Bill Pierce
Well-known
In your experience, how much does sensor size effect image quality? My limited experience contradicts the bigger is much better mantra. I was recently looking at some portraits taken on both APS-c and medium format. Yes, the medium format was “sharper,” but not by much. Certainly it was not something that would be noticed at all in normal viewing rather than my hyper critical close up view of eye lashes. These were portraits, not architecturals or landscapes, where, perhaps, the difference would make a difference. But the difference was amazingly slight.
I have spend a great deal of time sorting through processing programs and tif conversions to maximize the technical quality of my images. When I am producing the most detailed and sharp images I can from Fuji APS-c and Leica full frame files the APS-c files are sharper than the full frame files.
Of course, there is certainly a lot outside of sensor size that effects image quality. (But who wants to admit that they are often a bigger problem than their camera?) But what does your experience tell you? What are the advantages of large sensors, and what are the advantages of small sensors? How important overall is sensor size in your photography?
I have spend a great deal of time sorting through processing programs and tif conversions to maximize the technical quality of my images. When I am producing the most detailed and sharp images I can from Fuji APS-c and Leica full frame files the APS-c files are sharper than the full frame files.
Of course, there is certainly a lot outside of sensor size that effects image quality. (But who wants to admit that they are often a bigger problem than their camera?) But what does your experience tell you? What are the advantages of large sensors, and what are the advantages of small sensors? How important overall is sensor size in your photography?
xayraa33
rangefinder user and fancier
Even photos taken by the Micro Four Thirds system trump the quality of 35mm film cameras of yore using medium iso speed film in quality in large print enlargements.
goamules
Well-known
Are you talking digital, and looking at screens?
Because in my experience MF is QUITE a bit better in a large print than 35mm. As in anyone would notice the difference. And a large format contact print can be spectacular. But if you're just talking pixels and monitors and phones, I don't know.
Because in my experience MF is QUITE a bit better in a large print than 35mm. As in anyone would notice the difference. And a large format contact print can be spectacular. But if you're just talking pixels and monitors and phones, I don't know.
bmattock
Veteran
I like 35mm. I accept APS-C. But I love 6x9. Nothing beats real estate for me.
Pioneer
Veteran
Ignoring film for the moment, if you are interested in printing big then you are better served with a bigger sensor.
However, if you are perfectly happy with 8x10 or whole plate contact prints (Edward Weston style) then you are well served with just about any digital sensor currently manufactured.
Sharpness is a function of software when dealing with digital sensors and is really not the primary reason to use a larger sensor.
From the film side I think the discussion regarding the value of size has pretty well been settled way before I was even born.
However, if you are perfectly happy with 8x10 or whole plate contact prints (Edward Weston style) then you are well served with just about any digital sensor currently manufactured.
Sharpness is a function of software when dealing with digital sensors and is really not the primary reason to use a larger sensor.
From the film side I think the discussion regarding the value of size has pretty well been settled way before I was even born.
Ccoppola82
Well-known
Speaking of digital, I have a FF canon and a aps-c fuji. Image quality on both is excellent. The main thing that bothers me about aps-c is that a 23mm (35mm equivalent) at close distance shows massive distortion of a portrait etc. I find FF much better for that sort of thing. My son had a MASSIVE ear in a couple shots the other day on the Fuji that he would not have had on my canon FF.
RichC
Well-known
Significant - if you make large prints, which I do.
I rest my case: https://blog.mingthein.com/2013/06/25/a-question-of-sensor-size/
(Compare the vertical pairs of images, looking at the large images on Flickr)
The larger the sensor, the greater the tonal subtlety and variation. Sharpness ... you're picking on the least important factor here!
I rest my case: https://blog.mingthein.com/2013/06/25/a-question-of-sensor-size/
(Compare the vertical pairs of images, looking at the large images on Flickr)
The larger the sensor, the greater the tonal subtlety and variation. Sharpness ... you're picking on the least important factor here!
Ronald M
Veteran
Are you testing us?
MORE TO IT THAN sensor size, but all things being equal, bigger is better.
Equal means same print size over 10x, equal pixel size per unit of señsor area.
Take a PJ camera like Nikon D5. Smaller # of pixels , around 20 million, but they are large and capture more light so low light performance is improved. Less pixels compared to D850 which will make a less noisy large print because it has 45 million pixels. At higher ISO the 850 will be more noisy.
Exactly where in print size these things show and at what iso would take a book and lots of testing. I believe I have seen test show the current best crop dslr will make sharper prints that a D5 because smaller pixels will hold detail better up to some size.
Now 20 or 24 MP camera, one crop, one FF, and if they both have same AA filter, the crop will be better given same size print. Go up the iso scale, things may change.
MORE TO IT THAN sensor size, but all things being equal, bigger is better.
Equal means same print size over 10x, equal pixel size per unit of señsor area.
Take a PJ camera like Nikon D5. Smaller # of pixels , around 20 million, but they are large and capture more light so low light performance is improved. Less pixels compared to D850 which will make a less noisy large print because it has 45 million pixels. At higher ISO the 850 will be more noisy.
Exactly where in print size these things show and at what iso would take a book and lots of testing. I believe I have seen test show the current best crop dslr will make sharper prints that a D5 because smaller pixels will hold detail better up to some size.
Now 20 or 24 MP camera, one crop, one FF, and if they both have same AA filter, the crop will be better given same size print. Go up the iso scale, things may change.
Bill Pierce
Well-known
Rich -
You make an important point. It’s easy to assign numbers to some of the contributors to sharpness and difficult to quantify the improvement in tonality. I wish it were otherwise because this limits our discussion about what is certainly one of the important reasons to look at larger pixels. Sadly, web jpgs are not much of a demonstration of either sharpness or tonality.
Looking at big prints from earlier days of digital even in prints from full frame sensors you see the battle between detail and tonal scale. You and I probably disagree because I think there are full frame cameras today that capture fine detail in large prints. But we do agree that there is a tonal difference. Even if there were not so many different digital “medium formats” it would be difficult to pin down the subtle advantages of bigger pixels. I hope there is some very wise person on the forum that can explain this to us. We need a WISE PERSON ALERT.
(BTW, I think Ming has left Blad and is shooting Nikon. It's a crazy world.
farlymac
PF McFarland
Going from a 1/1.7 to an APS-C it was easy to tell the difference. Now if I had an FX camera, I don't know as if it would be as easy to tell between it and an APS-C DX.
PF
PF
pgk
Well-known
I and other photographers I talk to, who use 35mm and smaller formats, mostly agree that the smaller format images are less 'satisfying' in technical terms than the full frame. I don't think that this can easily be determined in terms of simple resolution figures or tonality, but I do think that all the numerous minor differences can cascade together and the visually educated observer of prints from these cameras may often appreciate the sum of the differences - not always though.
Dogman
Veteran
I've never shot full frame. APS-C was the format I started digital using--a Canon 30D. I now shoot Fujis and a Ricoh GRII, still APS-C, and an Olympus Micro 4/3 on occasions. There have also been a few pocket cameras with smaller formats.
What I have noticed is that image quality has improved markedly in the formats I've used over the years. I was an early adopter of the Micro 4/3 format and I also used Olympus's standard 4/3 DSLRs. Comparing those early 4/3 images to more recent ones, it's obvious the more recent photos are sharper with better tonality and lower noise. Comparing the early APS-C to more recent ones shows the same improvements. Some of this has to do with lenses. I'm using better lenses now than I did early on. Fuji and the Ricoh's lenses, as well as the PanaLeica M4/3 lenses I now use have the sharpest, most satisfying look to them of anything I've ever used. Today, I can see little difference between recent M4/3 and APS-C when optics are not included in the equation.
Full frame is tempting but I can't justify it. My largest prints are cropped images on 13x19 and even the older digital images print fine at this size. I shoot most of my photos in daylight or under bright artificial light so noise and dynamic range is not really an issue for me. Anyway, I used to shoot for a daily newspaper so I know what pushed Tri-X looks like and it can't compare in quality to high ISO digital. I don't think full frame offers me much improvement.
What I have noticed is that image quality has improved markedly in the formats I've used over the years. I was an early adopter of the Micro 4/3 format and I also used Olympus's standard 4/3 DSLRs. Comparing those early 4/3 images to more recent ones, it's obvious the more recent photos are sharper with better tonality and lower noise. Comparing the early APS-C to more recent ones shows the same improvements. Some of this has to do with lenses. I'm using better lenses now than I did early on. Fuji and the Ricoh's lenses, as well as the PanaLeica M4/3 lenses I now use have the sharpest, most satisfying look to them of anything I've ever used. Today, I can see little difference between recent M4/3 and APS-C when optics are not included in the equation.
Full frame is tempting but I can't justify it. My largest prints are cropped images on 13x19 and even the older digital images print fine at this size. I shoot most of my photos in daylight or under bright artificial light so noise and dynamic range is not really an issue for me. Anyway, I used to shoot for a daily newspaper so I know what pushed Tri-X looks like and it can't compare in quality to high ISO digital. I don't think full frame offers me much improvement.
ChrisPlatt
Thread Killer
Spoiled by classic 35mm gear
Spoiled by classic 35mm gear
I have tried various types of medium format cameras.
Image quality was slightly better for some types of subjects, under certain conditions.
However not one of them handled nearly as well as any 35mm camera I have owned.
IMO the tradeoff is just not worth it. Kick me if I ever buy another medium format rig.
Chris
Spoiled by classic 35mm gear
I have tried various types of medium format cameras.
Image quality was slightly better for some types of subjects, under certain conditions.
However not one of them handled nearly as well as any 35mm camera I have owned.
IMO the tradeoff is just not worth it. Kick me if I ever buy another medium format rig.
Chris
RichC
Well-known
I'd love to buy a medium-format camera, say a 100-150 MP Phase One. Sadly, I don't $50,000 spare...! Affordable medium-format cameras like the Pentax 645Z don't show sufficient increase in image quality to make buying them worthwhile and lose the convenience of 35mm systems (smaller, lighter, cheaper, massive lens choice, etc.).
I currently make do with a Sony A7R II (and before that a Nikon D800E) - both full frame and about 40 MP.
As I mentioned, I print large, and (a) I cannot make do with less than 40 MP and (b) - more germane to this thread - the difference between sensor sizes makes a clear and noticeable difference (smaller formats look "mushy" and "smeary"). This is not to do with sharpness per se but how well sensors resolve detail and tone.
This is an example of the size prints I make on occasion - this isn't mine, but I've been commissioned to make a print to fill this massive window next year!
People will stick their noses into large prints, so it's crucial it looks good at both inches and yards. Full-frame 35mm barely holds up at these sizes regardless of megapixels...
I currently make do with a Sony A7R II (and before that a Nikon D800E) - both full frame and about 40 MP.
As I mentioned, I print large, and (a) I cannot make do with less than 40 MP and (b) - more germane to this thread - the difference between sensor sizes makes a clear and noticeable difference (smaller formats look "mushy" and "smeary"). This is not to do with sharpness per se but how well sensors resolve detail and tone.
This is an example of the size prints I make on occasion - this isn't mine, but I've been commissioned to make a print to fill this massive window next year!
People will stick their noses into large prints, so it's crucial it looks good at both inches and yards. Full-frame 35mm barely holds up at these sizes regardless of megapixels...
Axel
singleshooter
I have a bunch of good pictures that are taken with small sensors.
And there are a lot of pictures that sadly were not taken because of none or a wrong camera.
Small sensors of today are mostly very good so there is no problem with printing or
publishing in digital galleries.
All planned or professional pictures can benefit from a larger sensor. Following the rule
"the bigger the better" can´t be wrong here.
And I am tired from bean counter discussions like "1" vs Fourthirds" or sth. else.
Assuming that every sensor needs a camera there are much more interesting facts
for me than sensorsize before it comes to purchase.
And there are a lot of pictures that sadly were not taken because of none or a wrong camera.
Small sensors of today are mostly very good so there is no problem with printing or
publishing in digital galleries.
All planned or professional pictures can benefit from a larger sensor. Following the rule
"the bigger the better" can´t be wrong here.
And I am tired from bean counter discussions like "1" vs Fourthirds" or sth. else.
Assuming that every sensor needs a camera there are much more interesting facts
for me than sensorsize before it comes to purchase.
taemo
eat sleep shoot
This thread is quite fitting to what I'm trying to decide right now.
Presently I am shooting with a Sony A7 24MP (full frame) and been debating to upgrade to the Fuji GFX 50mp (medium format).
Reason for this is because I want to print big and want as much detail as I can and by going MF I wouldn't need to stitch as much.
Take for example this 15 photo stitch taken with Leica M240 24mp giving me 100mp
The disadvantage with going MF though is the initial cost and size of the camera and lenses.
Ultimately the final print will dictate the format that you will be using.
From what I've learned from others is that if you are just going to stare on a phone/computer screen and unless you plan to print bigger than 32x24 and close viewing, FF is more than capable.
Realistically, 24mp FF or APS-C is more than sufficient for my need.
For example this was taken with a Fuji APS-C 24mp and printed 32x24
And this was taken with a Sony FF 24mp cropped down to 12mp and printed 48x36
I shoot quite a bit with my Iphone as well and have no complaint since I know that I wouldn't need to print these bigger than 8x10.
Presently I am shooting with a Sony A7 24MP (full frame) and been debating to upgrade to the Fuji GFX 50mp (medium format).
Reason for this is because I want to print big and want as much detail as I can and by going MF I wouldn't need to stitch as much.
Take for example this 15 photo stitch taken with Leica M240 24mp giving me 100mp
The disadvantage with going MF though is the initial cost and size of the camera and lenses.
Ultimately the final print will dictate the format that you will be using.
From what I've learned from others is that if you are just going to stare on a phone/computer screen and unless you plan to print bigger than 32x24 and close viewing, FF is more than capable.
Realistically, 24mp FF or APS-C is more than sufficient for my need.
For example this was taken with a Fuji APS-C 24mp and printed 32x24
And this was taken with a Sony FF 24mp cropped down to 12mp and printed 48x36
I shoot quite a bit with my Iphone as well and have no complaint since I know that I wouldn't need to print these bigger than 8x10.
charjohncarter
Veteran
Slightly off the question, I have a 36 mp ff camera that is new to me. This camera has 'Pixel Shift' so being new I wanted to try the new stuff. I shot a couple and really couldn't see any difference with the final jpegs.
I decided to read up on the post processing to see if I was doing something wrong: I was. First, you have to be on a solid tripod (and I mean solid like a tripod on cement), then you have to use mirror up with a remote trigger, it is recommended that you take at least five shots to pick the best one. After the shooting you HAVE to convert the RAW file using an editing program that will do pixel shift; until just recently Adobe didn't and Lightroom doesn't (but maybe they've added it now). So I used RAW Therapee selecting pixel shift and saving in TIFF, as when it is put in jpeg, compression loses the benefit.
So, at least for me it is a toy, a product photographer in a ground floor studio may use it. But I have to say the IQ is amazing but like Bill says not enough to go crazy over.
EDIT: taemo, your middle shot above is just great.
I decided to read up on the post processing to see if I was doing something wrong: I was. First, you have to be on a solid tripod (and I mean solid like a tripod on cement), then you have to use mirror up with a remote trigger, it is recommended that you take at least five shots to pick the best one. After the shooting you HAVE to convert the RAW file using an editing program that will do pixel shift; until just recently Adobe didn't and Lightroom doesn't (but maybe they've added it now). So I used RAW Therapee selecting pixel shift and saving in TIFF, as when it is put in jpeg, compression loses the benefit.
So, at least for me it is a toy, a product photographer in a ground floor studio may use it. But I have to say the IQ is amazing but like Bill says not enough to go crazy over.
EDIT: taemo, your middle shot above is just great.
ptpdprinter
Veteran
From the digital side too. Most of the discussion has to do with minimum necessary rather than maximum quality, which is why print size is always the qualifier.
willie_901
Veteran
Nothing Beats Lens And, Or Sensor Surface Area
Nothing Beats Lens And, Or Sensor Surface Area
It is best to think about the lens and sensor together. The surface area for both affect the maximum possible signal level. In comparisons where lens differences (aperture and T factors) are irrelevant, then sensor area alone affects the maximum attainable image quality.
...
As sensor area increases, the maximum possible signal level increases.
In terms of technical image quality (where composition, creativity, use of light and many other human factors are ignored) sensor surface area determines the maximum signal-to-noise ratio for the data. The signal contains the information we need (the spatial illuminance for the scene). The noise determines the uncertainty (error levels) for that information.
Whenever sensor underexposure is unavoidable, SNR makes a difference. The most obvious impact is on shadow region detail. Of course, at low ambient light levels, sensor underexposure reaches levels where even highlight regions are have low SNRs. With newer cameras most of this noise is photon noise.
In bright scenes SNR is also important. Analog dynamic range is directly determined by SNR. SNR advantage that may be obvious in low light is also relevant for the brightest scenes.
Small sensors mean smaller and lighter cameras. In some circumstances this is a significant advantage. The extreme examples are phone cameras.
But smaller sensors do not necessarily mean smaller and lighter lenses. Increasing sensor area and increasing aperture area both increase signal levels.
So, signal levels can be identical for a large sensor with a small maximum apertures lens and a small sensor and a large maximum aperture lens.The size and weight advantage of smaller sensor cameras can be offset by the size and weight of larger lenses.
But maximum possible lens apertures have practical limitations. When DOF is not considered, large sensors will have superior SNRs.
Right now another advantage of smaller sensors is cost. A mirrorless medium format camera (44 x 33 mm sensor) is more expensive than many 24 X 36 mm cameras. The Phase 1 IQ3 has 54 x 40 mm sensor. This is the highest SNR and dynamic range still camera one can buy. It also costs $50K.
...
It is very important.
I own APS-C cameras because the compromise between maximum SNR and convenience (size, weight, cost and OVF with RF style composition) is perfect for my interests.
If I was interested in landscape photography I would own a mirrorless medium format camera.
If I was interested in action photography I would own a camera with a 24 X 36 mm sensor.
In 2008 I started to use APS-C DSLRs. When I switched from a D200 to a D300 I noticed a significant difference in technical IQ due to SNR differences. Then I switched to a D700. Again the improvement in technical IQ was obvious.
I noticed similar technical IQ differences between the X100, XT-1/X-100T and the X-Pro 2. Here the difference SNR is due to improvements in sensor assembly electronics.
When I first started using the XT-1 I discovered it slightly out-performed the D700 at base ISO. At the time analog dynamic range at base ISO and shadow region IQ was very important. The D700's surface area advantage was offset by the improvements in sensor noise levels.
Nothing Beats Lens And, Or Sensor Surface Area
It is best to think about the lens and sensor together. The surface area for both affect the maximum possible signal level. In comparisons where lens differences (aperture and T factors) are irrelevant, then sensor area alone affects the maximum attainable image quality.
...
As sensor area increases, the maximum possible signal level increases.
In terms of technical image quality (where composition, creativity, use of light and many other human factors are ignored) sensor surface area determines the maximum signal-to-noise ratio for the data. The signal contains the information we need (the spatial illuminance for the scene). The noise determines the uncertainty (error levels) for that information.
- More surface area means more signal
- More signal means lower relative levels of photon (shot) noise
- For contemporary platforms, our camera's electronic noise levels are similar. Increasing sensor area does not necessarily increase electronic noise levels
Whenever sensor underexposure is unavoidable, SNR makes a difference. The most obvious impact is on shadow region detail. Of course, at low ambient light levels, sensor underexposure reaches levels where even highlight regions are have low SNRs. With newer cameras most of this noise is photon noise.
In bright scenes SNR is also important. Analog dynamic range is directly determined by SNR. SNR advantage that may be obvious in low light is also relevant for the brightest scenes.
Small sensors mean smaller and lighter cameras. In some circumstances this is a significant advantage. The extreme examples are phone cameras.
But smaller sensors do not necessarily mean smaller and lighter lenses. Increasing sensor area and increasing aperture area both increase signal levels.
So, signal levels can be identical for a large sensor with a small maximum apertures lens and a small sensor and a large maximum aperture lens.The size and weight advantage of smaller sensor cameras can be offset by the size and weight of larger lenses.
But maximum possible lens apertures have practical limitations. When DOF is not considered, large sensors will have superior SNRs.
Right now another advantage of smaller sensors is cost. A mirrorless medium format camera (44 x 33 mm sensor) is more expensive than many 24 X 36 mm cameras. The Phase 1 IQ3 has 54 x 40 mm sensor. This is the highest SNR and dynamic range still camera one can buy. It also costs $50K.
...
It is very important.
I own APS-C cameras because the compromise between maximum SNR and convenience (size, weight, cost and OVF with RF style composition) is perfect for my interests.
If I was interested in landscape photography I would own a mirrorless medium format camera.
If I was interested in action photography I would own a camera with a 24 X 36 mm sensor.
In 2008 I started to use APS-C DSLRs. When I switched from a D200 to a D300 I noticed a significant difference in technical IQ due to SNR differences. Then I switched to a D700. Again the improvement in technical IQ was obvious.
I noticed similar technical IQ differences between the X100, XT-1/X-100T and the X-Pro 2. Here the difference SNR is due to improvements in sensor assembly electronics.
When I first started using the XT-1 I discovered it slightly out-performed the D700 at base ISO. At the time analog dynamic range at base ISO and shadow region IQ was very important. The D700's surface area advantage was offset by the improvements in sensor noise levels.
Bill Pierce
Well-known
William - Thank you.