Effect of different size sensors
Jun 24, 2019 14:34:29 #
I am unclear about what the effect is of different size sensors and am hoping someone can explain it to me. I own two cameras - Sony a7r2 with which I use Sony G and G master lenses and the Sony Rx10 miv which has a fixed 24-600 Zeiss lens. The A7r2 has a ff sensor and produces a file size of 42mb in ff mode and a 20mb file in APS-C mode. The rx10 has a one-inch sensor and produces a 20mb file. So, comparing the a7r2 in APS-C mode to the RX10 what difference would the sensor make? I don’t see a difference, but I haven’t tried printing any of the images in huge sizes
Jun 24, 2019 14:40:17 #
Good question. I'm interested in seeing the answer to this.
Jun 24, 2019 15:23:47 #
It is misleading to compare sensor size with file size. Most cameras can produce jpg files of various sizes, selected in its menu by the user. What is more relavent is the actual size of the sensor, FF, APS-C, 1", etc. This indicates the physical size of the sensor. A handy chart is here:
https://www.ephotozine.com/article/sensor-size-and-crop-factor-explained-with-sample-photos-17813
In addition to physical size, the number of (mega)pixels is important. That is the number of light converting elements that are on the sensor.
A FF and an APS-C may both have 24MP, an array of 6000 by 4000 individual pixels. The difference is the pixels on the FF sensor would be much larger, (864/329* (for Canon))=2.6 times larger! Larger pixels can collect more light so are more sensitive in low light situations. It also means there would be less noise for any given ISO setting.
On the other hand, a FF sensor may use smaller pixels and so have many more than the APS-C sensor, but of similar size as the APS-C. In such a case, it would have much more detail or resolution, but similar light sensitivity. Greater pixel count can give greater file size photos.
Or the camera designer could make a FF sensor having more and somewhat larger pixels giving a lesser advantage in both resolution and sensitivity.
* Area in mm² of sensor, See linked chart
https://www.ephotozine.com/article/sensor-size-and-crop-factor-explained-with-sample-photos-17813
In addition to physical size, the number of (mega)pixels is important. That is the number of light converting elements that are on the sensor.
A FF and an APS-C may both have 24MP, an array of 6000 by 4000 individual pixels. The difference is the pixels on the FF sensor would be much larger, (864/329* (for Canon))=2.6 times larger! Larger pixels can collect more light so are more sensitive in low light situations. It also means there would be less noise for any given ISO setting.
On the other hand, a FF sensor may use smaller pixels and so have many more than the APS-C sensor, but of similar size as the APS-C. In such a case, it would have much more detail or resolution, but similar light sensitivity. Greater pixel count can give greater file size photos.
Or the camera designer could make a FF sensor having more and somewhat larger pixels giving a lesser advantage in both resolution and sensitivity.
* Area in mm² of sensor, See linked chart
Jun 24, 2019 15:25:40 #
In general the larger sensor would have less noise when compared to the smaller sensor shot at the same ISO. The larger sensor camera also has less depth of field for a given aperture and angle of view. Since the 24-600mm specified for your RX10 is really 35mm equivalent and not actual focal length you can test and see when you zoom the RX10 and the A7r2 to give you the same view and shot at the same aperture the depth of field for the RX10 would be significantly greater.
Jun 24, 2019 15:38:01 #
dfrost01 wrote:
...... Sony a7r2 with which I use Sony G and G master lenses and the Sony Rx10 miv which has a fixed 24-600 Zeiss lens. The A7r2 has a ff sensor and produces a file size of 42mb in ff mode and a 20mb file in APS-C mode. The rx10 has a one-inch sensor and produces a 20mb file......
That seems like a pretty basic question for someone with such sophisticated (and expensive) equipment. Did you inherit it from a rich uncle?
Generally, and in simple terms bigger sensor means better low light performance (ISO), bigger prints, shallower DoF for given aperture and focal length, more weight and more money.
Smaller sensor usually means lighter, less money (though not necessarily) and narrower field of vision for any given focal length -which some folks interpret as greater 'reach'. Also, I think it is harder to design good quality ultra wide angle fast lenses for smaller sensors. (I could be wrong - it may be the exact opposite)
Jun 24, 2019 18:53:07 #
a6k
Loc: Detroit & Sanibel
I hope that I am not adding to the confusion, but the physical dimensions of the sensor are a strong influence on image quality but not determinative.
The pixel density (pixels per square mm) or, more importantly, its reciprocal (pixel size in microns) is determinative within the constraints of the manufacturer's technology. The optimal pixel size with current technology is still roughly between 4 and 6 microns. Since almost all the sensors are using the same light-sensitive chemistry, there is no giant difference that would change this simple truth.
The A7R2 has a pixel size of 4.5 microns. The RX10 m3 or m4 has a pixel size of 2.4 microns. Since they are both Sony's new or newest sensor technology, the A7R2 has considerably better dynamic range and image quality overall. Dynamic range is largely a function of pixel size (more size=more range) but the explanation is TMI for this response.
The second feature, (again, influenced by but not strictly controlled by physical dimensions) is pixels horizontal and vertical which does determine the total mega pixels. The A7R2 and A7R3 have 7952 x 5304 pixels. The RX10 m3 or m4 have 5472 x 3648. You didn't mention the crop frame a6xxx series, but they are in between on both pixel size and dimensions at 6000 x 4000.
Ignoring sharpness, if you have more pixels then you can make a bigger print at the same DPI (dots per inch) which for prints up to perhaps 11x14 or even 16x20 would usually be 300 DPI but that is not a hard/fast rule. So if all else were equal, the A7R2 could make a bigger print and in most cases, lens dependent, a somewhat better one.
Example: 7952/300 = 26.5 inches. 5472/300= 18.2 inches. The way I'm equating pixels in the image file to dots per inch in the print has its critics but I think it is a practical, simple way to look at it. The use of 300 DPI is very flexible, of course. In fact, although you will usually not see commercial printers use a higher number, some will use a lower number and without it being noticed. BUT if you make a smaller print than those sizes they are down-sampling which simply means combining pixels via some program's logic. That can mean that if you make a 4x6 print or a 5x7 print, just for example, you have even less difference between the cameras' final product. Something similar happens when you look at them on a 1920x1200 screen; the picture is compressed to fit.
There are other Sony full frame camera with larger pixels and thus fewer megapixels which are thought of as better for low light. Bigger pixels are better for low light.
This is not a slam on the RX10m3 or m4. I had the m3 and now have the m4. They are excellent cameras and the RX10 is my go-to camera more often than the a6500. I have used the A7R2 and some of my highest quality pictures came from it. But I found it too heavy, etc.
The poster who pointed out that the RX10m3 or m4 is actually a 220 mm maximum length lens is correct but it has the same angle of view as an A7R2 with a 600 mm lens.
Now, one more thing. The A7R2 can allow more and better cropping and that can be an important aspect of wildlife and other kinds of photography when your lens just wasn't long enough.
BTW - I did not include file size in this because it is not directly in control of any of this and especially with JPG (vs raw) it's much less important.
BTW#2: I can't see any value at all to using APS-C mode on the A7R2; you'd just be throwing away pixels.
The pixel density (pixels per square mm) or, more importantly, its reciprocal (pixel size in microns) is determinative within the constraints of the manufacturer's technology. The optimal pixel size with current technology is still roughly between 4 and 6 microns. Since almost all the sensors are using the same light-sensitive chemistry, there is no giant difference that would change this simple truth.
The A7R2 has a pixel size of 4.5 microns. The RX10 m3 or m4 has a pixel size of 2.4 microns. Since they are both Sony's new or newest sensor technology, the A7R2 has considerably better dynamic range and image quality overall. Dynamic range is largely a function of pixel size (more size=more range) but the explanation is TMI for this response.
The second feature, (again, influenced by but not strictly controlled by physical dimensions) is pixels horizontal and vertical which does determine the total mega pixels. The A7R2 and A7R3 have 7952 x 5304 pixels. The RX10 m3 or m4 have 5472 x 3648. You didn't mention the crop frame a6xxx series, but they are in between on both pixel size and dimensions at 6000 x 4000.
Ignoring sharpness, if you have more pixels then you can make a bigger print at the same DPI (dots per inch) which for prints up to perhaps 11x14 or even 16x20 would usually be 300 DPI but that is not a hard/fast rule. So if all else were equal, the A7R2 could make a bigger print and in most cases, lens dependent, a somewhat better one.
Example: 7952/300 = 26.5 inches. 5472/300= 18.2 inches. The way I'm equating pixels in the image file to dots per inch in the print has its critics but I think it is a practical, simple way to look at it. The use of 300 DPI is very flexible, of course. In fact, although you will usually not see commercial printers use a higher number, some will use a lower number and without it being noticed. BUT if you make a smaller print than those sizes they are down-sampling which simply means combining pixels via some program's logic. That can mean that if you make a 4x6 print or a 5x7 print, just for example, you have even less difference between the cameras' final product. Something similar happens when you look at them on a 1920x1200 screen; the picture is compressed to fit.
There are other Sony full frame camera with larger pixels and thus fewer megapixels which are thought of as better for low light. Bigger pixels are better for low light.
This is not a slam on the RX10m3 or m4. I had the m3 and now have the m4. They are excellent cameras and the RX10 is my go-to camera more often than the a6500. I have used the A7R2 and some of my highest quality pictures came from it. But I found it too heavy, etc.
The poster who pointed out that the RX10m3 or m4 is actually a 220 mm maximum length lens is correct but it has the same angle of view as an A7R2 with a 600 mm lens.
Now, one more thing. The A7R2 can allow more and better cropping and that can be an important aspect of wildlife and other kinds of photography when your lens just wasn't long enough.
BTW - I did not include file size in this because it is not directly in control of any of this and especially with JPG (vs raw) it's much less important.
BTW#2: I can't see any value at all to using APS-C mode on the A7R2; you'd just be throwing away pixels.
Jun 24, 2019 21:05:03 #
dfrost01 wrote:
I am unclear about what the effect is of different... (show quote)
There are a lot of factors that affect image quality, but in general, the larger sensor, even if it has lots of very tiny pixels (like a Canon 5DS R or a Nikon D850, will give you better looking images. If you take an uncropped image from a 20mp 1" sensor ca,era and compare it to a 45-50 mp camera, two things will immediately be apparent - the higher pixel count will record more detail, and the fact that you have to magnify the 1" sensor considerably more than the full frame camera, will give the edge to the full frame camera.
It's an urban myth to think that you can't make really good prints from tiny sensors - Apple makes billboard-sized images from it's iPhone cameras, even the iPhone 6 which was only 8 mp. More pixels generally means greater detail capture and more flexibility when cropping - not bigger prints.
Jun 25, 2019 08:41:25 #
Glad you mentioned pixel density, which, as it relates to manufacturing technology is at a limit. The next limit, which we have started seeing discussed about high density sensors, is the physical diffraction limit. That is, as the pixel size gets to be a lower multiple of the wavelength of light, sharpness will be lost.
Without going into the mechanics of diffraction, you can see it’s effects shown as the loss of sharpness of a shadow The farther away it is from the object.
Without going into the mechanics of diffraction, you can see it’s effects shown as the loss of sharpness of a shadow The farther away it is from the object.
Jun 25, 2019 09:41:44 #
We all know a fact, larger pixels are capable of gathering light better while offering better resolution. The more pixels the smaller their size. The Nikon D3 or D700 with only 12 megapixels are capable of outstanding 20x30 inch enlargements. I had a Nikon D2X with only 4.1 megapixels and instead the enlargements were of first quality when shot in good light.
The smaller sensor in a point and shoot camera could be cramped with lots of pixels but that is not in favor of a better resolution. When you look at the images in your monitor they do not seem to have a big difference but they do when you print the files.
The smaller sensor in a point and shoot camera could be cramped with lots of pixels but that is not in favor of a better resolution. When you look at the images in your monitor they do not seem to have a big difference but they do when you print the files.
Jun 25, 2019 09:42:16 #
Have read many good articles on the relationship between sensor size and various aspects of image quality — from detail captured to noise-handling at high iso. What I would like to see and understand more is the relationship between those IQ aspects and the significant advances in sensor technology over the last 10+ years. How can that element be quantified and compared? By observation from my personal use of FF Nikon from the D3 through the D810, and comparing the results with my current M4/3’s Olympus, I can easily see the advances in sensor processing technology. But, its impact on IQ seems to often be overlooked or minimized in the obsession over sheer sensor size.
Jun 25, 2019 09:46:36 #
I am in agreement with William. Advances in sensor technologies are allowing us to get much better files from smaller sensors.
Jun 25, 2019 12:06:11 #
I looked at the diagram in the guide to sensor size and pixel size. Resolution increases as pixel size dropsand light collection is directly proportional to sensor size. Is there a sweet spot where you get maximum number of pixels with maximum size on a sensor?
Jun 25, 2019 12:55:12 #
fetzler
Loc: North West PA
The resulting file size depends on the compression and the number of megapixels on the camera. Because compression is very important in determining the file size it is somewhat difficult to correlate file size with the number of megapixels. Low levels of compression are better. File size is mostly not worth thinking about unless you are having computer performance issues.
On might consider the pixels on the sensor to be little photon counting buckets. The repeatability of the count depends on how many photons are counted. For a given light intensity larger pixels will count more photons. This bit of physics translates into larger pixels giving lower noise. A 24 MP FF camera will give lower noise that a 24MP micro 4/3 camera (I don't think these exist now)
The technology employed is also important. Newer sensors are generally better.
If one is interested in landscape photography, for instance, one might well consider medium format. This choice is expensive, however.
There are applications that favor smaller sensors. It is important to note that a 24 MP APS-C camera and a 16 MP micro 4/3 camera have the same size pixels ( indeed 20MP micro 4/3 pixels are close enough). Noise differences in these cameras is not great.
If a smaller sensor is used then less magnification is required to fill the frame of the sensor. Less magnification translates into greater DOF for given camera setting. Macro photography favors smaller sensors. Medium and Large format photography of small objects is very difficult or impossible.
Finally the more MP that you have, the more you have to work with for cropping and the production of large prints. For prints that are for close up viewing 300 pixels per inch are enough. For video display and for large prints viewed at a distance the number of pixels per inch can be considerably less.
On might consider the pixels on the sensor to be little photon counting buckets. The repeatability of the count depends on how many photons are counted. For a given light intensity larger pixels will count more photons. This bit of physics translates into larger pixels giving lower noise. A 24 MP FF camera will give lower noise that a 24MP micro 4/3 camera (I don't think these exist now)
The technology employed is also important. Newer sensors are generally better.
If one is interested in landscape photography, for instance, one might well consider medium format. This choice is expensive, however.
There are applications that favor smaller sensors. It is important to note that a 24 MP APS-C camera and a 16 MP micro 4/3 camera have the same size pixels ( indeed 20MP micro 4/3 pixels are close enough). Noise differences in these cameras is not great.
If a smaller sensor is used then less magnification is required to fill the frame of the sensor. Less magnification translates into greater DOF for given camera setting. Macro photography favors smaller sensors. Medium and Large format photography of small objects is very difficult or impossible.
Finally the more MP that you have, the more you have to work with for cropping and the production of large prints. For prints that are for close up viewing 300 pixels per inch are enough. For video display and for large prints viewed at a distance the number of pixels per inch can be considerably less.
Jun 25, 2019 13:38:45 #
amfoto1
Loc: San Jose, Calif. USA
Lots of good info above...
Something else you might note: Your RX10 actually doesn't have a "24-600mm" lens. The actual focal length range of it's lens is 8.8mm to 220mm.
It is common practice with cameras of that type (non-interchangeble lens) to refer to "35mm film equivalent" focal lengths. This is done in order to compare various cameras, because they use a lot of different size sensors in those cameras.
Thanks to it's moderate size sensor, the RX10's 8.8-220mm lens "acts like a 24-600mm lens" would on your a7RII (which has a sensor making images the same size as "35mm film").
The 1" sensor used in your RX10 is actually one of the larger type for this kind of camera. However, it's quite a bit smaller than an APS-C sensor... and a whole lot smaller than the "full frame" sensor in your a7RII. But it's considerably bigger than what's used in many other non-interchangeable lens digital that claim even more extreme zoom ranges. For example, the Nikon P900 uses a tiny sensor so that it's 4.3-357mm lens will "act like 24-2000mm, 35mm equivalent".
The actual dimensions of a 1" sensor are: 8.8x13.2mm... or 116 square mm area. They fit 20 million pixels onto that sensor, which means there are over 172,000 pixel sites per square mm. (The Nikon P900 mentioned above has a 16MP 1/2.3" sensor that has about 30 sq. mm area. This means it has well over 500,000 pixels per sq. mm!)
In comparison, your a7RII's "full frame" sensor is 24x36mm and 864 sq. mm. With 46 million pixels, this works out to about 53,000 pixels per sq. mm. (When set to APS-C crop, the camera only uses approx. 40% of it's sensor area, but the density of pixel sites per sq. mm remain the same.)
In other words, the a7RII's sensor is far, far less crowded. The RX10 has more than 3X the density of pixel sites. So the a7RII's individual pixels can be bigger, which makes them better gathering light. They also can have more space between them, which reduces heat build up and minimizes "cross talk" between them, both of which are leading causes of digital image "noise".
As a result, in comparison to the RX10, your a7RII will capture more detail and can be used at much higher ISO settings before noise becomes a problem. I.e., you can shoot in much lower light conditions with the a7RII. The images produced by the a7RII also will be much more enlargeable, should you wish to print them. The a7RII images also will tolerate more cropping, if needed.
Something else you might note: Your RX10 actually doesn't have a "24-600mm" lens. The actual focal length range of it's lens is 8.8mm to 220mm.
It is common practice with cameras of that type (non-interchangeble lens) to refer to "35mm film equivalent" focal lengths. This is done in order to compare various cameras, because they use a lot of different size sensors in those cameras.
Thanks to it's moderate size sensor, the RX10's 8.8-220mm lens "acts like a 24-600mm lens" would on your a7RII (which has a sensor making images the same size as "35mm film").
The 1" sensor used in your RX10 is actually one of the larger type for this kind of camera. However, it's quite a bit smaller than an APS-C sensor... and a whole lot smaller than the "full frame" sensor in your a7RII. But it's considerably bigger than what's used in many other non-interchangeable lens digital that claim even more extreme zoom ranges. For example, the Nikon P900 uses a tiny sensor so that it's 4.3-357mm lens will "act like 24-2000mm, 35mm equivalent".
The actual dimensions of a 1" sensor are: 8.8x13.2mm... or 116 square mm area. They fit 20 million pixels onto that sensor, which means there are over 172,000 pixel sites per square mm. (The Nikon P900 mentioned above has a 16MP 1/2.3" sensor that has about 30 sq. mm area. This means it has well over 500,000 pixels per sq. mm!)
In comparison, your a7RII's "full frame" sensor is 24x36mm and 864 sq. mm. With 46 million pixels, this works out to about 53,000 pixels per sq. mm. (When set to APS-C crop, the camera only uses approx. 40% of it's sensor area, but the density of pixel sites per sq. mm remain the same.)
In other words, the a7RII's sensor is far, far less crowded. The RX10 has more than 3X the density of pixel sites. So the a7RII's individual pixels can be bigger, which makes them better gathering light. They also can have more space between them, which reduces heat build up and minimizes "cross talk" between them, both of which are leading causes of digital image "noise".
As a result, in comparison to the RX10, your a7RII will capture more detail and can be used at much higher ISO settings before noise becomes a problem. I.e., you can shoot in much lower light conditions with the a7RII. The images produced by the a7RII also will be much more enlargeable, should you wish to print them. The a7RII images also will tolerate more cropping, if needed.
Jun 25, 2019 14:36:44 #
a6k
Loc: Detroit & Sanibel
amfoto1 wrote:
Lots of good info above...//snip//
As a result, in comparison to the RX10, your a7RII will capture more detail and can be used at much higher ISO settings before noise becomes a problem. I.e., you can shoot in much lower light conditions with the a7RII. The images produced by the a7RII also will be much more enlargeable, should you wish to print them. The a7RII images also will tolerate more cropping, if needed.
As a result, in comparison to the RX10, your a7RII will capture more detail and can be used at much higher ISO settings before noise becomes a problem. I.e., you can shoot in much lower light conditions with the a7RII. The images produced by the a7RII also will be much more enlargeable, should you wish to print them. The a7RII images also will tolerate more cropping, if needed.
True!❗️(and with a good explanation)
If you want to reply, then register here. Registration is free and your account is created instantly, so you can post right away.