Two aspects are attractive to me: The new body with the built-in grip bump on front and the backside sensor. One of the weaknesses in the Nikon 1 lineup for me has been poor high ISO performance.
I am interested in how Nikon can do this huge pixel density and maintain ISO performance. It would appear to me if my calculations are correct, the pixel density on this new J5 is 9 1/2 times as dense as the D4s, about 4 1/4 times as dense as the D810.
Can anyone offer an explanation as to how this is achieved?
Also, could it be that the J5 and previous iterations are the test mule for a crop sensor or full frame mirrorless?
I am interested in how Nikon can do this huge pixel density and maintain ISO performance. It would appear to me if my calculations are correct, the pixel density on this new J5 is 9 1/2 times as dense as the D4s, about 4 1/4 times as dense as the D810.
Can anyone offer an explanation as to how this is achieved?
Also, could it be that the J5 and previous iterations are the test mule for a crop sensor or full frame mirrorless?
I doubt it is achieved. The marketing department can slap any ISO it wants onto the spec and tell engineering to make that the native ISO, regardless of how crappy the result.
1. It probably isn't achieved. Your comparison assumes the high ISO images would have roughly the same amount of noise. That assumption most likely is in error. Nikon is probably using very strong noise reduction to achieve passable images in a Nikon 1 at ISO beyond 3,200. It likely won't show up much when you look at images on a cell phone or printed at 5x7 or 8x10 but will be very obvious when printed higher. We will see what the DxOMark high ISO test reveals but I would be surprised it if was much more than ISO 1,000 which itself is about a doubling of the previous Nikon 1 sensors. Still, we should remember that IQ is probably quite fine when the Nikon 1 is shot at ISO 160 to ISO 800. The J5 may have an auto ISO setting which allows you to limit the camera from going higher than ISO 800 (or 1,600) and that may be the way to shoot this body.
2. An old idea I have not heard much about lately: pixel binning. At higher ISO "bin" many pixels together and process them as one pixel?
3. I think Nikon has developed software which selectively can apply stronger noise reduction to areas in an image which are a solid color containing no detail such as blue sky or black shadows. Noise is more easily seen in these areas so if Nikon can selectively reduce that noise It looks like cleaner higher ISO to most people.
5. It is a lie.
However, it does remain true that the J5 pixel pitch is amazing. What would it yield in megapixels if applied to an FX sensor? About 154 mp in a backside illuminated sensor. Even if the IQ of such an FX sensor was only good to ISO 800 or 1,000 it should be worth producing. Cannon's 50mpixel sensor would seem puny by comparison.
When I started reading the specs of the camera, I was like wow, Nikon is actually listening to its customers. I mean, Compact, decent, but not the best ISO performance, 4k video, that has been a long time coming. 20FPS. 171 af points (if they actually work :-) ).
Then reality set in and I was like.... how good is that high ISO performance. Those 4k video samples are horrible, The first thing I noticed was flat colors and blown out highlights everywhere....
So now I am more concerned about my next DSLR upgrade from my D700 to a D900. Will they get it right or screw it up?
Don't worry. They will get it right on DX and FX sensors. The most recent releases of those sensors have been spectacular. It is just the CX sensor which I think Nikon is pushing beyond the acceptable limits they have set for ISO designations in DX and FX. In other words, Nikon has much lower criteria for what is acceptable IQ in a CX sensor at ISO 6,400 than for what is acceptable IQ in DX or FX at ISO of 6.400. I think Nikon is measuring CX high ISO performance with a different yardstick than they use to measure DX or FX high ISO performance. We will be able to judge better when we have some direct comparisons such as provided by DxOMark and Imaging-Resource.
Okay, until you know which sensors are "front-side" (FSI) or "back-side" (BSI) illuminated, you don't know if you're comparing apples and oranges. Nikon states the 1j5 is FSI, but I can't definitively find that information for the D4s or 810, although when Nikon describes the technologies behind the D4 sensor, they use FSI-oriented improvement words such as "...and anti-reflective coating is used on various parts – all of which results in minimized ghost and flare." (src: http://imaging.nikon.com/lineup/dslr/d4/features01.htm) - this implies anti-reflective coatings are used in the circuitry layer to optimize the circuitry pass-through, a challenge of FSI.
Aptina has a really good paper describing the trades between BSI and pixel size: https://www.aptina.com/news/FSI-BSI-WhitePaper.pdf. Essentially, large-pixel arrays can achieve good low-light performance in FSI without many stupid pet tricks, but if you want to increase the MP count on a given format, you need to get the sensors out from behind the circuitry, hence BSI
Edit: Okay, I'm not reading the posts well enough. My apologies... Although, it would seem you can go for a higher pixel density if you're less worried about having to tunnel through the readout circuitry layer. I'd be really interested in what they're doing to cost-effectively thin the substrate - semiconductor processes don't lend themselves to after-application "machining".
Oh, to answer the original question: Yes, I want one. First mirrorless camera I'd consider. Gotta figure out if this old man would miss his viewfinder, though...
Meh, just another soccer mum camera. Nikon needs to stop messing around with small sensor mirrorless and step up their game.
I agree, but I still may get one if the price drops a little (say $400 without the included lens). And I love your signature
I also like the way the silver one looks. It looks like it's all grown up and mature now, not some dinky plastic thing.
Although at $500, I could get a lot of other cool stuff for that price. So no, I'm definitely not buying this thing. I do like the dual dial setup and the proper PASM dial, which earlier 1 cameras didn't have (I know the V series does, but that's pretty expensive too).
$500 is at least not too too expensive, but still unreasonable for those who already own DSLRs. Besides, I have a Sony RX100 Mk 1.
"Gotta figure out if this old man would miss his viewfinder, though..." I suspect the new body shape and back side illumination sensor will soon be out in a Nikon 1 V version with an EVF which I hope is attached and not detachable as in the V3 version. Both the grip and the viewfinder should be part of the V body and not add on pieces.
However, it does remain true that the J5 pixel pitch is amazing. What would it yield in megapixels if applied to an FX sensor? About 154 mp in a backside illuminated sensor. Even if the IQ of such an FX sensor was only good to ISO 800 or 1,000 it should be worth producing. Cannon's 50mpixel sensor would seem puny by comparison.
How many megapixels of resolution can a CX, DX, or FX lens have for a reasonable cost (say less than the cost of a Zeiss Otus)?
There are practical and engineering constraints at play and lens are the true bottleneck on a D810 for all but the best lenses.
What is that bottleneck? My guess is 60, 30 and 7 or 8 megapixels for FX, DX and CX respectively. I would not compromise anything to exceed the bottleneck. 20 megapixels on a CX may not be a literal lie, but it certainly is a practical one.
Ok, someone correct my math if wrong, but I get a pixel pitch of about 2.3um...? If so, that's not an unrealistic number.
I went looking for ~2.3um cameras to compare, found the Fuji X10 with a 2.2um, albeit forward-illuminated. For this camera, dpreview says you can do 12.8K, but it don't look too good, stay at 1600 or below. So, I'll be interested to see a comparative test for this camera with the back-illluminated sensor, as the sensor construction removes a lot of impediments to collecting photons, as well as allowing a larger effective sensel size. May not do 12.8K to your liking, but it should be decent.
Moments of Light - D610 D7K S5pro 70-200f4 18-200 150f2.8 12-24 18-70 35-70f2.8 : C&C very welcome! Being a photographer is a lot like being a Christian: Some people look at you funny but do not see the amazing beauty all around them - heartyfisher.
I really wonder if Nikon is selling enough 1 series to justify all the development in new CX models. When are they going to accept they have bet on a dead horse ? Is it the Japanese market which is keeping this format alive ?
In the meantime, I see more and more D700/800/810 owners ( usually old-timers ) carrying Sony R7s and leaving heavy equipment home ( that is what they say when I start chatting ). That is one missed market opportunity by Nikon there. I bet most Nikon professionals would go for any Nikon branded body that resembled the R7 , as a second/third/fourth camera ...
They sell better in Asia, where they have done well, then North America.
I think if Nikon was motivated, they could produce a very small full fram DSLR that could compete with the Sony in form factor. The only advantages of the Sony system is the flange to focal length distance of the mirror less system and the mirror box. Think about the small size of the D3xxx and D5xxx series. Sure they are DX which leaves a little more room for internal electronics, but they still have to accommodate the f-mount which is the true bottleneck for a smaller size.
And if they wanted to make it smaller still, they could get rid of the mirror, while still accommodating the f-mount.
Personally, I would love to see a DX mirror less interchangeable lens system from Nikon with an optimally designed mount (the f-mount is to big and only exists in DX to facilitate the transition from film to digital). To me the CX system is too much of a compromise on image quality. Also, it is a consumer grade. I would be interested in a professional/ prosumer grade. That would be my small walk around camera system that is currently partially met with my Coolpix A.
However, it does remain true that the J5 pixel pitch is amazing. What would it yield in megapixels if applied to an FX sensor? About 154 mp in a backside illuminated sensor. Even if the IQ of such an FX sensor was only good to ISO 800 or 1,000 it should be worth producing. Cannon's 50mpixel sensor would seem puny by comparison.
How many megapixels of resolution can a CX, DX, or FX lens have for a reasonable cost (say less than the cost of a Zeiss Otus)?
There are practical and engineering constraints at play and lens are the true bottleneck on a D810 for all but the best lenses.
What is that bottleneck? My guess is 60, 30 and 7 or 8 megapixels for FX, DX and CX respectively. I would not compromise anything to exceed the bottleneck. 20 megapixels on a CX may not be a literal lie, but it certainly is a practical one.
Huh? Not sure I understand. Are you saying that there is no lens that can resolve more than 8mp on a CX sized sensor?
I should add that I am surprised Nikon keeps increasing the pixel count on Nikon 1 sensors since they already perform poorly at high ISO and I would think increasing pixel count would have a negative effect on improving high ISO IQ. People aren't going to be shooting with a Nikon 1 if they are going to print poster size. Also, as I understand it diffraction limitation occurs sooner with more megapixels in the sensor. Most Nikon 1 zoom lenses are designed from f3.5 to f5.6. At some point wouldn't increased mp in the sensor result in image degradation due to diffraction with the lens stopped down just one stop to f8?
Diffraction is unrelated to the number of pixels, only the sensor size matters. The number and size of the pixels also has very little to do high ISO capability.
These are both urban legends that have persisted on this forum. I thought I killed them on one of the D400 threads :-)
However, it does remain true that the J5 pixel pitch is amazing. What would it yield in megapixels if applied to an FX sensor? About 154 mp in a backside illuminated sensor. Even if the IQ of such an FX sensor was only good to ISO 800 or 1,000 it should be worth producing. Cannon's 50mpixel sensor would seem puny by comparison.
How many megapixels of resolution can a CX, DX, or FX lens have for a reasonable cost (say less than the cost of a Zeiss Otus)?
There are practical and engineering constraints at play and lens are the true bottleneck on a D810 for all but the best lenses.
What is that bottleneck? My guess is 60, 30 and 7 or 8 megapixels for FX, DX and CX respectively. I would not compromise anything to exceed the bottleneck. 20 megapixels on a CX may not be a literal lie, but it certainly is a practical one.
Huh? Not sure I understand. Are you saying that there is no lens that can resolve more than 8mp on a CX sized sensor?
Yes, subject to it being a very rough estimate. Of course, tests would need to be performed to refine the estimate.
Let's assume that you are using an FX lens attached to a Nikon 1 with an adapter. The 20 megapixel camera is using 1/7.29th (1/(2.7^2) of the image area. That means that unless the lens is sharp enough to outperform a 145.8 megapixel FX sensor (20 * 7.29) in the centre area, the centre 1-7.29th area is not going to outperform the 20 megapixel CX sensor.
All things being equal, the potential resolution of a lens is determined by its format size. But all things are not equal, as CX lenses are cheap consumer grades. I bet that anything more than 5 megapixels is a waste. That is why I used the FX example above. Do you know of any FX lenses that can outperform a 145.8 megapixel FX sensor? Not a fair question I admit, as none exist.
My point, while open to argument on the technical details, is meant to illustrate that for a consumer buying a Nikon 1 camera using Nikon 1 lenses, their results are not likely to be any different regardless of whether their camera is 10 or 20 megapixels.
Nice links ! i like my 1V1 too and I use it with my 18-200 dx kit lense, works well. if I want sharper I use my 70-200 F4 or 150 sigma macro. I am keen to get the 1j5 once the body is available at a lower price :-) the Nikon1 system is a lot of fun!
Regarding how well lenses resolve on these small sensors? no idea... i dont have one yet that is higher than my 10mp 1V1..
Post edited by heartyfisher on
Moments of Light - D610 D7K S5pro 70-200f4 18-200 150f2.8 12-24 18-70 35-70f2.8 : C&C very welcome! Being a photographer is a lot like being a Christian: Some people look at you funny but do not see the amazing beauty all around them - heartyfisher.
Comments
New J5 116.16 sq mm sensor 20.8MP = 179,063 px/mm sq
ISO 12,800
D4s 860.4 sq mm sensor 16.2 MP = 18,828 px/mm sq
ISO 25,600
D810 861.6 sensor 36.3 MP = 42,131 px/mm sq
ISO 12,800
I am interested in how Nikon can do this huge pixel density and maintain ISO performance. It would appear to me if my calculations are correct, the pixel density on this new J5 is 9 1/2 times as dense as the D4s, about 4 1/4 times as dense as the D810.
Can anyone offer an explanation as to how this is achieved?
Also, could it be that the J5 and previous iterations are the test mule for a crop sensor or full frame mirrorless?
http://www.extremetech.com/extreme/149742-how-back-illuminated-sensors-work-and-why-theyre-the-future-of-digital-photography
1. It probably isn't achieved. Your comparison assumes the high ISO images would have roughly the same amount of noise. That assumption most likely is in error. Nikon is probably using very strong noise reduction to achieve passable images in a Nikon 1 at ISO beyond 3,200. It likely won't show up much when you look at images on a cell phone or printed at 5x7 or 8x10 but will be very obvious when printed higher. We will see what the DxOMark high ISO test reveals but I would be surprised it if was much more than ISO 1,000 which itself is about a doubling of the previous Nikon 1 sensors. Still, we should remember that IQ is probably quite fine when the Nikon 1 is shot at ISO 160 to ISO 800. The J5 may have an auto ISO setting which allows you to limit the camera from going higher than ISO 800 (or 1,600) and that may be the way to shoot this body.
2. An old idea I have not heard much about lately: pixel binning. At higher ISO "bin" many pixels together and process them as one pixel?
3. I think Nikon has developed software which selectively can apply stronger noise reduction to areas in an image which are a solid color containing no detail such as blue sky or black shadows. Noise is more easily seen in these areas so if Nikon can selectively reduce that noise It looks like cleaner higher ISO to most people.
5. It is a lie.
However, it does remain true that the J5 pixel pitch is amazing. What would it yield in megapixels if applied to an FX sensor? About 154 mp in a backside illuminated sensor. Even if the IQ of such an FX sensor was only good to ISO 800 or 1,000 it should be worth producing. Cannon's 50mpixel sensor would seem puny by comparison.
Then reality set in and I was like.... how good is that high ISO performance. Those 4k video samples are horrible, The first thing I noticed was flat colors and blown out highlights everywhere....
So now I am more concerned about my next DSLR upgrade from my D700 to a D900. Will they get it right or screw it up?
Okay, until you know which sensors are "front-side" (FSI) or "back-side" (BSI) illuminated, you don't know if you're comparing apples and oranges. Nikon states the 1j5 is FSI, but I can't definitively find that information for the D4s or 810, although when Nikon describes the technologies behind the D4 sensor, they use FSI-oriented improvement words such as "...and anti-reflective coating is used on various parts – all of which results in minimized ghost and flare." (src: http://imaging.nikon.com/lineup/dslr/d4/features01.htm) - this implies anti-reflective coatings are used in the circuitry layer to optimize the circuitry pass-through, a challenge of FSI.
Aptina has a really good paper describing the trades between BSI and pixel size: https://www.aptina.com/news/FSI-BSI-WhitePaper.pdf. Essentially, large-pixel arrays can achieve good low-light performance in FSI without many stupid pet tricks, but if you want to increase the MP count on a given format, you need to get the sensors out from behind the circuitry, hence BSI
Edit: Okay, I'm not reading the posts well enough. My apologies... Although, it would seem you can go for a higher pixel density if you're less worried about having to tunnel through the readout circuitry layer. I'd be really interested in what they're doing to cost-effectively thin the substrate - semiconductor processes don't lend themselves to after-application "machining".
Although at $500, I could get a lot of other cool stuff for that price. So no, I'm definitely not buying this thing. I do like the dual dial setup and the proper PASM dial, which earlier 1 cameras didn't have (I know the V series does, but that's pretty expensive too).
$500 is at least not too too expensive, but still unreasonable for those who already own DSLRs. Besides, I have a Sony RX100 Mk 1.
There are practical and engineering constraints at play and lens are the true bottleneck on a D810 for all but the best lenses.
What is that bottleneck? My guess is 60, 30 and 7 or 8 megapixels for FX, DX and CX respectively. I would not compromise anything to exceed the bottleneck. 20 megapixels on a CX may not be a literal lie, but it certainly is a practical one.
I went looking for ~2.3um cameras to compare, found the Fuji X10 with a 2.2um, albeit forward-illuminated. For this camera, dpreview says you can do 12.8K, but it don't look too good, stay at 1600 or below. So, I'll be interested to see a comparative test for this camera with the back-illluminated sensor, as the sensor construction removes a lot of impediments to collecting photons, as well as allowing a larger effective sensel size. May not do 12.8K to your liking, but it should be decent.
Being a photographer is a lot like being a Christian: Some people look at you funny but do not see the amazing beauty all around them - heartyfisher.
In the meantime, I see more and more D700/800/810 owners ( usually old-timers ) carrying Sony R7s and leaving heavy equipment home ( that is what they say when I start chatting ). That is one missed market opportunity by Nikon there. I bet most Nikon professionals would go for any Nikon branded body that resembled the R7 , as a second/third/fourth camera ...
I think if Nikon was motivated, they could produce a very small full fram DSLR that could compete with the Sony in form factor. The only advantages of the Sony system is the flange to focal length distance of the mirror less system and the mirror box. Think about the small size of the D3xxx and D5xxx series. Sure they are DX which leaves a little more room for internal electronics, but they still have to accommodate the f-mount which is the true bottleneck for a smaller size.
And if they wanted to make it smaller still, they could get rid of the mirror, while still accommodating the f-mount.
Personally, I would love to see a DX mirror less interchangeable lens system from Nikon with an optimally designed mount (the f-mount is to big and only exists in DX to facilitate the transition from film to digital). To me the CX system is too much of a compromise on image quality. Also, it is a consumer grade. I would be interested in a professional/ prosumer grade. That would be my small walk around camera system that is currently partially met with my Coolpix A.
The number and size of the pixels also has very little to do high ISO capability.
These are both urban legends that have persisted on this forum. I thought I killed them on one of the D400 threads :-)
Let's assume that you are using an FX lens attached to a Nikon 1 with an adapter. The 20 megapixel camera is using 1/7.29th (1/(2.7^2) of the image area. That means that unless the lens is sharp enough to outperform a 145.8 megapixel FX sensor (20 * 7.29) in the centre area, the centre 1-7.29th area is not going to outperform the 20 megapixel CX sensor.
All things being equal, the potential resolution of a lens is determined by its format size. But all things are not equal, as CX lenses are cheap consumer grades. I bet that anything more than 5 megapixels is a waste. That is why I used the FX example above. Do you know of any FX lenses that can outperform a 145.8 megapixel FX sensor? Not a fair question I admit, as none exist.
My point, while open to argument on the technical details, is meant to illustrate that for a consumer buying a Nikon 1 camera using Nikon 1 lenses, their results are not likely to be any different regardless of whether their camera is 10 or 20 megapixels.
https://photographylife.com/photographing-butterflies-with-the-1-nikon-cx-70-300
https://photographylife.com/sample-bird-images-using-1-nikon-cx-70-300-lens
Regarding how well lenses resolve on these small sensors? no idea... i dont have one yet that is higher than my 10mp 1V1..
Being a photographer is a lot like being a Christian: Some people look at you funny but do not see the amazing beauty all around them - heartyfisher.