanced and now i set to AF priority. Anyway I am quite sure that i push the shutter when i hear the confirmation sound of the autofocus so i think it's unlikely that i pushed the button before the focus action was finished

In that case there is another delay: between you hearing the AF confirmation beep and your pressing the button is a enough time for a fast moving object to move out of focus. By setting your camera to AF Priority, you don't wait for the confirmation beep (in fact, I switch the beep off). You push the shutter release through, and the very moment the AF is obtained, the picture is taken. Sharps as it gets. Only limiting factor than the quality of your lens (and you could and should test this with stationary objects).

That's very interesting information, i din't think about the possibility of the subject moving between the confirmation and the shutter button pressing but, since i shoot in af-c, shouldn't the camera offset the subject movement?

Your camera has 3 modes of focus/shutter actuation:

• Focus priority. Basically your camera wants to confirm that the lens is actually in focus, then it'll allow your shutter to go off and take a picture.
• Release priority. Camera doesn't do a check, it just takes a picture regardless.
• Balanced. Somewhere in between focus and release.

As the 24-240 has a pretty slow focus motor, it's possible the lens hasn't actually acquired focus.

At this point, I think a static test indoors with a tripod is the best bet. You're not really helping your investigation by refusing to do it to confirm that the lens isn't actually problematic.

 

[AlesSor]

So,



today was able to do the tripod test... I put the camera on a tripod and try to get 212mm (but got 191). shoot 2 pictures with different shutter speed. What do you guys think?

Note that the focus point was on the book and i used 2 seconds self timer to get the camera steady...

100% crop

other picture:

100% crop

 

[4Photos]

What are the EXIF data? Hard to interpret without.

 

[Locutus001]

I think people already gave the correct answers to your problem.

First of all probably the lens itself ^.^

But even though it might be redundant I'd like to point out (again) that the shot you have taken in the resolution it is supposed to be is sharp. If you want to pixel peep for sharpness, you can only push it that far. The boy playing soccer, standing infront of the goal in itself is a sharp picture. Only the face of the boy zoomed into... a different story. But if I take pictures with my A7RIII or A7CII with a 135mm GM... I also can only crop that much. The only difference is that the sharpness my 135mm GM can produce is probably much higher than a zoom lens of that big of a distance at some random mm count.

With the 85mm f1.8 (a very nice lens by the way) I am sure you will get vastly different results.

 

[AlesSor]

sorry, I don't know why data wasn0t incorporated:

pictures are: iso 640 - 1/40 f7.1 ; iso 4000 - 1/250 f7.1

 

[Mr_Win]

the lens is soft at longer focal lengths, just let it go already, it's been well documented on 24 and 42mp cameras.

The Sony FE 24-240mm f3.5-6.3 OSS is a small and relatively light super-zoom which offers a huge 10x zoom-range with the wide end starting at 24mm focal length. Together with a useful 1:3.6 magnification the lens is an ideal travel companion which covers a wide array of photographic opportunities and isn't too expensive either. But I had expected more from the optical images stabilization which could partly compensate for the lowly focal ratio. And sharpness of the lens at focal lengths beyond 105mm is not a good match for todays high resolution sensors. But still: the versatility of the FE 24-240mm f3.5-6.3 OSS is unmatched which earns it a recommendation!

- camera labs

Starting with the best, in the center of the frame: With a wide-open aperture, center of the frame results are reasonably sharp at 24mm and 50mm, become soft by 100mm, and are quite soft at 240mm. The middle of the image circle results are quite sharp by f/5.6 at 24mm and 50mm. At 100mm, f/5.6 is wide open and f/8 results are decent at this focal length. At 240mm, f/11 results appear similar to wide-open 24mm results which are decent and considerably better than the 240mm f/6.3 results.

-TDP

The bottom line is that some photographers will ignore these kinds of lenses while others prefer the versatility they can offer. (After all, we are talking about a 10x full-frame zoom lens here.) You can get interesting results and have the reassurance that you will never miss a shot due to not having the correct focal length mounted. However, in terms of pure optical quality, you are constrained to specific settings. It is a matter of understanding how much you are willing to sacrifice for the sake of versatility.

-mirrorlessons

you are pixel peeping a lens that gives "acceptable" sharpness, and your images are acceptable, but they arent crispy. I've used the lens extensively, and it's soft compared to shorter range zooms, and if you must have a lot of range, the tamron 28-200 is far sharper, and even then, it's not as sharp as shorter range stuff. AFAIK, the nikon is the only other lens that's sharp and it goes from 24-200.