[BigMackCam]

In recent months I've been involved in several forum discussions where it's been stated that greater sensor real estate (i.e. larger sensor size) results in better image quality... for instance, a medium format camera will produce better image quality than a so-called "full frame" camera, which in turn offers better image quality than APS-C, followed by Micro 4/3 and so-on.

I fully accept this can be the case, yet I find myself questioning the validity of such a blanket statement in all circumstances, and I feel it needs qualification when presented. Just one of my previous arguments has been that the desired reproduction medium and dimensions can render the advantage irrelevant. One response to this argument was that I had chosen to compromise by limiting my output in this way - yet in my view, if my intended output target is, for example, a 24" QHD wide gamut LCD screen, that's not a compromise... it's just my desired output medium, so it's not a compromise if that's what I wanted or needed.

Furthermore, I feel there are many other variables that contribute to image quality and potentially challenge the notion that greater sensor real estate always results in better image quality... sensor type / age / resolution / dynamic range, AA filter, lens performance, ISO level, subject matter and desired creative intent (e.g. DoF at FoV), file format and bit depth, raw processing software (and user abilities therein), and more.

I wanted to start this thread so we could discuss this concept in a constructive and respectful manner, and perhaps reach a more informative description of when and why larger sensors are better than smaller (or otherwise, agreement that the blanket statement described is, in fact, always correct). My opening gambit is that greater sensor real estate can matter, but this depends on a number of contributing factors, and - depending on the use case - a bigger sensor isn't always better. However, I'm open-minded and willing to be convinced otherwise.

Some qualifying background information on my shooting conditions that, in part, contribute to my own views:

I shoot with all sorts of equipment for a variety of reasons - compact fixed-lens cameras with tiny sensors, the Pentax Q and Q7, Pentax (and Samsung-rebranded) APS-C DSLRs, Sony SLT and E-mount full frame. I don't shoot medium format. In APS-C, my cameras range from 6MP to 24MP, whilst my full frame gear is all 24MP. I use a wide range of glass on my ILCs, from 1950s vintage primes to modern, high-performance primes and zooms, including a handful of especially good lenses (by especially good, I mean those that both render well and demonstrate impressive optical test data). But it's fair to say that I'm far more interested in lens rendering than technical performance data. I shoot all manner of subjects and styles, from shallow depth-of-field portraiture and still life, to maximum DoF products, landscapes and architecture. I occasionally (but not often) shoot birds - sometimes in flight - and other moving subjects, a little macro, but rather casually. I don't shoot astro. I shoot raw almost exclusively, processing my files in Darktable, RawTherapee and Lightroom as required, occasionally carrying out further post-processing in GIMP. I'm reasonably proficient in all of these. Most often, I reproduce my photos on a 24" QHD wide-gamut LCD monitor, 17" near-full-RGB laptop, 15.6" wide-gamut 4K laptop and 10" sub-RGB tablet, viewed at typical distances. I have a few prints (of varying sizes) of my photos on the walls of my home, displayed for casual viewing, with no expectation of examining them from a couple of inches away.

So... let's discuss

---

Last edited by BigMackCam; 08-12-2019 at 02:47 PM.

[victormeldrew]

I think that in general the statement should be that "all other factors being equal...larger sensors are capable of providing higher image quality"

The 'all other factors' takes into account pixel count, lens, processing, AA filters etc.

The 'are capable of' takes into account that you might wish to throw away that additional quality with a reduced quality output.

[AstroDave]

Well, for starters, how do we (you!) define "image quality"?? You take a crack at this, but some folks might be happy with subjective quality (how it looks to them) while others might want to be more objective (with my technical background, I probably fall into this category).

Here are some technical areas, generally subject to actual measurements: More dynamic range? More pixels? Better high ISO performance (=? less noise)? Truer color fidelity? (How do you check that!?) Wider color gamut?

[pschlute]

Full Frame..... Medium Format..... pfffffff

Attached Images

 

[photoptimist]

Good topic. Although it is generally true that bigger is generally better for IQ, you are right that it does depend on some variables:

Output media limitations: Bigger is still better even if the output medium limits final IQ. It's just that there is the matter of diminishing improvements for larger sensors or that the IQ improvements from more real estate only show up under the most difficult conditions (e.g., dark scenes requiring ISO 51,200).

Sensor performance: Clearly yes! The newest small sensors have better IQ than the oldest large sensors. IQ is strongly affected by SNR (Signal to noise). Bigger sensors collect a lot more signal (better SNR). But older sensors have much worse noise (worse SNR).

DoF at FOV: Tricky! In theory all DoF levels are possible with all FOVs and all sensor sizes. In practice, the shallowest DoFs are only available for the most popular sensor sizes and FOVs (e.g, not large format and not the smallest formats). Also, in practice, the deepest DoFs may not be available for some sensor sizes and FOVs because the available lenses only go to f/16 or f/22 (but might need f/64 to match small sensor DoF).

ISO level: Hmmm. I suppose that at ISO 100 and a low-contrast scene, the advantages of larger sensors might be minimal. But either high contrast scenes or high ISO scenes clearly benefit from more real estate.

Frame-rate: On the other hand, smaller is better if perceptions of final picture IQ depends on frame-rate such as when capturing unpredictable transient situations At any given level of sensor technology development, smaller sensors are faster to read out.

[boriscleto]

Full Frame..... Medium Format..... pfffffff

[swanlefitte]

Does a symbol need higher quality?
Is this 'a' higher quality than this 'A'?
From the purely defined symbol there is no difference.
So the argument for real estate is better must come from additional information. If and only if more information is desirable then bigger is better.
One could argue that when function is defined there is only compromise when that function can't be met 100% such as a symbol.
When an image is for more than function then any additional information can be said to be better. Then everything is a compromise because perhaps waiting for better equipment to be invented before taking the image vs dieing is already a compromise.

[GUB]

For a comparison where a given lens being used - ie a FF Pentax lens used on apsc vs FF then the answer is absolutely yes in terms of presented resolution.

[selfnoise]

[boriscleto]

But the donut bokeh...

[RGlasel]

I don't think it is possible to look at the components of image quality in isolation from the image itself. Sometimes a photograph is appealing because it is not perfect. How can an aspect of image quality be evaluated across different shooting circumstances, different display media and different photographic objectives? It's more than just the right tool for the task, different photographers get better or worse results with the same equipment. It would make more sense to discuss the PA (photographic aptitude) of different photographers; generally speaking the very best photographers are capable of getting excellent results in a variety of shooting circumstances, with a variety of photographic objectives and they tend to stick with the tools they are comfortable with as opposed to trying different equipment for every different situation and every different objective. At least you can say with confidence that a high PA photographer will get better results in most situations than a low PA photographer.

Even if we narrow down our focus to a single factor, like pixel density or resolution, if a camera that scores better in that factor is compromised in its ergonomics or selection of compatible lenses, that camera will not take better photographs, except in situations that are too limited to be of interest to anyone. If someone limits themselves to only taking photographs that maximize the IQ of their camera(s), they are a technician, not a photographer. Within very broad limits, sensor size is arbitrary, everything else is designed to be compatible with a specific sensor size (or maybe a couple of closely related sensor sizes, but the principle is still the same), the sensor size isn't designed to be compatible with a particular focal length or specific camera dimensions.

[fehknt]

I believe it does matter, because physics. Photographic techniques convert light into a recording medium, and every recording medium I know of has a "saturation" of too much light where it can't record any more.

So, my basic reasoning is that _if_ you are limited in your imaging by noise at base ISO (e.g., you have plenty of light to work with or can increase it in a studio setting or something, and are viewing at an enlargement factor that makes noise problematic for you) then the increased sensor area can allow for more electrons to be captured providing a better overall SNR.

I think the reasoning that gets confused a lot is "limited in imaging by noise at base ISO" -- and the expectation that somehow lenses can be perfectly matched across formats. So I think what is often compared is something like, an 85/1.4 35mm vs. a similar (in terms of min DoF) 120/2 (made up numbers, should be about right) on a larger format -- will have same DoF and total light gathering capability wide open though the 120/2 image will be dimmer of course -- you'd have to apply one stop of gain in post and exactly give up the one stop of "better noise performance" unless you also changed the shutter speed to be longer. Now, with that shutter speed the smaller format would "blow out" and lose information while the larger format, spreading that light out over more area, can capture all of it. However, modern techniques such as Pixel Shift and HDR or multi-frame image stacking etc. can all be used to capture more data over more time with the same imaging system at the cost of needing more computation/time. With pixel-shift, you gain a stop of data at the cost of artifacts, which would be rendered as blur in the "normal" exposure of equal brightness and information (given same image rendering choices) you could do with a larger format. Or, if you want to change DoF, maybe there's a larger format lens that has a wider equiv. aperture and you could get more light per time captured keeping the shutter speed the same.

The basic principle boils down to: more area = more recording capacity per shot. But you don't need more recording capacity if you're not able to fill up what you have already! (I.e., the way you shoot involves shooting above base ISO.)

One of the reasons I believe film can be superior to digital is related to sensor size, but the other is related to film speed -- you can get some low speed film that has exceptionally low grain/noise (mmmm, velvia 50), which will saturate slower than digital sensors (base ~100 usually).

I think a lot of the perception comes from the early days of computer aided design, before the crazy types of and purity of glass, etc. In this context, where lens flaws also get enlarged, lower enlargement ratios would be far superior particularly on the wide angles where the light simply has to bend less for larger formats to get the same FoV. With the amazing modern giant hunks of glass we have, I think a most of the problems with smaller formats have been reduced significantly.


... Finally, I want to just say that understanding the technical aspects of it probably just make my life worse because it's easy to focus on and matters like 1% for creating better images at this point in my photography. If you're happy with the images you create in the form you choose to display them... what else really matters? It sounds like from your context paragraph, that your current systems are plenty for your choices. I found that 24x30 enlargements from APSC were a little iffy and I wanted to move to FF. Anything beyond that is fun for me but not what I consider "production" because I can't get modern glass for it (and the digital 645 systems don't represent what seems like a good tradeoff on my shooting which is largely non-studio based).

(My personal context: I started on 35mm film, went to APSC digital, then moved to FF on the K-1, added the pentax 6x7, and am considering getting into 4x5 or 8x10 but daaaang things get expensive and heavy.)

[biz-engineer]

The larger the sensor is the better the image quality, when excluding usability constrains such as depth of field, shutter speed, real lens and available light. Shooting conditions (hand held, available light, max lens aperture), define when it is possible to obtain the better IQ potential of the larger sensor. Shake reduction (SR for Pentax, VR for Nikon etc) for non moving subjects, does shift the shooting triangle so that more of the IQ potential of a larger sensor can be used. SR on full frame makes full frame compete with medium format.

---------- Post added 13-08-19 at 07:26 ----------

ISO level: Hmmm. I suppose that at ISO 100 and a low-contrast scene, the advantages of larger sensors might be minimal. But either high contrast scenes or high ISO scenes clearly benefit from more real estate.

It can be the opposite of what you are saying. It is only true to some extent that the larger sensor compensate IQ for a higher ISO, noise gets averaged across more sensor area/pixels count, but for digital sensor the true dynamic range of the A/D conversion gets reduced by higher ISO level regardless of sensor size. The IQ improvement a larger sensor goes down to zero when the larger sensor is constrained by shooting conditions. At ISO100, both small and large sensors deliver 14bits deep color information, so the larger sensor deliver better overall image IQ. The user don't see a difference in IQ between an image from a larger sensor and an image from a larger sensor because the enlargement is insufficient to appreciate the difference.

[fehknt]

I think we're actually saying the same thing even though you sound like you're disagreeing. If you keep FoV, DoF and shutter speed the same (so all exposure params except ISO), then the larger format will have a dimmer image because the light is spread over a wider area and you must compensate by increasing ISO/gain. If you can increase the light intensity recorded somehow so you can run the sensor at ISO100 again then you're recording more light, so you get a better SNR of course! If you took the same larger format and tried to condense it onto a smaller area, you'd find now that you're already at base ISO and you blow the highlights.

So, yes, we both agree that IFF you can run the sensors both at base ISO, then the larger one will do better. I carry it one step further to try to argue that for an equiv image, above base ISO you'll have to be at a higher ISO on the larger sensor by exactly the difference in sensor size, so there's no benefit anymore in terms of noise.

[biz-engineer]

If you keep FoV, DoF and shutter speed the same (so all exposure params except ISO), then the larger format will have a dimmer image because the light is spread over a wider area and you must compensate by increasing ISO/gain.

That's all about what we call "equivalence". But a lab / studio environment is not the same as a travel environment. When traveling, the smaller equivalent format is more convenient to use, hand held, and it will produce images that are almost as good as the same images that would be produced with a larger sensor. That one thing tells me that I got it wrong with selecting a single full frame system to cover all needs, I'd have been better served by a compact apsc camera dedicated to travel by plane and casual / street shooting, and a medium format system with carrying tripod and lenses in a car to go on landscape location. Also, apsc mirrorless makes a lot of sense for travel kits, and reducing the system of medium format camera via mirrorless doesn't make sense at all when you need a large tripod anyway or be in a studio. Equivalence is a thing, usability is another. The camera market is a kind of Chinese buffer, all kind of systems are offered but only some of them make sense for use cases.