Just how much of an effect does sensor size have on low light performance and color reproduction? This insightful video from The Slanted Lens aims to find out, making a few interesting discoveries along the way.
Clearly, the typical uses of the medium format Hasselblad X1D II 50C and the micro four-thirds Panasonic GH5 are dramatically different, but it’s fascinating to see just how diverse their performance is when facing the same task. Jay P. Morgan and Kenneth Merrill conjure some tough challenges for the smaller sensor cameras, and as you’d expect, the image quality really suffers when pushed in circumstances that the medium format and full-frame cameras find relatively acceptable.
Obviously, there are a lot of other elements that are not taken into consideration in this video, not least being the price of each system. The Hasselblad X1D II 50C paired with the XCD 45mm lens will cost you more than four times as much than the GH5 paired with a similar lens.
Despite the poor performance of the smallest sensor, the differences are not as massive as you might expect. Furthermore, there’s a strong argument that micro four-thirds is more than adequate for most consumers, especially when the cost is taken into consideration. Image quality will be good enough for the vast majority of shooting situations, and the portability and convenience of the system make it far more accessible.
How well did your favored sensor size perform? Let us know your thoughts in the comments below.
Looks about right. You can really see how inferior the Sigma 16mm is compared to the other lenses in this.
For my use anyway I think the Panasonic held up OK considering how extreme of cases they were using.
I think the should have changes the apertures though. No one would shoot any of the other cameras at f8 for that shot. they used f8 to get enough depth of field on the Hasselblad. no reason for that on a crop sensor and probably not even on the full frame. seems like they were robbing light form all the other cameras and on the GH5 they were even introducing diffraction. Also who underexposes a shot by 7 stops either blow the highlight in the windows out or use a flash.
It was a poor test. For MFT, they ought to have used a DSC-centric camera, such as the Olympus OM-D E-M1 mk II, (or E-M1X) and not a video-centric choice such as the Panasonic. Furthermore, they ought to have tested “systems,” using Hasselblad lenses on the Hasselblad, the two Sony with appropriate Sony lenses, and the Olympus with an M.Zuiko lens.
Additionally, for the APS-C, they ought to have used a manufacturer which is vested in the APS-C format. All formats ought to be “flagship” models. Had they used the “flagship” Sony α9 II, they would now have three out of four cameras all at 20-24 Mpx, a much more fair comparison of pro models IF sensor size was the thing being compared, (and the price of the F-type would have been closer to the Hasselblad).
Additionally, all cameras ought to have been tested at their base sensitivity value. To say, “camera A had a base Sv of ISO 200/24°, so we set all cameras to an EI of,…” is not a good way of doing it, because that is NOT how people use their cameras.
This test told me nothing. What I do know is that there are many pros shooting APS-C and MFT systems, so it does not really matter.
M1X and the GH5 use the same sensor, so results would have been near identical. As far as using system lenses, ehh at f/8 they all look the same anyway.
I don't think this was as wildly disproportionate as people make it out to be. No matter what lenses, and what apertures they used, the conclusion would be the same:
Bigger chips give cleaner files.
Lens choices make a HUGE difference, even when all at f/8.0. There are several YT videos comparing different lenses on the same sensor to establish the fact.
I will even argue that a great lens on a small sensor beats a mediocre lens on a big sensor, (resolution being equal).
Also, same sensor does not mean same quality. Sony, Nikon, & Pentax proved this once. Three cameras, same Sony sensor, yet the Pentax slightly outperformed the Nikon, (not by any reasonably compelling margin), but they both outperformed the Sony by far. That is why we compare “systems.”
If they all look the same anyway the clear winner is the cheapest camera not the most expensive.
Since the obviously do not all look the same I'll expand a little. I do agree that the results would have placed the cameras in the same order, but the test is "slanted". They set it up so there winner would get the best results that it could. Then they kicked all of the competition in the knee before the race to widen the gaps. The only setting that needed to be constant was shutter to get the same motion blur. They should have taken dozens of shots with each camera at different settings and picked the best of each. This would have given much more realistic real word results and I'm sure the gaps would have narrowed. If all you wanted to know was the order what is the point of running the test,.
Same sensor doesn't mean same processor. Also yeah lenses make a world of difference, which is why it's also utter garbage they use $300 budget third-party lenses on the Sony bodies (one of them is even a format mismatch)! They use a $2700 lens on the MF body, but can't scrounge more than $300 for the FF and APS-C or $400 for the MFT (which is also not a lens designed specifically for the format used)?
Originally, I did not watch the video, but read TSL article. I just watched the less detailed video, and realised how ignorant they are.
«Bigger sensor, more DR hence, more detail.»
Wrong! More Mpx, more detail. The two 50+Mpx sensors had more detail than the 20Mpx sensor. I wonder why? Could it be because it was made up of more dots? Nah! It MUST be because it was smaller!
«Bigger sensor causes less glaring of highlights.»
Wrong! A six-blade, straight aperture causes glaring in six directions, (and bad glaring to boot), while a nine-blade, rounded aperture causes more even glaring in eighteen directions.
If they really took an objective look at the LA CARE marquee, (or the other one), they would see how badly blurred they are in the Hasselblad, when any of the six directions of the aperture blades are casually observed. The other lenses gave a more even blur, and less of it. But of course it was not the lens, it was the sensor size!
As for what they said concerning the «loss of colour,» on the smaller sensors, versus the Hasselblad, I can only see that on one camera, (and maybe just a touch on another). They are calling things as “obvious” which I just did not see, either in the video, nor on their website.
They clearly had an agenda here, despite one of them getting it right; more megapixels, more detail, as he spoke about stitching images.
«APS-C/MFT good enough for consumer/snapshots.»
Yes, and an excellent choice for quite a vast number of professionals who use them. When I was shooting landscapes on 35mm film, I longed for a medium format for that finer grain. Someone taught me how to stitch frames together in the darkroom for increased resolution, and I thought, although the technique worked, was too much work. Easier to buy a 5×4 view camera, or a Pentax 6x7, than to stitch 35mm film frames together.
In today's world of digital processing, that has all changed. For those few times —very few, and very far between— in which I require more than the 24Mpx of my APS-C, I can easily stitch frames, and get a better image than a single-shot medium format. Push comes to shove, I can rent for a week, (or a weekend).
So what did I learn from this?
① More pixels, more detail.
② Match exposure index to sensor sensitivity for the most dynamic range.
③ Expose for your subject, (not your scene), for the most dynamic range.
④ Corollary of ③, understand what part(s) of the scene is really the subject.
⑤ Rounded aperture blades are better than straight blades.
⑥ Good glass trumps good camera, and good photographer trumps good glass.
[EDIT]
⑦ Black swans are black, white swans are white. Expose for the light incident on the subject, not the subject
[/EDIT]
No,… wait…! I did not learn any of that from this. This simply reiterated what I learned in the eighties, changing “AgHalide crystals,” for, “pixels,” and “film,” for, “sensor.”
they really had some space left for the subject, between all the advertisements. Amazing.
They must have been bored, the whole concept is unreal. I used to use MFT and the Olympus EM5 was pretty good in low light.
I moved to full frame for completely different reasons but about a year ago I swapped from a D610 to a D850 not because of the relative performance of the sensors in low light but purely because the D850 will acquire focus in lower light.
Hasselblad? Well, in ‘normal’ use, by the time that camera has started up, the subject will have got bored and wandered off!
It's too bad you didn't stick with Olympus - the E-M1X will focus at -6 EV, the D850 at -4 EV.
It would make better sense to either set each cameras as it's best used, or along the same equivalence line. Everything at F/8 ISO 200 don't make sense. I also wonder how much vibration they get from those camera straps in the wind.
f/8 on MFT doesn't soften the image *that* much. For most of the lenses 5.6 is overall optimal anyway, with 2.8 being optimal for center performance.