Reasons for wanting higher mega pixel count

Started 4 months ago | Discussions
KE_DP
KE_DP Veteran Member • Posts: 5,621
Re: Reasons for wanting higher mega pixel count

Considering Moore's Law photography data in gigapixels are only a matter of time - maybe even terapixels.

 KE_DP's gear list:KE_DP's gear list
Sony a7R III Sony a7R IV
JimKasson
JimKasson Forum Pro • Posts: 26,549
Re: Reasons for wanting higher mega pixel count
2

KE_DP wrote:

Considering Moore's Law photography data in gigapixels are only a matter of time - maybe even terapixels.

There are some considerable difficulties building a lens that can provide terapixels of information -- as opposed to noise -- in a single capture. The simplest one is diffraction.

-- hide signature --
 JimKasson's gear list:JimKasson's gear list
Nikon D5 Sony a7 III Nikon Z7 Fujifilm GFX 100 Sony a7R IV +4 more
sportyaccordy Forum Pro • Posts: 15,317
Re: Reasons for wanting higher mega pixel count

KE_DP wrote:

Considering Moore's Law photography data in gigapixels are only a matter of time - maybe even terapixels.

Terapixels enter the realm of individual photons. Maybe they could do away with the color filters and just capture the wavelength at each photo site.

-- hide signature --

Sometimes I take pictures with my gear- https://www.flickr.com/photos/41601371@N00/

 sportyaccordy's gear list:sportyaccordy's gear list
NEX-5T Sony a7R II Canon EOS 5D Mark IV Canon EF 24-85mm f/3.5-4.5 USM Sigma 70-200mm F2.8 EX DG Macro HSM II +6 more
KE_DP
KE_DP Veteran Member • Posts: 5,621
Re: Reasons for wanting higher mega pixel count

JimKasson wrote:

KE_DP wrote:

Considering Moore's Law photography data in gigapixels are only a matter of time - maybe even terapixels.

There are some considerable difficulties building a lens that can provide terapixels of information -- as opposed to noise -- in a single capture. The simplest one is diffraction.

That assumes lenses will still be used in the future - forever.    

 KE_DP's gear list:KE_DP's gear list
Sony a7R III Sony a7R IV
Entropy512 Senior Member • Posts: 4,601
Re: Reasons for wanting higher mega pixel count
2

KE_DP wrote:

Considering Moore's Law photography data in gigapixels are only a matter of time - maybe even terapixels.

In addition to Jim's comments, Moore's Law has been tending to slow down due to physical limits of how small you can make things before the fact that some parts of the chip are mere atoms thick becomes problematic.

Also once the pixel pitch becomes on the order of the wavelength of the light you're trying to record, things start getting interesting.  I recall seeing some papers indicating they're seeing some unique challenges with phone sensors down in the 1 micron pitch range - visible light is roughly 0.4 to 0.7 microns in wavelength.

The A7RIV is around 3.8 micron pixel pitch.  1 micron FF would take you to around 866 MPixels.

-- hide signature --

Context is key. If I have quoted someone else's post when replying, please do not reply to something I say without reading text that I have quoted, and understanding the reason the quote function exists.

 Entropy512's gear list:Entropy512's gear list
Sony a6000 Pentax K-5 Pentax K-01 Sony a6300 Canon EF 85mm f/1.8 USM +5 more
Trollmannx Senior Member • Posts: 5,658
Re: Reasons for wanting higher mega pixel count

JimKasson wrote:

Trollmannx wrote:

Or maybe change direction and include large format cameras - 9x12 cm or 4x5 inch - guess we are talking 300-400 MP full frame to beat such monsters (at least based on something like 6 micron photographic film grain lumps as the basis for that kind of images).

I have a 72x96mm digital camera, but it's only 144 MP (counting each the red, green, and blue sensors individually; as I remember that's the CIPA standard).

Thank you for that piece of information - guess CIPA stand well above humble me... 

JimKasson
JimKasson Forum Pro • Posts: 26,549
Re: Reasons for wanting higher mega pixel count

Trollmannx wrote:

JimKasson wrote:

Trollmannx wrote:

Or maybe change direction and include large format cameras - 9x12 cm or 4x5 inch - guess we are talking 300-400 MP full frame to beat such monsters (at least based on something like 6 micron photographic film grain lumps as the basis for that kind of images).

I have a 72x96mm digital camera, but it's only 144 MP (counting each the red, green, and blue sensors individually; as I remember that's the CIPA standard).

Thank you for that piece of information - guess CIPA stand well above humble me...

I din't mean to say you were wrong. There are higher res versions. Or, rather, were, as they're not made anymore.

And anyway, I'll bet you were talking about film.

-- hide signature --
 JimKasson's gear list:JimKasson's gear list
Nikon D5 Sony a7 III Nikon Z7 Fujifilm GFX 100 Sony a7R IV +4 more
Trollmannx Senior Member • Posts: 5,658
Re: Reasons for wanting higher mega pixel count

JimKasson wrote:

Trollmannx wrote:

JimKasson wrote:

Trollmannx wrote:

Or maybe change direction and include large format cameras - 9x12 cm or 4x5 inch - guess we are talking 300-400 MP full frame to beat such monsters (at least based on something like 6 micron photographic film grain lumps as the basis for that kind of images).

I have a 72x96mm digital camera, but it's only 144 MP (counting each the red, green, and blue sensors individually; as I remember that's the CIPA standard).

Thank you for that piece of information - guess CIPA stand well above humble me...

I din't mean to say you were wrong. There are higher res versions. Or, rather, were, as they're not made anymore.

And anyway, I'll bet you were talking about film.

Yes - the smoking gun is "photographic film grain lumps"... 

JimKasson
JimKasson Forum Pro • Posts: 26,549
Re: Reasons for wanting higher mega pixel count

Trollmannx wrote:

JimKasson wrote:

Trollmannx wrote:

JimKasson wrote:

Trollmannx wrote:

Or maybe change direction and include large format cameras - 9x12 cm or 4x5 inch - guess we are talking 300-400 MP full frame to beat such monsters (at least based on something like 6 micron photographic film grain lumps as the basis for that kind of images).

I have a 72x96mm digital camera, but it's only 144 MP (counting each the red, green, and blue sensors individually; as I remember that's the CIPA standard).

Thank you for that piece of information - guess CIPA stand well above humble me...

I din't mean to say you were wrong. There are higher res versions. Or, rather, were, as they're not made anymore.

And anyway, I'll bet you were talking about film.

Yes - the smoking gun is "photographic film grain lumps"...

Thinking about it, I don't think that 6 um is a limit for film resolution; we certainly produced finer feature sizes than that back in the day with visible-wavelength optical lithography and contact printing for making semiconductors.

Jim

-- hide signature --
 JimKasson's gear list:JimKasson's gear list
Nikon D5 Sony a7 III Nikon Z7 Fujifilm GFX 100 Sony a7R IV +4 more
DejayRezme Regular Member • Posts: 131
Re: Nonlinear noise reduction

JimKasson wrote:

Entropy512 wrote:

As Jim Kasson established in a blog post a while back with some good examples, with modern nonlinear NR methods, a higher resolution sensor is better as long as you don't go to such a high resolution as to start negatively impacting fill factor.

Here's a link to a comparison with linear methods:

https://blog.kasson.com/the-last-word/big-vs-little-pixels/

And here's a pointer to what happens with nonlinear NR:

https://blog.kasson.com/the-last-word/noise-reduction-with-nonlinear-tools-and-downsampling/

Jim

Thanks that is really interesting, although I'm mostly curious about 4k video. I haven't read all of it but it seems that a 42MP A7IV could beat a 12MP A7SIII for video. Can these nonlinear noise reduction algorithms be implemented fast enough to downsample 8k video on the fly to 4k video?

I would think that since flagship smartphones can easily run 3D games in 1080p they should be able to run reasonably complex image algorithms on a single input image.

With all this talk about computational photography I kinda wish nvidia would release a FF camera with an open OpenCL based eco system.

xh43k
xh43k Forum Member • Posts: 94
Re: Reasons for wanting higher mega pixel count

JimKasson wrote:

KE_DP wrote:

Considering Moore's Law photography data in gigapixels are only a matter of time - maybe even terapixels.

There are some considerable difficulties building a lens that can provide terapixels of information -- as opposed to noise -- in a single capture. The simplest one is diffraction.

I do have tiny zeiss lenses in my eyes after cataract surgery (im 30yo) and they seem sharp.

Resolution of an eye according to quick google search seems to be 324 megapixels.

Considering the size of the lenses in my eyes I think it could be 'viable' - of course, with much higher price.

 xh43k's gear list:xh43k's gear list
Sony a7 III Tamron 28-75mm F2.8 III +1 more
Trollmannx Senior Member • Posts: 5,658
Re: Reasons for wanting higher mega pixel count

JimKasson wrote:

Trollmannx wrote:

JimKasson wrote:

Trollmannx wrote:

JimKasson wrote:

Trollmannx wrote:

Or maybe change direction and include large format cameras - 9x12 cm or 4x5 inch - guess we are talking 300-400 MP full frame to beat such monsters (at least based on something like 6 micron photographic film grain lumps as the basis for that kind of images).

I have a 72x96mm digital camera, but it's only 144 MP (counting each the red, green, and blue sensors individually; as I remember that's the CIPA standard).

Thank you for that piece of information - guess CIPA stand well above humble me...

I din't mean to say you were wrong. There are higher res versions. Or, rather, were, as they're not made anymore.

And anyway, I'll bet you were talking about film.

Yes - the smoking gun is "photographic film grain lumps"...

Thinking about it, I don't think that 6 um is a limit for film resolution; we certainly produced finer feature sizes than that back in the day with visible-wavelength optical lithography and contact printing for making semiconductors.

Jim

Using ordinary film under a starry sky 6 micron was about the diameter of the faintest stars to be positively identified - smaller and things became very uncertain.

Just used this as a starting point for film grain vs pixel size - have to start somewhere. The good thing is getting input and correction from there on - that is appreciated.

Litography is not strictly what ordinary photographers are into, but ok - how small is the tiniest photosite ever made (sub micron diameter)?

(unknown member) Senior Member • Posts: 1,233
Re: Reasons for wanting higher mega pixel count

xh43k wrote:

JimKasson wrote:

KE_DP wrote:

Considering Moore's Law photography data in gigapixels are only a matter of time - maybe even terapixels.

There are some considerable difficulties building a lens that can provide terapixels of information -- as opposed to noise -- in a single capture. The simplest one is diffraction.

I do have tiny zeiss lenses in my eyes after cataract surgery (im 30yo) and they seem sharp.

Resolution of an eye according to quick google search seems to be 324 megapixels.

Considering the size of the lenses in my eyes I think it could be 'viable' - of course, with much higher price.

Add a wifi implant to transfer photos and I'm in!

Steven Rodgers New Member • Posts: 22
Then 12MP is enough?
1

I have a high end 4k monitor.

IIRC  sony uses 6K and downsized to get the sharpest 4K video.

Take into account some aggressive cropping and 12MP or 16MP would be enough for your two points.  A 12 MP sensor may have less noise due to the larger pixel size, and 6MP is enough for sharp 4K images.  And I've seen highly detailed 30 inch prints from 8MP images, so that is covered too.

I think the real reason for more MP is the same reason we buy Dodge Demons.  No one here will admit this I bet.   I just bought an A7RIII and I drive a car with a large v8 so I am in that group.

Entropy512 Senior Member • Posts: 4,601
Re: Nonlinear noise reduction
1

DejayRezme wrote:

JimKasson wrote:

Entropy512 wrote:

As Jim Kasson established in a blog post a while back with some good examples, with modern nonlinear NR methods, a higher resolution sensor is better as long as you don't go to such a high resolution as to start negatively impacting fill factor.

Here's a link to a comparison with linear methods:

https://blog.kasson.com/the-last-word/big-vs-little-pixels/

And here's a pointer to what happens with nonlinear NR:

https://blog.kasson.com/the-last-word/noise-reduction-with-nonlinear-tools-and-downsampling/

Jim

Thanks that is really interesting, although I'm mostly curious about 4k video. I haven't read all of it but it seems that a 42MP A7IV could beat a 12MP A7SIII for video. Can these nonlinear noise reduction algorithms be implemented fast enough to downsample 8k video on the fly to 4k video?

There is no A7M4, and there is no mark IV anything that is 42MP...

If you're coming up with some sort of theoretical construct where the next A7 family camera gets the R2/R3 resolution - that's extremely unlikely.  But as far as 42MP downsampled to 4k - with the right hardware, probably.  With anything remotely close to Sony's current BIONZ family - no.  There seems to be a 500 MPixel/sec sustained demosaicer/scaler limit even in the R4, even if the R3 can hit much higher readout rates within a given frame.

I would think that since flagship smartphones can easily run 3D games in 1080p they should be able to run reasonably complex image algorithms on a single input image.

Would need to do some memory bandwidth analysis here.

With all this talk about computational photography I kinda wish nvidia would release a FF camera with an open OpenCL based eco system.

That would be...  Interesting. 

That is SORT OF how a recently announced camera that (I think) had some DPR coverage appeared to behave - a high-bandwidth PCI Express camera feeding an Intel NUC.  Intel's integrated GPUs aren't stellar, but I think they're probably beyond what a BIONZ can do for certain operations.

-- hide signature --

Context is key. If I have quoted someone else's post when replying, please do not reply to something I say without reading text that I have quoted, and understanding the reason the quote function exists.

 Entropy512's gear list:Entropy512's gear list
Sony a6000 Pentax K-5 Pentax K-01 Sony a6300 Canon EF 85mm f/1.8 USM +5 more
JimKasson
JimKasson Forum Pro • Posts: 26,549
Re: Reasons for wanting higher mega pixel count

Trollmannx wrote:

JimKasson wrote:

Trollmannx wrote:

JimKasson wrote:

Trollmannx wrote:

JimKasson wrote:

Trollmannx wrote:

Or maybe change direction and include large format cameras - 9x12 cm or 4x5 inch - guess we are talking 300-400 MP full frame to beat such monsters (at least based on something like 6 micron photographic film grain lumps as the basis for that kind of images).

I have a 72x96mm digital camera, but it's only 144 MP (counting each the red, green, and blue sensors individually; as I remember that's the CIPA standard).

Thank you for that piece of information - guess CIPA stand well above humble me...

I din't mean to say you were wrong. There are higher res versions. Or, rather, were, as they're not made anymore.

And anyway, I'll bet you were talking about film.

Yes - the smoking gun is "photographic film grain lumps"...

Thinking about it, I don't think that 6 um is a limit for film resolution; we certainly produced finer feature sizes than that back in the day with visible-wavelength optical lithography and contact printing for making semiconductors.

Jim

Using ordinary film under a starry sky 6 micron was about the diameter of the faintest stars to be positively identified - smaller and things became very uncertain.

Just used this as a starting point for film grain vs pixel size - have to start somewhere. The good thing is getting input and correction from there on - that is appreciated.

Litography is not strictly what ordinary photographers are into, but ok - how small is the tiniest photosite ever made (sub micron diameter)?

I was limiting it to optical wavelengths and contact printers to make it more relevant to photography, but you're right that there are emulsions that can retain detail well under a micron.

-- hide signature --
 JimKasson's gear list:JimKasson's gear list
Nikon D5 Sony a7 III Nikon Z7 Fujifilm GFX 100 Sony a7R IV +4 more
JimKasson
JimKasson Forum Pro • Posts: 26,549
Re: Reasons for wanting higher mega pixel count

xh43k wrote:

Resolution of an eye according to quick google search seems to be 324 megapixels.

That number is way too high considering that the fovea only subtends about one degree of visual field.

Or are you talking about a stitched image due to saccade?

-- hide signature --
 JimKasson's gear list:JimKasson's gear list
Nikon D5 Sony a7 III Nikon Z7 Fujifilm GFX 100 Sony a7R IV +4 more
xh43k
xh43k Forum Member • Posts: 94
Re: Reasons for wanting higher mega pixel count

JimKasson wrote:

xh43k wrote:

Resolution of an eye according to quick google search seems to be 324 megapixels.

That number is way too high considering that the fovea only subtends about one degree of visual field.

Or are you talking about a stitched image due to saccade?

tbh I didn't even read the article, it just popped right out in google

I looked in to several articles and in few of them this repeats:

the human eye is 576 megapixels—but really only about 7 megapixels matter.

 xh43k's gear list:xh43k's gear list
Sony a7 III Tamron 28-75mm F2.8 III +1 more
JimKasson
JimKasson Forum Pro • Posts: 26,549
Re: Reasons for wanting higher mega pixel count

xh43k wrote:

JimKasson wrote:

xh43k wrote:

Resolution of an eye according to quick google search seems to be 324 megapixels.

That number is way too high considering that the fovea only subtends about one degree of visual field.

Or are you talking about a stitched image due to saccade?

tbh I didn't even read the article, it just popped right out in google, when checking it I found this

How many pixels are needed to match the resolution of the human eye? Each pixel must appear no larger than 0.3 arc-minute. Consider a 20 x 13.3-inch print viewed at 20 inches. The Print subtends an angle of 53 x 35.3 degrees, thus requiring 53*60/.3 = 10600 x 35*60/.3 = 7000 pixels, for a total of ~74 megapixels to show detail at the limits of human visual acuity.

And then in different article they say it's only 7 megapixels.. so I guess nobody knows exactly

It sounds like the key differences are not a lack of knowledge, but in the way the question is parsed.

Jim

-- hide signature --
 JimKasson's gear list:JimKasson's gear list
Nikon D5 Sony a7 III Nikon Z7 Fujifilm GFX 100 Sony a7R IV +4 more
Paul Barnard
Paul Barnard Veteran Member • Posts: 3,147
Re: Then 12MP is enough?

Steven Rodgers wrote:

I think the real reason for more MP is the same reason we buy Dodge Demons. No one here will admit this I bet. I just bought an A7RIII and I drive a car with a large v8 so I am in that group.

My wife tells me I'm compensating for something...

 Paul Barnard's gear list:Paul Barnard's gear list
Sony a7R III Sony FE 55mm F1.8 Sony FE 70-200 F4 Zeiss Batis 25mm F2 Sony 1.4x Teleconverter (2016) +4 more
Keyboard shortcuts:
FForum MMy threads