4
\$\begingroup\$

My friend and I took near-identical photos of the same neon light sign. I'm on a Canon SL2 with a 50mm prime with f1.8. He's using a Sony a7c with an 85mm f1.4 lens.

Looking at the images a bit more closely in Lightroom (untouched), I noticed that my image (on the right below) is not quite distinguishing the neon lights as cleanly as his is (on the left); my neon lights look more blown out. On his photo you can also see more details on the white area behind the neon lines.

What's causing this discrepancy? The aperture and ISO are roughly similar. Is it simply the difference between a cropped and full frame sensor, or does his camera have better dynamic range? I'm just getting started in photography so I'd like to figure this puzzle out as something to aim for.

enter image description here

\$\endgroup\$

7 Answers 7

14
\$\begingroup\$

Tl;dr: The Sony exposed less, reducing bloom. The Sony's lens plays in a different league. A larger sensor helps.

The first thing to note is that the exposures are not equivalent. The reason the bright parts in the neon sign are washed out is that they are overexposed. This overexposure is expected for such a scene because the dynamic range is too large and the camera finds a compromise sacrificing the brightest and darkest areas in the image. A quick spread sheet I made confirms this, giving a (dimensionless) exposure of 0.816 for the Sony and 0.965 for your Canon: Exposure computation I'd actually say that your camera gets it right because the brick wall is more adequately exposed. My suspicion is that the bright glare from an outside light source in your friend's shot increases the overall light measured by the sensor, which accordingly leads to a lesser exposure. Incidentally, we are less interested in the brick wall than in the neon sign which is why we subjectively like the darker exposure better.

That increased exposure in the neon sign in your image may lead to some "leaching" of the very bright areas into the surrounding parts of the image (I think the technical term is "blooming") which increases the perception of bad optical definition because, as mentioned, unfortunately our attention is drawn exactly to this problematic piece.

Secondly it appears indeed, as other answers note, that the Canon image is less sharp overall. Here are a number of possible reasons which actually may all contribute to different degrees.

  1. I cannot see an obvious, conspicuous motion blur (more on that later) but rather a general diffuse "unsharpness".
  2. The autofocus may have generally struggled in the low-light condition.
  3. The autofocus may not be perfect in the first place. Expensive cameras have procedures to calibrate the autofocus in order to compensate for production tolerances. Any (auto)focus issue has a larger impact at large apertures.
  4. Using a lens at full aperture is putting it to the test. The EF 50mm/1.8 is a cheap lens. 30 years ago I noticed that my comparable Nikon 50mm/1.8 lens should not have been advertised as such because the optical quality was too inferior at f/1.8. By contrast, the Sigma 85mm f1.4 DG DN Art is a $900 lens aiming at a professional market. "It’s ridiculous how sharp this lens is, even when wide open.."
  5. Additionally, the Canon has a smaller sensor which may more aggressively de-noise the image at low light conditions.

Now to the motion blur. Interestingly, it appears to me that it is your friend's Sony image that has noticeable motion blur. Consider the following detail from that shot: enter image description here The duplicated lines are typical indications of horizontal movement. A close look reveals duplication even in the vertical lines of the brick wall (but not in the horizontal ones). It is astonishing that the image still looks reasonable sharp overall, probably due to the lack of narrow, high-contrast lines in other places.

By contrast, your Canon picture shows no obvious double lines that I can see. If there is motion blur, it is uniform in all directions.

\$\endgroup\$
1
  • \$\begingroup\$ Thank you for the incredibly detailed answer! To your point of My suspicion is that the bright glare from an outside light source in your friend's shot increases the overall light measured by the sensor, one thing my comparison image didn't show is some very blurred-out white lights on a tree that we were both shooting through, so maybe that would explain part of the outside light source? My image had those blurred lights too, they were just cropped from this comparison. \$\endgroup\$ Nov 11, 2022 at 4:10
4
\$\begingroup\$

Accuracy of focus and camera movement during the exposure. 1/80 vs. 1/125 plus the ability to steadily handhold the camera.

\$\endgroup\$
5
  • \$\begingroup\$ I see; how could I improve focus accuracy? I know the SL2 only has 9 focus points compared to the 400+ on the Sony... \$\endgroup\$ Nov 10, 2022 at 4:33
  • \$\begingroup\$ Likely the main ingredient is shutter speed and handholding. \$\endgroup\$ Nov 10, 2022 at 5:04
  • 1
    \$\begingroup\$ Relevant other question about sharpness: photo.stackexchange.com/q/50006/9161 \$\endgroup\$ Nov 10, 2022 at 16:11
  • \$\begingroup\$ @SaaruLindestøkke That question was much more useful before the link to the example images became dead and had to be edited out. \$\endgroup\$
    – Michael C
    Nov 15, 2022 at 0:35
  • \$\begingroup\$ Normally I'd agree with the poor technique due to long Tv and handholding, but the camera with the shorter Tv shows more evidence of camera movement. This is the rare case where it really is the gear. The $100 Canon EF 50mm f/1.8 is notoriously soft wide open. The $1,100 Sony 85/1.4 is notoriously sharp wide open. The Canon shot is about as good as can be expected with that lens wide open at f/1.8, even if the camera is tripod mounted and used with mirror lockup and a remote release. \$\endgroup\$
    – Michael C
    Nov 15, 2022 at 0:38
2
\$\begingroup\$

There are many things to observe about these two images:

  • The Sony image is more accurately focused but has some camera shake / motion blur.

  • The Canon image is definitely off-focus - you can look at the sharpness of the neon text and the brick outlines. The Canon image could have motion blur, but the misfocus is a more serious problem.

  • The Canon image is a bit brighter, which you can tell by comparing the red bricks across both photos.

  • The two fields of view are roughly equivalent (85mm full frame vs. 50mm × 1.6 APS-C crop factor = 80mm).

  • The depths of field are not equivalent. Full-frame 85mm @ f/1.4 would be equivalent to APS-C 53mm @ f/0.88. However, both images don't have much foreground-background separation, so this can be ignored for now.

  • The two camera bodies can have different color grading curves, so even if the same amount of light falls on both sensors, they can end up with different-looking JPEG images.

Next time, to make a fairer comparison between the two cameras, please do these things:

  • Shoot on a tripod to minimize camera shake.

  • Focus accurately, checking with the digital magnifier and/or using manual focus as necessary.

  • Use the same shutter speed, f-number, and ISO. Use different focal lengths if the crop factors are different, as in your case.

  • After matching the exposure parameters, also try tweaking the exposure a bit to match the rendered image brightness (i.e. what you see from both cameras' JPEGs).

  • Consider stopping down both lenses to increase sharpness and reduce bloom. Shooting wide open at f/1.4, f/1.8, etc. could produce excessive blurring artifacts.

\$\endgroup\$
2
\$\begingroup\$

I would say the difference is primarily due to handholding technique and image stabilization...

The Canon SL2 is dependent on 2 axis lens-based stabilization, which the Canon 50mm doesn't have; the Sony A7c has 5 axis in-body stabilization, which is effective with all lenses.

I would also guess that the much more expensive (~10x) 85/1.4 ART lens is much better corrected for off axis light (ghosting/flare/etc) and sharpness.

The slight difference in exposure is negligible, and quite likely attributable to the metering mode selected (or how the two cameras actually implement it).

\$\endgroup\$
1
  • \$\begingroup\$ A careful examination of both sides indicates the Sony shot has horizontal motion blur. The Canon shot does not. This seems to me to mainly be about the poor resolution of the Canon 50mm f/1.8 II (a $100 lens) @ f/1.8 compared to the much better Sigma 85mm f/1.4 (an $1,100 lens) @ f/1.4. \$\endgroup\$
    – Michael C
    Nov 13, 2022 at 18:02
2
\$\begingroup\$

Neon lights flicker with the frequency of the AC electric supply. 60 Hz in this case, or once every .017 seconds. I suspect that the slightly longer exposure on the right caught more of the brighter part of the cycle.

\$\endgroup\$
10
  • \$\begingroup\$ Flicker of neon lights is twice the line frequency. Thus, a 60 Hz mains will cause neon lights to flicker at 120 Hz. \$\endgroup\$
    – qrk
    Nov 13, 2022 at 1:15
  • \$\begingroup\$ @qrk In which case the 1/125 Tv would catch almost exactly one cycle, while the 1/80 Tv would capture roughly 1.66 cycles. Shot-to-shot variation at 1/125 would be very low. Shot-to-shot variation at 1/80 could be much more variable in either direction. \$\endgroup\$
    – Michael C
    Nov 13, 2022 at 18:06
  • \$\begingroup\$ @MichaelC If my maths is correct, the chance exposure variation is about 0.057 stops when shooting at 1/125 s for a light source that linearly varies brightness with a sinusoidal voltage source. At 1/80 s, the chance exposure variation is about 0.31 stops. However, neon signs are not linear devices, so the above is to be taken lightly. \$\endgroup\$
    – qrk
    Nov 14, 2022 at 6:33
  • \$\begingroup\$ @qrk The first (sony) image is obviously catching almost one complete cycle. You can tell this due to the camera motion. If the camera was moving from R to L, the shutter opened just as the neon was lighting up before peaking. The neon then bottomed out (and was almost totally dark), then began to brighten again just as the exposure ended. But we don't know if the neon was flickering due to AC or if it was blinking on and off very rapidly as some neon displays do to make them more "exciting". Have you ever been to Vegas? Seen the Flamingo at night? \$\endgroup\$
    – Michael C
    Nov 15, 2022 at 0:01
  • \$\begingroup\$ The second (Canon) image has caught a full cycle plus part of another. The overlap seems to be in the brighter part of the cycle. So it caught two peaks with only one trough in between. If the exposure had started half a cycle later, it would have caught two troughs with one peak in between them and been much darker for the same exposure duration. \$\endgroup\$
    – Michael C
    Nov 15, 2022 at 0:03
2
\$\begingroup\$

I've used the Canon EF 50mm f/1.8 II plenty in the past. A lot of folks talk about how fantastic a lens it is. They've even nicknamed it the Plastic Fantastic. That's mostly because it's the first prime lens they've ever used. It's not a great lens. When used wide open it's not even a good lens. It is a fantastic value for the price, particularly when stopped down to f/2.8 or narrower. It's also better than most kit zooms used wide open when the 50/1.8 is stopped down to around f/2.2 or f/2.5, which is still faster than any of those kit zooms.

Normally I'm the first to write answers that advise folks to improve technique and get the most you can out of the gear you have instead of thinking a camera upgrade will solve the photographer's own artistic limits or poor technique. Most issues with blurry pictures are due to poor technique. A better camera or lens doesn't make anyone a better photographer, it just reduces the limits imposed on the photographer by the gear.

But even the latest, greatest, best, and most expensive gear can impose limitations on the most skilled photographers. The reason those photographers are the best is because they know the limits of the tools they are using and are able to work just within the boundaries of those limits. They also understand which of the multiple tools in their bag are most appropriate for a particular shooting situation.

But in this case, honestly, the example on the right is about the best I think one can expect from any of the three versions of Canon EF 50mm f/1.8 lenses when used wide open. It really is that soft when used at f/1.8. Even if the camera were on a sturdy tripod, set to use mirror lockup, and triggered with a remote shutter release, I doubt one could do much, if any, better with any of the three versions of the Canon 50/1.8 when used wide open at f/1.8.

On the other hand, the Sigma 85mm f/1.4 ART is known to be a nicely sharp lens, even wide open at f/1.4. That's why it costs about ten times as much as the Canon 50/1.8.

Even stopped down to, say, f/4 and used on a tripod with remote release and mirror lockup I doubt the EF 50mm f/1.8 could do any better than the Sony α7C/Sigma 85/1.4 shot, other than regarding camera movement. If one looks carefully at the Sony shot, one can see evidence of rotational movement in the 'E' and both 'S' letters at the bottom of the frame. Perhaps it's a combination of camera movement and IBIS correction that counteracted camera movement in most, but not all, of the frame?

Sometimes it isn't all about technique. Some rare times it really is the gear.

You may have heard: Gear doesn't matter.

While very often true, it is only half the truth.

The full truth is: Gear doesn't matter... until it does.

When the gear does matter, it really does matter.

\$\endgroup\$
5
  • \$\begingroup\$ "ten times as much as the Canon 50/1.4" - did you mean 1.8? \$\endgroup\$
    – Zeus
    Nov 16, 2022 at 0:38
  • \$\begingroup\$ @Zeus current prices at B&H are $125 for the Canon and $1199 for the Sigma. Close enough to 10x for me. Oops, sorry - I see you're confused abot f/1.4 vs. f/1.8. So am I. \$\endgroup\$ Nov 16, 2022 at 1:01
  • \$\begingroup\$ The nickname I know for the Canon EF 50mm f/1.8 is the nifty fifty \$\endgroup\$
    – Nayuki
    Nov 16, 2022 at 14:07
  • \$\begingroup\$ @Zeus Yes. Thanks. \$\endgroup\$
    – Michael C
    Nov 17, 2022 at 1:14
  • \$\begingroup\$ @Nayuki Pretty much all of the budget 50/1.8 lenses from Canon, Nikon, Sony, Pentax, as well as in the past Minolta, Konica, etc. are known collectively as nifty fifties. \$\endgroup\$
    – Michael C
    Nov 17, 2022 at 1:16
0
\$\begingroup\$

In the settings of your camera, the exposure values ​​of the shutter speed are such that the shutter has been kept open for a longer time than Sony, Even considering that the f-stop value in Canon is a bit more closed than Sony but the ISO value in Canon is also higher than Sony. All these factors cause more exposure on the sensor and this makes it difficult to separate the details of the final image. And when more light passes through the aperture (in all brands and lenses), it will cause a drop in sharpness because it brings optical distortion, and if you reduce the f-stop value in Canon, you will definitely have higher details than You could see dark parts in the final image and even higher sharpness.

\$\endgroup\$
0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.