Does using FX lens on DX body reduce sharpness?

[RoyAmatore]

I saw Tony Northrup say that using a full frame lens on a crop sensor results in a loss of sharpness. Does this sound right to you, experts?

 

[Tony Beach]

I saw Tony Northrup say that using a full frame lens on a crop sensor results in a loss of sharpness. Does this sound right to you, experts?

Not at the pixel level. The argument is that you enlarge the DX format more than the FX format, but if you have a good lens you should be able to match results with a 24 or 21 MP FX camera versus a 24 or 21 MP DX camera. Hard to make direct comparisons though since you are covering a different FOV with each format; so there are several things you have to "balance" as you try to make the comparisons -- including equalizing the aperture, comparing sharpness and vignetting in the corners, and in some way changing the composition either by zooming the lens (which means it's no longer a direct comparison) or changing the perspective. Bottom line though is if the lens resolves to the level of the sensor there is no issue, it's only an issue when you are comparing a lens that is less than optimal on both formats.

 

[Leonard Migliore]

I saw Tony Northrup say that using a full frame lens on a crop sensor results in a loss of sharpness. Does this sound right to you, experts?

Well, if you compare a full frame image with a crop sensor image taken with the same lens, the full frame image will be sharper since the lens resolution is the same but you don't have to enlarge it as much. This may not be true if you have a full frame lens with very bad corners.

In a more hypothetical vein, a lens designed for crop sensor use can be sharper than a lens of the same focal length and aperture designed for full frame because it doesn't have to cover as much area; this allows the designer to optimize the image over a smaller diameter circle. But usually, full frame lenses are more expensive than crop sensor lenses and tend to be sharper all over.

So what did Tony say, exactly? In my experience his statements are accurate but may be subject to misinterpretation.

 

[lickity split]

I saw Tony Northrup say that using a full frame lens on a crop sensor results in a loss of sharpness. Does this sound right to you, experts?

Can you share the link to the video ?

 

[RoyAmatore]

Here's the link. http://www.dpreview.com/forums/reply?parent=57298753 He is discussing the Sony a6300 and speaks generally. "Significantly less sharp images" using full frame glass . He discusses this at !3:20 into the video. He also sites a link to a full video that explains it here:

.

 

[Tony Beach]

Here's the link. http://www.dpreview.com/forums/reply?parent=57298753 He is discussing the Sony a6300 and speaks generally. "Significantly less sharp images" using full frame glass . He discusses this at !3:20 into the video. He also sites a link to a full video that explains it here:

.

Sorry, but that's pretty weak on his part.

 

[RoyAmatore]

Here's the link.http://www.dpreview.com/forums/reply?parent=57298753He is discussing the Sony a6300 and speaks generally. "Significantly less sharp images" using full frame glass . He discusses this at !3:20 into the video. He also sites a link to a full video that explains it here:

.

Sorry, but that's pretty weak on his part.

 

[Tony Beach]

Here's the link.http://www.dpreview.com/forums/reply?parent=57298753He is discussing the Sony a6300 and speaks generally. "Significantly less sharp images" using full frame glass . He discusses this at !3:20 into the video. He also sites a link to a full video that explains it here:

.

Sorry, but that's pretty weak on his part.

No need to be sorry.

I'm sorry for Tony Northrup.

Not sure what you're trying to say, though.

There are numerous issues with DxO Mark's rating lenses the way they do, and to solely rely on that as an explanation of his friend's issue is wholly insufficient. Take that 18-55 kit lens and compare it at any aperture and any focal length to that 24-70 (IIRC, that was the comparison he was making), and you will see that the 24-70 is clearly better than the 24-70 on an APS-C camera. DxO Mark was probably saying something like the "perceptual" sharpness of the 18-55 is "best" wide open at 18mm, and maybe they were measuring the 24-70 wide open at 70mm -- that's an irrelevant comparison, and almost certainly not consistent with the experience of anyone that has actually used those lenses (lenses are always sharper when they are stopped down some).

 

[Digetydog]

I saw Tony Northrup say that using a full frame lens on a crop sensor results in a loss of sharpness. Does this sound right to you, experts?

He is a XXXXX if he believes that.

 

[JimPearce]

It has nothing to do with the lens, it is the sensor that shifts the spatial frequency of details upward. But a 300 f2.8 VR on my D7100 still out-resolves a 300 f2.8 VR on a D810 in DX crop mode (or cropped to DX).

 

[Flashlight]

I saw Tony Northrup say that using a full frame lens on a crop sensor results in a loss of sharpness. Does this sound right to you, experts?

Well, if you compare a full frame image with a crop sensor image taken with the same lens, the full frame image will be sharper since the lens resolution is the same but you don't have to enlarge it as much. This may not be true if you have a full frame lens with very bad corners.

After you compensate for FL, DoF and ISO the digital files from comparable cameras like the D7200 and D750 are basically equal. When displaying them or printing them one has not to be enlarged more than the other.

The enlargement argument was valid in the film days where the emulsion of 35mm film and 6x6cm film was basically the same (for say Ektachrome 64). So when enlarging 35mm film to the same size as 6x6cm film the grain became more visible while the total number of light sensitive crystals (overall slide/negative resolution) was lower to begin with on 35mm film. That problem is non-existent on two 24MP sensors (from the same generation) with different formats. It would be when comparing a 24MP FF sensor to a 10.6MP DX sensor.

Here's an example of D610 (sigma 90 macro) vs D7100 (nikon 60 macro):

In a more hypothetical vein, a lens designed for crop sensor use can be sharper than a lens of the same focal length and aperture designed for full frame because it doesn't have to cover as much area; this allows the designer to optimize the image over a smaller diameter circle. But usually, full frame lenses are more expensive than crop sensor lenses and tend to be sharper all over.

So what did Tony say, exactly? In my experience his statements are accurate but may be subject to misinterpretation.

--
Leonard Migliore

--
Philip

 

[RoyAmatore]

Here's the link.http://www.dpreview.com/forums/reply?parent=57298753He is discussing the Sony a6300 and speaks generally. "Significantly less sharp images" using full frame glass . He discusses this at !3:20 into the video. He also sites a link to a full video that explains it here:

.

Sorry, but that's pretty weak on his part.

No need to be sorry.

I'm sorry for Tony Northrup.

Not sure what you're trying to say, though.

There are numerous issues with DxO Mark's rating lenses the way they do, and to solely rely on that as an explanation of his friend's issue is wholly insufficient. Take that 18-55 kit lens and compare it at any aperture and any focal length to that 24-70 (IIRC, that was the comparison he was making), and you will see that the 24-70 is clearly better than the 24-70 on an APS-C camera. DxO Mark was probably saying something like the "perceptual" sharpness of the 18-55 is "best" wide open at 18mm, and maybe they were measuring the 24-70 wide open at 70mm -- that's an irrelevant comparison, and almost certainly not consistent with the experience of anyone that has actually used those lenses (lenses are always sharper when they are stopped down some).

 

[Leonard Migliore]

I saw Tony Northrup say that using a full frame lens on a crop sensor results in a loss of sharpness. Does this sound right to you, experts?

Well, if you compare a full frame image with a crop sensor image taken with the same lens, the full frame image will be sharper since the lens resolution is the same but you don't have to enlarge it as much. This may not be true if you have a full frame lens with very bad corners.

After you compensate for FL, DoF and ISO the digital files from comparable cameras like the D7200 and D750 are basically equal. When displaying them or printing them one has not to be enlarged more than the other.

The enlargement argument was valid in the film days where the emulsion of 35mm film and 6x6cm film was basically the same (for say Ektachrome 64). So when enlarging 35mm film to the same size as 6x6cm film the grain became more visible while the total number of light sensitive crystals (overall slide/negative resolution) was lower to begin with on 35mm film. That problem is non-existent on two 24MP sensors (from the same generation) with different formats. It would be when comparing a 24MP FF sensor to a 10.6MP DX sensor.

That's sensor resolution rather than lens resolution. Both 24 MP sensors are 4000 X 6000 pixels. On a D750 the 4000 pixels cover 24mm and on a D7200 they cover 16mm. But most lenses don't resolve that well. If you have a lens that resolves 100 lp/mm, that comes out to 2400 lp/ picture height on a D750 and 1600 lp/ picture height on a D7200. So you should get some benefit in sharpness out of the larger sensor.

 

[Flashlight]

I saw Tony Northrup say that using a full frame lens on a crop sensor results in a loss of sharpness. Does this sound right to you, experts?

Well, if you compare a full frame image with a crop sensor image taken with the same lens, the full frame image will be sharper since the lens resolution is the same but you don't have to enlarge it as much. This may not be true if you have a full frame lens with very bad corners.

After you compensate for FL, DoF and ISO the digital files from comparable cameras like the D7200 and D750 are basically equal. When displaying them or printing them one has not to be enlarged more than the other.

The enlargement argument was valid in the film days where the emulsion of 35mm film and 6x6cm film was basically the same (for say Ektachrome 64). So when enlarging 35mm film to the same size as 6x6cm film the grain became more visible while the total number of light sensitive crystals (overall slide/negative resolution) was lower to begin with on 35mm film. That problem is non-existent on two 24MP sensors (from the same generation) with different formats. It would be when comparing a 24MP FF sensor to a 10.6MP DX sensor.

That's sensor resolution rather than lens resolution. Both 24 MP sensors are 4000 X 6000 pixels. On a D750 the 4000 pixels cover 24mm and on a D7200 they cover 16mm. But most lenses don't resolve that well. If you have a lens that resolves 100 lp/mm, that comes out to 2400 lp/ picture height on a D750 and 1600 lp/ picture height on a D7200. So you should get some benefit in sharpness out of the larger sensor.

With the 35-70 f/2.8, 70-200 f/4, 105VR f/2.8 and the recent f/1.8 primes I can see no real difference. I have some older lenses but don't use them anymore. But this Tony is basing himself on DxO numbers where they give something like 13 P-MPix (whatever that is ) for the D7100 + 70-200 f/4 and 22 P-MPix for the D750 + 70-200 f/4. I really can't see such a difference in my tests and/or normal photography.

 

[Leonard Migliore]

With the 35-70 f/2.8, 70-200 f/4, 105VR f/2.8 and the recent f/1.8 primes I can see no real difference. I have some older lenses but don't use them anymore. But this Tony is basing himself on DxO numbers where they give something like 13 P-MPix (whatever that is ) for the D7100 + 70-200 f/4 and 22 P-MPix for the D750 + 70-200 f/4. I really can't see such a difference in my tests and/or normal photography.

Current lenses are pretty good. And current DX sensors are pretty good. You can make very nice 16X20 prints from them and it is, indeed, hard to tell any difference from FX. So yes, there isn't much difference in sharpness in normal photography. But I'm sure there is one. It just isn't in the range where you or I can see it.

At this point, it's more of a philosophical argument than an issue of photographic practice.

 

[Brian From WI]

Comparing the same lenses on the different formats leaves enough wiggle room to construe the results to your liking. I think that sample variations of various production runs of the same lens will effect IQ way more than any of this.

I don't discriminate, I'll use FX glass on DX.

 

[Tony Beach]

Here's the link.http://www.dpreview.com/forums/reply?parent=57298753He is discussing the Sony a6300 and speaks generally. "Significantly less sharp images" using full frame glass . He discusses this at !3:20 into the video. He also sites a link to a full video that explains it here:

.

Sorry, but that's pretty weak on his part.

No need to be sorry.

I'm sorry for Tony Northrup.

Not sure what you're trying to say, though.

There are numerous issues with DxO Mark's rating lenses the way they do, and to solely rely on that as an explanation of his friend's issue is wholly insufficient. Take that 18-55 kit lens and compare it at any aperture and any focal length to that 24-70 (IIRC, that was the comparison he was making), and you will see that the 24-70 is clearly better than the 24-70 on an APS-C camera. DxO Mark was probably saying something like the "perceptual" sharpness of the 18-55 is "best" wide open at 18mm, and maybe they were measuring the 24-70 wide open at 70mm -- that's an irrelevant comparison, and almost certainly not consistent with the experience of anyone that has actually used those lenses (lenses are always sharper when they are stopped down some).

I'm no expert, but his argument seemed based more on math, not lens test data.

I have no idea what math is involved here, and I only watched about a minute of the video starting at approximately where you said to start, but what I saw was a reliance on DxO Mark's evaluation of lens/camera performance (which I have looked at before myself and came to the conclusion that it's mostly nonsensical). Whatever the case, it's clear he was not doing any hands-on analysis of the lenses and cameras in question.

 

[Nicholas128]

Harder to understand is Tony's statement in the more detailed video (

), that when you put an FX lens on a DX camera, you have to multiply not only the focal length by the crop factor, but also the aperture. So he seems to be saying that an f/2.8 FX lens on a DX body becomes an f/4.2. I understand that the angle of view and DOF is different on the DX body, but I don't fully understand what he means by this and some others have said its nonsense. Can anyone shed some more light on this (pun intended :-~) ?

N

 

[Leonard Migliore]

Harder to understand is Tony's statement in the more detailed video (

), that when you put an FX lens on a DX camera, you have to multiply not only the focal length by the crop factor, but also the aperture. So he seems to be saying that an f/2.8 FX lens on a DX body becomes an f/4.2. I understand that the angle of view and DOF is different on the DX body, but I don't fully understand what he means by this and some others have said its nonsense. Can anyone shed some more light on this (pun intended :-~) ?

Sure. For some reason, Tony has dragged equivalence into his exposition. One of the principles of equivalence is that to get an equivalent image from cameras with different sensor dimensions, you have to divide the f/stop by the crop factor as well as dividing the focal length by the crop factor. So a 150mm f/2.8 lens on FX is equivalent to a 100mm f/(2.8/1.5) or f/1.8 lens on DX. This gives you the same field of view and depth of field. It's unfortunate that he just dropped that one on everybody and, in the context that he discussing, it's just irrelevant. It has no part in what he's talking about since if you take a 24-70 f/2.8 and put it on a DX camera, it is no longer equivalent to the same lens on FX.

But it is true, as Tony notes, that a 24-70 f/2.8 on FX is sort of equivalent to an 18-35 F/1.8 on DX.

The revealed truth of equivalence can be found here:

http://www.josephjamesphotography.com/equivalence/

Tony would have been better off not mentioning anything about f/stops.

--
Leonard Migliore

 

[n057]

Harder to understand is Tony's statement in the more detailed video (

), that when you put an FX lens on a DX camera, you have to multiply not only the focal length by the crop factor, but also the aperture. So he seems to be saying that an f/2.8 FX lens on a DX body becomes an f/4.2. I understand that the angle of view and DOF is different on the DX body, but I don't fully understand what he means by this and some others have said its nonsense. Can anyone shed some more light on this (pun intended :-~) ?

N

Some insist that because Depth-Of-Field (DOF) is less shallow on DX, one must compensate when calculating "equivalence". So they believe that to make the image identical, the aperture must be larger on DX.

As far as I am concerned, I am more interested in gathering light on DX, and when I set my 70-200 at 200mm/2.8, I am effectively shooting at a 300mm field-of-view at f/2.8. DOF is the least of my concerns when I struggle at speeds of 1/10s.


JC
Some cameras, some lenses, some computers