This topic is in the 2009 archive. You can read it and like its posts here. It is read only.
Archive 2009the d300s killer
p.11 #1the d300s killer
Kerry Pierce wrote:
It depends mainly on how hard you look for it and what level of deterioration that you find acceptable. The bottom line is that the higher the density, the faster diffraction will bite you. I like that calculator and have used it for years. Look at the VISUAL EXAMPLE: APERTURE VS. PIXEL SIZE and compare the grids generated. It's very easy to see the problem generated with smaller pixels. It's unfortunate that he hasn't updated that to show the newer sensors or allow for custom input for pixel size.
Kerry -
I think it's more accurate to say: The higher the pixel density, the sooner you notice the diffraction. Again, there's not much downside to high pixel density cameras in re: to diffraction (or even noise, but that's another argument,
) . Your final photo will never be worse than it was with a lower MP camera. From the site everyone is linking to:
Are smaller pixels somehow worse? Not necessarily. Just because the diffraction limit has been reached with large pixels does not mean the final photo will be any worse than if there were instead smaller pixels and the limit was surpassed;
So - When your sensor doesn't show diffraction due to large pixels, it's because it doesn't have enough resolution to. So, you aren't getting as much detail in the first place.
The summation goes on to say:
Both scenarios still have the same total resolution (although one will produce a larger file). Even though the resolution is the same, the camera with the smaller pixels will render the photo with fewer artifacts (such as color moiré and aliasing). Smaller pixels also provide the flexibility of having better resolution with larger apertures, in situations where the depth of field can be more shallow. When other factors such as noise and depth of field are considered, the answer as to which is better becomes more complicated.
So, even though you may end up with the same res photo, you get less "junk" in it with smaller pixels.
But the main thing is that a 12MP camera at F16 - despite diffraction - is never going to give you MORE detail than a 24Mp camera @ F16. (and I can say for sure 24MP actaully has a LOT more detail)
Sure, I could be wrong about all this, but these are the facts as I understand them. Always willing to learn, so fire away
It depends mainly on how hard you look for it and what level of deterioration that you find acceptable. The bottom line is that the higher the density, the faster diffraction will bite you. I like that calculator and have used it for years. Look at the VISUAL EXAMPLE: APERTURE VS. PIXEL SIZE and compare the grids generated. It's very easy to see the problem generated with smaller pixels. It's unfortunate that he hasn't updated that to show the newer sensors or allow for custom input for pixel size.
Kerry -
I think it's more accurate to say: The higher the pixel density, the sooner you notice the diffraction. Again, there's not much downside to high pixel density cameras in re: to diffraction (or even noise, but that's another argument,
) . Your final photo will never be worse than it was with a lower MP camera. From the site everyone is linking to:Are smaller pixels somehow worse? Not necessarily. Just because the diffraction limit has been reached with large pixels does not mean the final photo will be any worse than if there were instead smaller pixels and the limit was surpassed;
So - When your sensor doesn't show diffraction due to large pixels, it's because it doesn't have enough resolution to. So, you aren't getting as much detail in the first place.
The summation goes on to say:
Both scenarios still have the same total resolution (although one will produce a larger file). Even though the resolution is the same, the camera with the smaller pixels will render the photo with fewer artifacts (such as color moiré and aliasing). Smaller pixels also provide the flexibility of having better resolution with larger apertures, in situations where the depth of field can be more shallow. When other factors such as noise and depth of field are considered, the answer as to which is better becomes more complicated.
So, even though you may end up with the same res photo, you get less "junk" in it with smaller pixels.
But the main thing is that a 12MP camera at F16 - despite diffraction - is never going to give you MORE detail than a 24Mp camera @ F16. (and I can say for sure 24MP actaully has a LOT more detail)
Sure, I could be wrong about all this, but these are the facts as I understand them. Always willing to learn, so fire away
Sep 10, 2009 at 08:02 AM
p.11 #2the d300s killer
Steve Perry wrote:
Kerry -
I think it's more accurate to say: The higher the pixel density, the sooner you notice the diffraction. Again, there's not much downside to high pixel density cameras in re: to diffraction (or even noise, but that's another argument,) . Your final photo will never be worse than it was with a lower MP camera. From the site everyone is linking to:
Are smaller pixels somehow worse? Not necessarily. Just because the diffraction limit has been reached with large pixels does not mean the final photo will be any worse than if there were instead smaller pixels
So - When your sensor doesn't show diffraction due to large pixels, it's because it doesn't have enough resolution to. So, you aren't getting as much detail in the first place.
The summation goes on to say:
Both scenarios still have the same total resolution (although one will produce a larger file). Even though the resolution is the same, the camera with the smaller pixels will render the photo with fewer artifacts (such as color moiré and aliasing). Smaller pixels also provide the flexibility of having better resolution with larger apertures, in situations where the depth of field can be more shallow. When other factors such as noise and depth of field are considered, the answer as to which is better becomes more complicated.
So, even though you may end up with the same res photo, you get less "junk" in it with smaller pixels.
But the main thing is that a 12MP camera at F16 - despite diffraction - is never going to give you MORE detail than a 24Mp camera @ F16. (and I can say for sure 24MP actaully has a LOT more detail)
Sure, I could be wrong about all this, but these are the facts as I understand them. Always willing to learn, so fire away
...Show more →
You are correct from a resolution standpoint. However, noise may be an issue even if you downres the image. In general, for similar technology nodes, the same sensor size, and assuming you do not push the pixels too small (i.e. a factor of 2x or 3x at most) noise performance will be about the same after properly downresing the image. BUT, the sensor with less MP will have better DR, exposure latitude and less color reproduction issues (assuming all else is the same).
Kerry Pierce wrote
It depends mainly on how hard you look for it and what level of deterioration that you find acceptable. The bottom line is that the higher the density, the faster diffraction will bite you. I like that calculator and have used it for years. Look at the VISUAL EXAMPLE: APERTURE VS. PIXEL SIZE and compare the grids generated. It's very easy to see the problem generated with smaller pixels. It's unfortunate that he hasn't updated that to show the newer sensors or allow for custom input for pixel size.I think it's more accurate to say: The higher the pixel density, the sooner you notice the diffraction. Again, there's not much downside to high pixel density cameras in re: to diffraction (or even noise, but that's another argument,
) . Your final photo will never be worse than it was with a lower MP camera. From the site everyone is linking to:Are smaller pixels somehow worse? Not necessarily. Just because the diffraction limit has been reached with large pixels does not mean the final photo will be any worse than if there were instead smaller pixels
So - When your sensor doesn't show diffraction due to large pixels, it's because it doesn't have enough resolution to. So, you aren't getting as much detail in the first place.
The summation goes on to say:
Both scenarios still have the same total resolution (although one will produce a larger file). Even though the resolution is the same, the camera with the smaller pixels will render the photo with fewer artifacts (such as color moiré and aliasing). Smaller pixels also provide the flexibility of having better resolution with larger apertures, in situations where the depth of field can be more shallow. When other factors such as noise and depth of field are considered, the answer as to which is better becomes more complicated.
So, even though you may end up with the same res photo, you get less "junk" in it with smaller pixels.
But the main thing is that a 12MP camera at F16 - despite diffraction - is never going to give you MORE detail than a 24Mp camera @ F16. (and I can say for sure 24MP actaully has a LOT more detail)
Sure, I could be wrong about all this, but these are the facts as I understand them. Always willing to learn, so fire away
...Show more →
You are correct from a resolution standpoint. However, noise may be an issue even if you downres the image. In general, for similar technology nodes, the same sensor size, and assuming you do not push the pixels too small (i.e. a factor of 2x or 3x at most) noise performance will be about the same after properly downresing the image. BUT, the sensor with less MP will have better DR, exposure latitude and less color reproduction issues (assuming all else is the same).
Sep 10, 2009 at 08:10 AM
p.11 #3the d300s killer
Kerry Pierce wrote:
It depends mainly on how hard you look for it and what level of deterioration that you find acceptable. The bottom line is that the higher the density, the faster diffraction will bite you. I like that calculator and have used it for years. Look at the VISUAL EXAMPLE: APERTURE VS. PIXEL SIZE and compare the grids generated. It's very easy to see the problem generated with smaller pixels. It's unfortunate that he hasn't updated that to show the newer sensors or allow for custom input for pixel size.
This juvenile fanboy crap is just so much tiresome nonsense...Show more →
I'm aware that diffraction is an issue on higher pixel density, but to say it happens at f/8 is an error.
Finally, I don't appreciate being called a juvenile fanboy. Its not in the spirit of the Fred Miranda forums. We ALL try to help one another. My point being that a "cheap" Canon FF was introduced in the fall on 2005 and it was disparaged in the Nikon forum for a number of reasons, many valid, but both Canon & Nikon overcame issues like vignetting, and everyone today loves FF & FX cameras, well at least most, and issues with them are overlooked or overcome today. We will see high pixel density DX cameras from Nikon next year. I was simply drawing an analogy.
rsg_1 wrote
There are some other variables in place, like the Anti-Aliasing filter that also affects diffraction. By using the calculator above in the link I provided, in a worst case scenario, something which has the 7D's pixel density will see diffraction at about f/16.Earlier quote
What will really be interesting is all the same people clamoring about diffraction will be singing the praises of a Nikon D400 when it appears next year with a 16-18 MP sensor. It reminds me of all the "tsk-tsk when will people learn that a full frame DSLR doesn't work" type talk that emerged right here on this forum about 4 years ago with the introduction of the 5D.I'm aware that diffraction is an issue on higher pixel density, but to say it happens at f/8 is an error.
Finally, I don't appreciate being called a juvenile fanboy. Its not in the spirit of the Fred Miranda forums. We ALL try to help one another. My point being that a "cheap" Canon FF was introduced in the fall on 2005 and it was disparaged in the Nikon forum for a number of reasons, many valid, but both Canon & Nikon overcame issues like vignetting, and everyone today loves FF & FX cameras, well at least most, and issues with them are overlooked or overcome today. We will see high pixel density DX cameras from Nikon next year. I was simply drawing an analogy.
Sep 10, 2009 at 09:50 AM
p.11 #4the d300s killer
Andre Labonte wrote:
You are correct from a resolution standpoint. However, noise may be an issue even if you downres the image. In general, for similar technology nodes, the same sensor size, and assuming you do not push the pixels too small (i.e. a factor of 2x or 3x at most) noise performance will be about the same after properly downresing the image. BUT, the sensor with less MP will have better DR, exposure latitude and less color reproduction issues (assuming all else is the same).
Andre,
The argument is that if you can improve the quantum efficiency of the photosites, the same ISO with the same exposure, when compared to lower density sensor, will not reduce DR or deteriorate the IQ. How true this is will be found out at the end of this month. However, initial samples are very promising.
You are correct from a resolution standpoint. However, noise may be an issue even if you downres the image. In general, for similar technology nodes, the same sensor size, and assuming you do not push the pixels too small (i.e. a factor of 2x or 3x at most) noise performance will be about the same after properly downresing the image. BUT, the sensor with less MP will have better DR, exposure latitude and less color reproduction issues (assuming all else is the same).
Andre,
The argument is that if you can improve the quantum efficiency of the photosites, the same ISO with the same exposure, when compared to lower density sensor, will not reduce DR or deteriorate the IQ. How true this is will be found out at the end of this month. However, initial samples are very promising.
Sep 10, 2009 at 09:59 AM
p.11 #5the d300s killer
Steve Perry wrote:
Kerry -
I think it's more accurate to say: The higher the pixel density, the sooner you notice the diffraction. Again, there's not much downside to high pixel density cameras in re: to diffraction (or even noise, but that's another argument,) . Your final photo will never be worse than it was with a lower MP camera.
I don't see how that is meaningful for anyone. So, they spend a year or 2 developing a new sensor that isn't any worse than the one it replaces? It has some minor advantages, but is that what we really want, to pay for smoke, mirrors and minor advantages? Personally, I want to see significant, tangible benefits from the sensor in the next camera that I buy, just like I saw with my last 2 purchases. Otherwise, what is the point in expending all of that R&D on a new sensor? Just use the old one and pass along the savings to us.
But the main thing is that a 12MP camera at F16 - despite diffraction - is never going to give you MORE detail than a 24Mp camera @ F16. (and I can say for sure 24MP actaully has a LOT more detail)
Well, you can't say for sure that a 24MP DX camera will have a LOT more detail. You're looking at the d3x 24mp, which has less pixel density than the d300 and has a diffraction threshold comparable to the d200. That's not what we're discussing here.
We'll soon know, how well the 7d performs at smaller apertures and when diffraction will be biting the landscape, stock and other photographers that need to use small apertures.
I'm all for more MPs, when appropriate. FX has room to grow and I'm pretty sure that the d4 will have more MPs without losing much, if any noise performance. I am not at all convinced that DX has much room to grow, or that it really needs to.
Kerry -
I think it's more accurate to say: The higher the pixel density, the sooner you notice the diffraction. Again, there's not much downside to high pixel density cameras in re: to diffraction (or even noise, but that's another argument,
) . Your final photo will never be worse than it was with a lower MP camera. I don't see how that is meaningful for anyone. So, they spend a year or 2 developing a new sensor that isn't any worse than the one it replaces? It has some minor advantages, but is that what we really want, to pay for smoke, mirrors and minor advantages? Personally, I want to see significant, tangible benefits from the sensor in the next camera that I buy, just like I saw with my last 2 purchases. Otherwise, what is the point in expending all of that R&D on a new sensor? Just use the old one and pass along the savings to us.
But the main thing is that a 12MP camera at F16 - despite diffraction - is never going to give you MORE detail than a 24Mp camera @ F16. (and I can say for sure 24MP actaully has a LOT more detail)
Well, you can't say for sure that a 24MP DX camera will have a LOT more detail. You're looking at the d3x 24mp, which has less pixel density than the d300 and has a diffraction threshold comparable to the d200. That's not what we're discussing here.
We'll soon know, how well the 7d performs at smaller apertures and when diffraction will be biting the landscape, stock and other photographers that need to use small apertures.
I'm all for more MPs, when appropriate. FX has room to grow and I'm pretty sure that the d4 will have more MPs without losing much, if any noise performance. I am not at all convinced that DX has much room to grow, or that it really needs to.
Sep 10, 2009 at 10:47 AM
p.11 #6the d300s killer
rsg_1 wrote:
I'm aware that diffraction is an issue on higher pixel density, but to say it happens at f/8 is an error.
I think that the point was that the 7d must have diffraction affects sooner than sensors with less pixel density, specifically the d300, thus it would be closer to f/8 than the d300 is. The d300 is commonly believed to be f/11 max for avoiding diffraction issues.
Finally, I don't appreciate being called a juvenile fanboy. Its not in the spirit of the Fred Miranda forums. We ALL try to help one another.
The quoted paragraph below is what I don't appreciate. It adds absolutely nothing to the discussion and certainly is not in your "spirit of FM".
What will really be interesting is all the same people clamoring about diffraction will be singing the praises of a Nikon D400 when it appears next year with a 16-18 MP sensor. It reminds me of all the "tsk-tsk when will people learn that a full frame DSLR doesn't work" type talk that emerged right here on this forum about 4 years ago with the introduction of the 5D.
I'm aware that diffraction is an issue on higher pixel density, but to say it happens at f/8 is an error.
I think that the point was that the 7d must have diffraction affects sooner than sensors with less pixel density, specifically the d300, thus it would be closer to f/8 than the d300 is. The d300 is commonly believed to be f/11 max for avoiding diffraction issues.
Finally, I don't appreciate being called a juvenile fanboy. Its not in the spirit of the Fred Miranda forums. We ALL try to help one another.
The quoted paragraph below is what I don't appreciate. It adds absolutely nothing to the discussion and certainly is not in your "spirit of FM".
What will really be interesting is all the same people clamoring about diffraction will be singing the praises of a Nikon D400 when it appears next year with a 16-18 MP sensor. It reminds me of all the "tsk-tsk when will people learn that a full frame DSLR doesn't work" type talk that emerged right here on this forum about 4 years ago with the introduction of the 5D.
Sep 10, 2009 at 11:00 AM
p.11 #7the d300s killer
rsg_1 wrote:
Andre,
The argument is that if you can improve the quantum efficiency of the photosites, the same ISO with the same exposure, when compared to lower density sensor, will not reduce DR or deteriorate the IQ. How true this is will be found out at the end of this month. However, initial samples are very promising.
DR with smaller pixels at the same tech node will reduce DR but not necessarily increase noise. This is where exposure accuracy comes into play. On the other hand, a more advanced tech node might actually give better noise performance despite having smaller pixels. BUT, remember that this is NOT a linear relationship and there ARE REAL PHYSICAL LIMITS that will be reached. In other words, advancing the technology will only improve things to a point. At some point, the pixel size will be the determining factor.
Also, with regards to this statement by you in your response to Kerry:
"I'm aware that diffraction is an issue on higher pixel density, but to say it happens at f/8 is an error."
In fact, it does start to show up as early as f/5 for red light and f/9 for blue light for the 7D.
Andre Labonte wrote
You are correct from a resolution standpoint. However, noise may be an issue even if you downres the image. In general, for similar technology nodes, the same sensor size, and assuming you do not push the pixels too small (i.e. a factor of 2x or 3x at most) noise performance will be about the same after properly downresing the image. BUT, the sensor with less MP will have better DR, exposure latitude and less color reproduction issues (assuming all else is the same).The argument is that if you can improve the quantum efficiency of the photosites, the same ISO with the same exposure, when compared to lower density sensor, will not reduce DR or deteriorate the IQ. How true this is will be found out at the end of this month. However, initial samples are very promising.
DR with smaller pixels at the same tech node will reduce DR but not necessarily increase noise. This is where exposure accuracy comes into play. On the other hand, a more advanced tech node might actually give better noise performance despite having smaller pixels. BUT, remember that this is NOT a linear relationship and there ARE REAL PHYSICAL LIMITS that will be reached. In other words, advancing the technology will only improve things to a point. At some point, the pixel size will be the determining factor.
Also, with regards to this statement by you in your response to Kerry:
"I'm aware that diffraction is an issue on higher pixel density, but to say it happens at f/8 is an error."
In fact, it does start to show up as early as f/5 for red light and f/9 for blue light for the 7D.
Sep 10, 2009 at 11:28 AM
p.11 #8the d300s killer
Kerry Pierce wrote:
I think that the point was that the 7d must have diffraction affects sooner than sensors with less pixel density, specifically the d300, thus it would be closer to f/8 than the d300 is. The d300 is commonly believed to be f/11 max for avoiding diffraction issues.
The quoted paragraph below is what I don't appreciate. It adds absolutely nothing to the discussion and certainly is not in your "spirit of FM".
Sir,
If you are offended, then I apologize; however, I did not resort to name calling.
rsg_1 wrote
I'm aware that diffraction is an issue on higher pixel density, but to say it happens at f/8 is an error.Earlier quote
Finally, I don't appreciate being called a juvenile fanboy. Its not in the spirit of the Fred Miranda forums. We ALL try to help one another.Earlier quote
What will really be interesting is all the same people clamoring about diffraction will be singing the praises of a Nikon D400 when it appears next year with a 16-18 MP sensor. It reminds me of all the "tsk-tsk when will people learn that a full frame DSLR doesn't work" type talk that emerged right here on this forum about 4 years ago with the introduction of the 5D.Sir,
If you are offended, then I apologize; however, I did not resort to name calling.
Sep 10, 2009 at 11:46 AM
p.11 #9the d300s killer
men the only way to handle this now is with swords.
Sep 10, 2009 at 12:04 PM
p.11 #10the d300s killer
Andre
I think we do absolutely understand each other except in the terms of in the how and why. Your figures are absolutely correct - I'm not disputing this. You will notice a fine-contrast pixel-to-pixel degradation already at F8, even if very subtle.
I only objected to the use of the word "limit". F/8 is the aperture where diffraction gets "noticeable", not a hard "limit". To clarify - you don't hit a "hard limit" until diffraction has degraded the pixel-pixel contrast so bad that alternating black/white lines (1 pixel wide) are indistinguishable from a gray mass, and this means that Rayleigh must calculated to span almost two pixel-widths. 1.7 pixelwidths if you do the calculation on sineshaped input.
The practical example is that an 18MP picture at F/11 (well into diffraction affected territory) will still contain a lot more detail than the same F/11 picture taken with a 12MP camera.
I think we do absolutely understand each other except in the terms of in the how and why. Your figures are absolutely correct - I'm not disputing this. You will notice a fine-contrast pixel-to-pixel degradation already at F8, even if very subtle.
I only objected to the use of the word "limit". F/8 is the aperture where diffraction gets "noticeable", not a hard "limit". To clarify - you don't hit a "hard limit" until diffraction has degraded the pixel-pixel contrast so bad that alternating black/white lines (1 pixel wide) are indistinguishable from a gray mass, and this means that Rayleigh must calculated to span almost two pixel-widths. 1.7 pixelwidths if you do the calculation on sineshaped input.
The practical example is that an 18MP picture at F/11 (well into diffraction affected territory) will still contain a lot more detail than the same F/11 picture taken with a 12MP camera.
Sep 10, 2009 at 12:58 PM
p.11 #11the d300s killer
theSuede wrote:
The practical example is that an 18MP picture at F/11 (well into diffraction affected territory) will still contain a lot more detail than the same F/11 picture taken with a 12MP camera.
If that turns out to be true, then a lot of people will be happy, including me. Assuming, of course, that your definition of the term "a lot more detail" is the same as or at least close to mine. 
The practical example is that an 18MP picture at F/11 (well into diffraction affected territory) will still contain a lot more detail than the same F/11 picture taken with a 12MP camera.
If that turns out to be true, then a lot of people will be happy, including me. Assuming, of course, that your definition of the term "a lot more detail" is the same as or at least close to mine.
Sep 10, 2009 at 01:15 PM
p.11 #12the d300s killer
Kerry Pierce wrote:
I think it's more accurate to say: The higher the pixel density, the sooner you notice the diffraction. Again, there's not much downside to high pixel density cameras in re: to diffraction (or even noise, but that's another argument,) . Your final photo will never be worse than it was with a lower MP camera.I don't see how that is meaningful for anyone. So, they spend a year or 2 developing a new sensor that isn't any worse than the one it replaces? It has some minor advantages, but is that what we really want, to pay for smoke, mirrors and minor advantages? Personally, I want to see significant, tangible benefits from the sensor in the next camera that I buy, just like I saw with my last 2 purchases. Otherwise, what is the point in expending all of that R&D on a new sensor? Just use the old one and pass along...Show more →
I mostly agree.
My point is that I don't think we should let diffraction scare us off - until it gets to the point where it really is a wash - you've captured all the detail you can with the system, so more MP just because you can is a waste. I don't think we're there yet.
I'm also excited about a few more MP where I can get them just because of the striking difference in detail between my old D3 and new D3x files. When I edit old D3 files, I'm like the digital photo version of the old Burger King lady asking, "Where's the beef?" - or in this case detail. No matter what F stop, the D3x has more detail, even though it hits diffraction limits sooner.
So, I for one, want the manufactures to push the MP limit to the limit - and absolutely no further and ONLY if it doesn't come at the expense of quality. 18MP might be about the top end for a DX size sensor, we'll have to see.
Steve Perry wrote
Kerry -I think it's more accurate to say: The higher the pixel density, the sooner you notice the diffraction. Again, there's not much downside to high pixel density cameras in re: to diffraction (or even noise, but that's another argument,
) . Your final photo will never be worse than it was with a lower MP camera.Earlier quote
But the main thing is that a 12MP camera at F16 - despite diffraction - is never going to give you MORE detail than a 24Mp camera @ F16. (and I can say for sure 24MP actaully has a LOT more detail)I mostly agree.
My point is that I don't think we should let diffraction scare us off - until it gets to the point where it really is a wash - you've captured all the detail you can with the system, so more MP just because you can is a waste. I don't think we're there yet.
I'm also excited about a few more MP where I can get them just because of the striking difference in detail between my old D3 and new D3x files. When I edit old D3 files, I'm like the digital photo version of the old Burger King lady asking, "Where's the beef?" - or in this case detail. No matter what F stop, the D3x has more detail, even though it hits diffraction limits sooner.
So, I for one, want the manufactures to push the MP limit to the limit - and absolutely no further and ONLY if it doesn't come at the expense of quality. 18MP might be about the top end for a DX size sensor, we'll have to see.
Sep 10, 2009 at 02:04 PM
p.11 #13the d300s killer
Please, don't quote my overstatements... :-) I might get embarrassed.
I really don't think of the difference between 15 and 18MP as even worthwhile in itself - if the 7D gave nothing more than just that into the equation "more cost - new camera" compared to the previous 50D. But it also includes a better overall noise into equation, new AF, new metering and so on. THAT is what makes the 7D interesting, not the 18MP.
I do think though that there is a difference between 12 and 18MP. That (a 50% increase of picture elements) is a noticeable difference in resolution. And in no way until now has it been worth it (noise-wise and so on) to go 12MP+ in a crop camera.
BTW - another example of diffraction "limited" and downsampled vs "not limited":
10MP APS-C vs 15MP APS-C at F/16. This should affect the 15MP camera (the 50D) a lot more than the 10MP - let's see what happens when we scale the 15MP camera down to match the 400D's 10MP size.
Oops. due to some limitations of the available raw-converters, it seems that the downsampled 18MP F/16 is actually sharper and contains less artefacts than the 10MP picture. Even if the 15MP camera is 22% further into the "diffraction limited" region.
Do note that 10/15 is the same relational difference as 12/18. Sorry about the large picture, but at least it isn't that "heavy" at ~300kb.
http://pici.se/pictures/FkoHPoLAZ.jpg
I really don't think of the difference between 15 and 18MP as even worthwhile in itself - if the 7D gave nothing more than just that into the equation "more cost - new camera" compared to the previous 50D. But it also includes a better overall noise into equation, new AF, new metering and so on. THAT is what makes the 7D interesting, not the 18MP.
I do think though that there is a difference between 12 and 18MP. That (a 50% increase of picture elements) is a noticeable difference in resolution. And in no way until now has it been worth it (noise-wise and so on) to go 12MP+ in a crop camera.
BTW - another example of diffraction "limited" and downsampled vs "not limited":
10MP APS-C vs 15MP APS-C at F/16. This should affect the 15MP camera (the 50D) a lot more than the 10MP - let's see what happens when we scale the 15MP camera down to match the 400D's 10MP size.
Oops. due to some limitations of the available raw-converters, it seems that the downsampled 18MP F/16 is actually sharper and contains less artefacts than the 10MP picture. Even if the 15MP camera is 22% further into the "diffraction limited" region.
Do note that 10/15 is the same relational difference as 12/18. Sorry about the large picture, but at least it isn't that "heavy" at ~300kb.
http://pici.se/pictures/FkoHPoLAZ.jpg
Sep 10, 2009 at 02:07 PM
p.11 #14the d300s killer
theSuede wrote:
Andre
I think we do absolutely understand each other except in the terms of in the how and why. Your figures are absolutely correct - I'm not disputing this. You will notice a fine-contrast pixel-to-pixel degradation already at F8, even if very subtle.
I only objected to the use of the word "limit". F/8 is the aperture where diffraction gets "noticeable", not a hard "limit". To clarify - you don't hit a "hard limit" until diffraction has degraded the pixel-pixel contrast so bad that alternating black/white lines (1 pixel wide) are indistinguishable from a gray mass, and this means that Rayleigh must...Show more →
Yep! We agree. In physics terms the diffraction "limit" is f/10 to f/18 depending of wavelength. In practical photographic terms, maximum sharpness is obtained at f/5 to f/9 depending of wavelength and assuming that the lens suffers a loss of sharpness due spherical aberation for wider aperatures ... which is typical.
Andre
I think we do absolutely understand each other except in the terms of in the how and why. Your figures are absolutely correct - I'm not disputing this. You will notice a fine-contrast pixel-to-pixel degradation already at F8, even if very subtle.
I only objected to the use of the word "limit". F/8 is the aperture where diffraction gets "noticeable", not a hard "limit". To clarify - you don't hit a "hard limit" until diffraction has degraded the pixel-pixel contrast so bad that alternating black/white lines (1 pixel wide) are indistinguishable from a gray mass, and this means that Rayleigh must...Show more →
Yep! We agree. In physics terms the diffraction "limit" is f/10 to f/18 depending of wavelength. In practical photographic terms, maximum sharpness is obtained at f/5 to f/9 depending of wavelength and assuming that the lens suffers a loss of sharpness due spherical aberation for wider aperatures ... which is typical.
Sep 10, 2009 at 03:29 PM
p.11 #15the d300s killer
Kerry Pierce wrote:
I think that the point was that the 7d must have diffraction affects sooner than sensors with less pixel density, specifically the d300, thus it would be closer to f/8 than the d300 is. The d300 is commonly believed to be f/11 max for avoiding diffraction issues.
For a given print size the 7D image will not be worse than for example a D300 image due to diffraction effects assuming the same aperture in both cases. At wider apertures you should get more detail from the 7D though. BUT if you pixel peep at 100%, then yes, diffraction softening will be seen sooner i.e. at a wider aperture on the 7D.
I feel this is all somewhat academic. Some lenses will not be up to the demands of the 7D sensor, some hard disks will start squealing from the size of the files, some PCs will run photoshop slowly when processing 7D files, but sometimes, just sometimes, the extra pixels will be useful, but only when doing big enlargements. Not having used the 7D in anger, I don't feel qualified to comment on whether or not the extra pixies are worth having.
I think that the point was that the 7d must have diffraction affects sooner than sensors with less pixel density, specifically the d300, thus it would be closer to f/8 than the d300 is. The d300 is commonly believed to be f/11 max for avoiding diffraction issues.
For a given print size the 7D image will not be worse than for example a D300 image due to diffraction effects assuming the same aperture in both cases. At wider apertures you should get more detail from the 7D though. BUT if you pixel peep at 100%, then yes, diffraction softening will be seen sooner i.e. at a wider aperture on the 7D.
I feel this is all somewhat academic. Some lenses will not be up to the demands of the 7D sensor, some hard disks will start squealing from the size of the files, some PCs will run photoshop slowly when processing 7D files, but sometimes, just sometimes, the extra pixels will be useful, but only when doing big enlargements. Not having used the 7D in anger, I don't feel qualified to comment on whether or not the extra pixies are worth having.
Sep 10, 2009 at 03:59 PM
p.11 #16the d300s killer
Steve Perry wrote:
I think it's more accurate to say: The higher the pixel density, the sooner you notice the diffraction. Again, there's not much downside to high pixel density cameras in re: to diffraction (or even noise, but that's another argument,) . Your final photo will never be worse than it was with a lower MP camera.I don't see how that is meaningful for anyone. So, they spend a year or 2 developing a new sensor that isn't any worse than the one it replaces? It has some minor advantages, but is that what we really want, to pay for smoke, mirrors and minor advantages? Personally, I want to see significant, tangible benefits from the sensor in the next camera that I buy, just like I saw with my last 2 purchases. Otherwise, what is the point in expending all of that R&D on a new sensor? Just use the old one and pass along the savings to us.
Well, you can't say for sure that a 24MP DX camera will have a LOT more detail. You're looking at the d3x 24mp, which has less pixel density than the d300 and has a diffraction threshold comparable to the d200. That's not what we're discussing here.
We'll soon know, how well the 7d performs at smaller apertures and when diffraction will be biting the landscape, stock and other photographers that need to use small apertures.
I'm all for more MPs, when appropriate. FX has room to grow and I'm pretty sure that the d4 will have more MPs without losing much, if any noise performance. I am not at all convinced that DX has much room to grow, or that it really needs to.I mostly agree.
My point is that I don't think we should let diffraction scare us off - until it gets to the point where it really is a wash - you've captured all the detail you can with the system, so more MP just because you can is a waste. I don't think we're there yet.
I'm also excited about a few more MP where I can get them just because of the striking difference in detail between my old D3 and new D3x files. When I edit old D3 files, I'm like the digital photo version of the...Show more →
Steve, I think the other thing to think about is the fact that DX or the similar Canon 1.6 sensor are closer to that limit than FX. In FX there is more room to grow the MP count without adverse effects at this time.
Kerry Pierce wrote
Steve Perry wrote
Kerry -I think it's more accurate to say: The higher the pixel density, the sooner you notice the diffraction. Again, there's not much downside to high pixel density cameras in re: to diffraction (or even noise, but that's another argument,
) . Your final photo will never be worse than it was with a lower MP camera.Earlier quote
But the main thing is that a 12MP camera at F16 - despite diffraction - is never going to give you MORE detail than a 24Mp camera @ F16. (and I can say for sure 24MP actaully has a LOT more detail)We'll soon know, how well the 7d performs at smaller apertures and when diffraction will be biting the landscape, stock and other photographers that need to use small apertures.
I'm all for more MPs, when appropriate. FX has room to grow and I'm pretty sure that the d4 will have more MPs without losing much, if any noise performance. I am not at all convinced that DX has much room to grow, or that it really needs to.
My point is that I don't think we should let diffraction scare us off - until it gets to the point where it really is a wash - you've captured all the detail you can with the system, so more MP just because you can is a waste. I don't think we're there yet.
I'm also excited about a few more MP where I can get them just because of the striking difference in detail between my old D3 and new D3x files. When I edit old D3 files, I'm like the digital photo version of the...Show more →
Steve, I think the other thing to think about is the fact that DX or the similar Canon 1.6 sensor are closer to that limit than FX. In FX there is more room to grow the MP count without adverse effects at this time.
Sep 10, 2009 at 04:41 PM
p.11 #17the d300s killer
Steve Perry wrote:
My point is that I don't think we should let diffraction scare us off -
I don't either. I'm not a single issue kinda guy. If I were an APS-C canon shooter, I'd be on the 7d waiting list right now. Not because of the MPs though....
until it gets to the point where it really is a wash - you've captured all the detail you can with the system, so more MP just because you can is a waste. I don't think we're there yet.
I don't know if we are or not. I do know that I wasn't impressed by the 50d resolution vs the d300. I'm more than happy to let the canon boys beta test this one.
The problem is that if nikon is following, we may never know if they could have given us another stop or more of real noise and DR improvement for DX. Personally, I think that issue is a lot more important to the DX masses than MPs will ever be. Few people print as large as you apparently do. But there are a ton of them that buy DX cameras and use slow consumer lenses to try to take photos of their kids in poor light.
Given a choice, I'd pay $1800 for a d300 body and all features with an 8mp sensor that topped the d3 for noise and DR before I'd pay $1500 for an 18mp d300.
I'm also excited about a few more MP where I can get them just because of the striking difference in detail between my old D3 and new D3x files.
Yes, but again that's an apples to oranges comparison. Going from 12 to 24 on FX without a big hit on noise or diffraction, is certainly a nice thing for folks that print large. We have yet to see if it's meaningful for DX.
So, I for one, want the manufactures to push the MP limit to the limit - and absolutely no further and ONLY if it doesn't come at the expense of quality. 18MP might be about the top end for a DX size sensor, we'll have to see.
I don't have any problem with that statement other than the fact that I hate paying for air sandwiches. They're not very filling.
My point is that I don't think we should let diffraction scare us off -
I don't either. I'm not a single issue kinda guy. If I were an APS-C canon shooter, I'd be on the 7d waiting list right now. Not because of the MPs though....
until it gets to the point where it really is a wash - you've captured all the detail you can with the system, so more MP just because you can is a waste. I don't think we're there yet.
I don't know if we are or not. I do know that I wasn't impressed by the 50d resolution vs the d300. I'm more than happy to let the canon boys beta test this one.
The problem is that if nikon is following, we may never know if they could have given us another stop or more of real noise and DR improvement for DX. Personally, I think that issue is a lot more important to the DX masses than MPs will ever be. Few people print as large as you apparently do. But there are a ton of them that buy DX cameras and use slow consumer lenses to try to take photos of their kids in poor light.
Given a choice, I'd pay $1800 for a d300 body and all features with an 8mp sensor that topped the d3 for noise and DR before I'd pay $1500 for an 18mp d300.
I'm also excited about a few more MP where I can get them just because of the striking difference in detail between my old D3 and new D3x files.
Yes, but again that's an apples to oranges comparison. Going from 12 to 24 on FX without a big hit on noise or diffraction, is certainly a nice thing for folks that print large. We have yet to see if it's meaningful for DX.
So, I for one, want the manufactures to push the MP limit to the limit - and absolutely no further and ONLY if it doesn't come at the expense of quality. 18MP might be about the top end for a DX size sensor, we'll have to see.
I don't have any problem with that statement other than the fact that I hate paying for air sandwiches. They're not very filling.
Sep 10, 2009 at 05:10 PM
p.11 #18the d300s killer
theSuede wrote:
I really don't think of the difference between 15 and 18MP as even worthwhile in itself - if the 7D gave nothing more than just that into the equation "more cost - new camera" compared to the previous 50D. But it also includes a better overall noise into equation, new AF, new metering and so on. THAT is what makes the 7D interesting, not the 18MP.
I agree completely.
BTW - another example of diffraction "limited" and downsampled vs "not limited":
Sorry, but I've no interest in the downsample, no worse type stuff. There's no way I'd buy a camera knowing that I had to downsample all of the images to get the performance that I wanted, even if it was relatively easy to do with a PS action.
Oops. due to some limitations of the available raw-converters, it seems that the downsampled 18MP F/16 is actually sharper and contains less artefacts than the 10MP picture. Even if the 15MP camera is 22% further into the "diffraction limited" region.
Are those 100% or higher crops?
I really don't think of the difference between 15 and 18MP as even worthwhile in itself - if the 7D gave nothing more than just that into the equation "more cost - new camera" compared to the previous 50D. But it also includes a better overall noise into equation, new AF, new metering and so on. THAT is what makes the 7D interesting, not the 18MP.
I agree completely.
BTW - another example of diffraction "limited" and downsampled vs "not limited":
Sorry, but I've no interest in the downsample, no worse type stuff. There's no way I'd buy a camera knowing that I had to downsample all of the images to get the performance that I wanted, even if it was relatively easy to do with a PS action.
Oops. due to some limitations of the available raw-converters, it seems that the downsampled 18MP F/16 is actually sharper and contains less artefacts than the 10MP picture. Even if the 15MP camera is 22% further into the "diffraction limited" region.
Are those 100% or higher crops?
Sep 10, 2009 at 05:17 PM
p.11 #19the d300s killer
Those are 100% of the 400D's 10MP, 50D downsampled to same size. I wouldn't say that this would limit me in any conceivable way, "having" to downsample shots taken at F16... As I mentioned earlier - checking my database statistics for this month, 99pointsomething percent of my shots are taken at <F/5.6 :-)
I could do the other way around, upsample the 400D to 15MP, and then you'd REALLY see some difference. Do keep in mind that we're deep into "diffraction limited" apertures with F/16 in a 50D - which if calculated with the same figures as used earlier should be diffraction limited at F/10.
I think I'll do just that. Wait a minute.
I could do the other way around, upsample the 400D to 15MP, and then you'd REALLY see some difference. Do keep in mind that we're deep into "diffraction limited" apertures with F/16 in a 50D - which if calculated with the same figures as used earlier should be diffraction limited at F/10.
I think I'll do just that. Wait a minute.
Sep 10, 2009 at 05:28 PM
p.11 #20the d300s killer
Well, that's more than a minute, but I had to answer a few emails also. Anyway.
Two F/16 pictures, one taken at 1.6x crop 10MP and upsampled, the other one at 1.6x crop 15MP.
I'd say there's a slight edge to the 15MP version, mostly because of the very slight loss of resolution in the 10MP picture, and due to the raw-interpolation artefacts (stray resolution - short "breaks" of cross-aligned 3-pixel areas and such) getting a bit more magnified in the upsampled 10MP picture.
http://pici.se/pictures/TZgUxkNoO.jpg
Two F/16 pictures, one taken at 1.6x crop 10MP and upsampled, the other one at 1.6x crop 15MP.
I'd say there's a slight edge to the 15MP version, mostly because of the very slight loss of resolution in the 10MP picture, and due to the raw-interpolation artefacts (stray resolution - short "breaks" of cross-aligned 3-pixel areas and such) getting a bit more magnified in the upsampled 10MP picture.
http://pici.se/pictures/TZgUxkNoO.jpg
Sep 10, 2009 at 06:01 PM