Canon’s Testing a 75MP+ Pro DSLR

Started Jul 21, 2013 | Discussions
RedFox88 Forum Pro • Posts: 28,541
Re: The question is...

SubPrime wrote:

RedFox88 wrote:

That is so hilarious. Canon owned the high pixel count 35mm dSLR market from 2002 through 2007. They aren't going to "give up" on it calling it not their market.

I am only reporting what I was told. I agree it makes sense for Canon to roll out a high res camera.

They announced a working aps-h image sensor about 2 years ago that was 120 MP.

Yes, that was a one off R&D project, not related to any camera or product as I recall.

Just because it didn't go into a consumer product doesn't mean it wasn't valuable for them to produce and have working.

Canon is certainly working on something big to be their flagship camera again. It seems they want to get it right and make it big and not do the 38 MP.

So the rumors would suggest. The fact remains that there are diminishing returns to packing more pixels into a sensor, both from tech POV of the sensor and the lenses.

Adding pixels increases the resolution.  The only way to gain resolution is to redesign lenses which result, from Canon's offerings, in significant price increases.

Again, I am only reporting what I was told.

And salesmen want to sell what they have in stock too.

True, but again, no one in eh market for a 1DX is going to wait 12 months for something bigger.

SubPrime Senior Member • Posts: 1,238
Re: The question is...

RedFox88 wrote:

Adding pixels increases the resolution. The only way to gain resolution is to redesign lenses which result, from Canon's offerings, in significant price increases.

Yes, adding pixels does increase resolution, but if the lenses are not up to the task , you're simply resolving the flaws.  So if your corners are soft, then those soft corners will be made ore obvious.

And yes, the lenses will be significantly more expensive, but then, so will the camera.

SubPrime Senior Member • Posts: 1,238
Re: 75MP+ Pro DSLR: for architecture and landscape?

David Hull wrote:

SubPrime wrote:

NancyP wrote:

This sounds like a nightmare for action, wildlife, and PJ photographers, and a serious headache for wedding, fashion, and portrait photographers. One frame per second, maybe?

Of course. Do you really think Canon would release a 1DX killer?

I don't think you need to worry about that. The 1Dx is designed to maximize AF performance and firing rate. Anything with 75MP is not likely to match the 1Dx in the latter.

I agree entirely, which is why I was bemused by the suiggestiont that such a camera would be any use to an action photographer, PJ photographers, or wedding photographer.

SubPrime Senior Member • Posts: 1,238
Re: Per-pixel comparisons are meaningless...

Great Bustard wrote:

...for photos made from different numbers of pixels. In any case, rather than rebut you point-by-point (obviously a "pointless" exercise), I'll just ask you a single, and simple, question:

Would a landscape photographer produce "better" photos with FF or APS-C, or does it not matter at all?

Let us know.

It shouldn't matter in the end other than the fact the FF sensor will capture a better quality image in the end - provided the lenses are up to the task.

SubPrime Senior Member • Posts: 1,238
Re: Canon’s Testing a 75MP+ Pro DSLR

David Hull wrote:

What people seem to miss here is that the lens is what it is and the resolution of the sensor really has no influence on that. The lens has a special frequency response (in lines/mm). If the camera resolution is significantly more than that then the camera resolution won’t significantly influence the resolution of the cascaded system (sensor+lens).

Good point (in theory) but it would make the limits of the lens all the more obvious.   I can only refer to the example of the Nikon 14-24G lens, which looked outstanding (corner to corner) on a 24 mpx sensor, but less so on a D800. So at this point, throwing a higher res sensor at the lens will only make the flaws more obvious.

Your theory also excludes the diminishing returns that come with squeezing that many photo sites onto a 35 mm sensor.

It is as if you played your CD player through a phone line it would have a significant influence on the sound you heard. If on the other hand you played it through a channel that had bandwidth out to 500 kHz all you would hear is how good the CD machine sounded.

False analogy because our only trained ears will pick up the inadequacies of the CD machine, whereas untrained eyes would easily pick u the inadequacies of the lens.

I don’t get the “sensor out resolves the lens thing”; ideally that is what you want at least up to the point where you reach some point of diminishing returns and are wasting money on the sensor (as in my Audio example above).

I agree. And it it more than likely that this will be

Even 75MP would be more like a mosaic than an analog image of the lens' projection, for a moderately sharp lens.

Wrong. Before you get to the mosaic, you begin to see the inadequacies of most lenses.

Of course, but the imperfections of the sensor are gone. You can finally toss the AIF for example (since the lens has become one).

Fair enough, but with the imperfections of the sensor gone, you then have to deal with the imperfections of the lens.

Rick Knepper
Rick Knepper Forum Pro • Posts: 16,216
Tp Press Correspondent
2

Yes, I have a D800E. I would replace my 5D2 with a Canon version of the D800E.

And, since I have a D800E, I have the freedom to pass on any behemoth that does not meet my criteria i.e. first and foremost, a 1 series body, probably 2nd, the usual hefty 1 series price tag for hi rez.

-- hide signature --

Rick Knepper, photographer, non-professional, shooting for pleasure, check my profile for gear list and philosophy. TJ said, "Every generation needs a new revolution".

 Rick Knepper's gear list:Rick Knepper's gear list
Pentax 645Z Canon EOS 5DS R Fujifilm GFX 50S Fujifilm GF 23mm F4 Fujifilm GF 32-64mm F4 +8 more
altair8800 Senior Member • Posts: 1,903
Re: Canon’s Testing a 75MP+ Pro DSLR

SubPrime wrote:

David Hull wrote:

What people seem to miss here is that the lens is what it is and the resolution of the sensor really has no influence on that. The lens has a special frequency response (in lines/mm). If the camera resolution is significantly more than that then the camera resolution won’t significantly influence the resolution of the cascaded system (sensor+lens).

Good point (in theory) but it would make the limits of the lens all the more obvious. I can only refer to the example of the Nikon 14-24G lens, which looked outstanding (corner to corner) on a 24 mpx sensor, but less so on a D800. So at this point, throwing a higher res sensor at the lens will only make the flaws more obvious.

I did not see the example , but the 24 Mp sensor would be APS-C (DX) which does not use the corners of the lens. That 24 Mp sensor pixel density would give 56 Mp FF. A 56 Mp FF sensor should look "outstanding" with that lens, except in the corners.

Dan

SubPrime Senior Member • Posts: 1,238
Re: Canon’s Testing a 75MP+ Pro DSLR

altair8800 wrote:

SubPrime wrote:

David Hull wrote:

What people seem to miss here is that the lens is what it is and the resolution of the sensor really has no influence on that. The lens has a special frequency response (in lines/mm). If the camera resolution is significantly more than that then the camera resolution won’t significantly influence the resolution of the cascaded system (sensor+lens).

Good point (in theory) but it would make the limits of the lens all the more obvious. I can only refer to the example of the Nikon 14-24G lens, which looked outstanding (corner to corner) on a 24 mpx sensor, but less so on a D800. So at this point, throwing a higher res sensor at the lens will only make the flaws more obvious.

I did not see the example , but the 24 Mp sensor would be APS-C (DX) which does not use the corners of the lens. That 24 Mp sensor pixel density would give 56 Mp FF. A 56 Mp FF sensor should look "outstanding" with that lens, except in the corners.

Dan

The camera I was referring to was the D3X, which is FF 24 mpx, so yes, it does use the corners.  The lens in very good on the D800/E, but not sure I would describe it as "outstanding".

RedFox88 Forum Pro • Posts: 28,541
Re: The question is...
1

SubPrime wrote:

RedFox88 wrote:

Adding pixels increases the resolution. The only way to gain resolution is to redesign lenses which result, from Canon's offerings, in significant price increases.

Yes, adding pixels does increase resolution, but if the lenses are not up to the task , you're simply resolving the flaws. So if your corners are soft, then those soft corners will be made ore obvious.

Not sure what you are so against.  What is at the corners of images is rarely of important or the main subject.  Need edge/corner sharpness and they make macro lenses for that.

And yes, the lenses will be significantly more expensive, but then, so will the camera.

The d800 cost $300 less than the d700 when it was launched while Canon's long telephoto lenses went from $6,000 or $8,000 to 10 to $12,000 when they updated them to a II version.  You are flatly wrong.  Camera companies have been adding pixels to cameras for years and prices have been going down and down.

altair8800 Senior Member • Posts: 1,903
Re: Canon’s Testing a 75MP+ Pro DSLR

SubPrime wrote:

altair8800 wrote:

SubPrime wrote:

David Hull wrote:

What people seem to miss here is that the lens is what it is and the resolution of the sensor really has no influence on that. The lens has a special frequency response (in lines/mm). If the camera resolution is significantly more than that then the camera resolution won’t significantly influence the resolution of the cascaded system (sensor+lens).

Good point (in theory) but it would make the limits of the lens all the more obvious. I can only refer to the example of the Nikon 14-24G lens, which looked outstanding (corner to corner) on a 24 mpx sensor, but less so on a D800. So at this point, throwing a higher res sensor at the lens will only make the flaws more obvious.

I did not see the example , but the 24 Mp sensor would be APS-C (DX) which does not use the corners of the lens. That 24 Mp sensor pixel density would give 56 Mp FF. A 56 Mp FF sensor should look "outstanding" with that lens, except in the corners.

Dan

The camera I was referring to was the D3X, which is FF 24 mpx, so yes, it does use the corners. The lens in very good on the D800/E, but not sure I would describe it as "outstanding".

I forgot about that one. Was it only in the corners the D800/E was not as good as D3X?

Great Bustard Forum Pro • Posts: 41,906
Re: Per-pixel comparisons are meaningless...
1

SubPrime wrote:

Great Bustard wrote:

...for photos made from different numbers of pixels. In any case, rather than rebut you point-by-point (obviously a "pointless" exercise), I'll just ask you a single, and simple, question:

Would a landscape photographer produce "better" photos with FF or APS-C, or does it not matter at all?

Let us know.

It shouldn't matter in the end other than the fact the FF sensor will capture a better quality image in the end - provided the lenses are up to the task.

How is that so, if the corners are not as good as crop, as you claim?  Are you saying that the corners don't matter for landscapes?  Or are you saying that the other advantages of FF matter more than the corners do?

David Hull
David Hull Veteran Member • Posts: 6,201
Re: Canon’s Testing a 75MP+ Pro DSLR

SubPrime wrote:

David Hull wrote:

What people seem to miss here is that the lens is what it is and the resolution of the sensor really has no influence on that. The lens has a special frequency response (in lines/mm). If the camera resolution is significantly more than that then the camera resolution won’t significantly influence the resolution of the cascaded system (sensor+lens).

Good point (in theory) but it would make the limits of the lens all the more obvious. I can only refer to the example of the Nikon 14-24G lens, which looked outstanding (corner to corner) on a 24 mpx sensor, but less so on a D800. So at this point, throwing a higher res sensor at the lens will only make the flaws more obvious.

Your theory also excludes the diminishing returns that come with squeezing that many photo sites onto a 35 mm sensor.

OK, but what you seem to be advocating amounts to masking the limits of the lens with limitations in the camera. I assume that they wouldn't be putting 75 MP on the sensor if they couldn't get it to work -- assuming they are actually putting 75MP on a 35mm sensor (it is a only a rumor we are discussing )

It is as if you played your CD player through a phone line it would have a significant influence on the sound you heard. If on the other hand you played it through a channel that had bandwidth out to 500 kHz all you would hear is how good the CD machine sounded.

False analogy because our only trained ears will pick up the inadequacies of the CD machine, whereas untrained eyes would easily pick u the inadequacies of the lens.

I disagree. If you play it through a 300 Hz to 3 kHz phone line, even an untrained ear will hear that something is amiss (my guess that would be true even if you are playing Metallica on the CD machine). It is a pretty easy experiment to run, actually.

I don’t get the “sensor out resolves the lens thing”; ideally that is what you want at least up to the point where you reach some point of diminishing returns and are wasting money on the sensor (as in my Audio example above).

I agree. And it it more than likely that this will be

Even 75MP would be more like a mosaic than an analog image of the lens' projection, for a moderately sharp lens.

Wrong. Before you get to the mosaic, you begin to see the inadequacies of most lenses.

Of course, but the imperfections of the sensor are gone. You can finally toss the AIF for example (since the lens has become one).

Fair enough, but with the imperfections of the sensor gone, you then have to deal with the imperfections of the lens.

-- hide signature --
 David Hull's gear list:David Hull's gear list
Canon EOS 5D Mark III Canon EOS M5
David Hull
David Hull Veteran Member • Posts: 6,201
Re: 75MP+ Pro DSLR: for architecture and landscape?

SubPrime wrote:

David Hull wrote:

SubPrime wrote:

NancyP wrote:

This sounds like a nightmare for action, wildlife, and PJ photographers, and a serious headache for wedding, fashion, and portrait photographers. One frame per second, maybe?

Of course. Do you really think Canon would release a 1DX killer?

I don't think you need to worry about that. The 1Dx is designed to maximize AF performance and firing rate. Anything with 75MP is not likely to match the 1Dx in the latter.

I agree entirely, which is why I was bemused by the suiggestiont that such a camera would be any use to an action photographer, PJ photographers, or wedding photographer.

This just in (fom Canon Rumors):

We have confirmed that a test camera with around 80mp is definitely in the wild. A few things to note about the camera. It’s apparently in an EOS-1 body and has a larger than 3.2″ LCD on the back. The camera also apparently shoots at a very high frame rate.

Ha... shows what we know

-- hide signature --
 David Hull's gear list:David Hull's gear list
Canon EOS 5D Mark III Canon EOS M5
bobn2
bobn2 Forum Pro • Posts: 58,081
Re: Canon’s Testing a 75MP+ Pro DSLR
3

altair8800 wrote:

SubPrime wrote:

altair8800 wrote:

SubPrime wrote:

David Hull wrote:

What people seem to miss here is that the lens is what it is and the resolution of the sensor really has no influence on that. The lens has a special frequency response (in lines/mm). If the camera resolution is significantly more than that then the camera resolution won’t significantly influence the resolution of the cascaded system (sensor+lens).

Good point (in theory) but it would make the limits of the lens all the more obvious. I can only refer to the example of the Nikon 14-24G lens, which looked outstanding (corner to corner) on a 24 mpx sensor, but less so on a D800. So at this point, throwing a higher res sensor at the lens will only make the flaws more obvious.

I did not see the example , but the 24 Mp sensor would be APS-C (DX) which does not use the corners of the lens. That 24 Mp sensor pixel density would give 56 Mp FF. A 56 Mp FF sensor should look "outstanding" with that lens, except in the corners.

Dan

The camera I was referring to was the D3X, which is FF 24 mpx, so yes, it does use the corners. The lens in very good on the D800/E, but not sure I would describe it as "outstanding".

I forgot about that one. Was it only in the corners the D800/E was not as good as D3X?

It is always and everywhere better on the D800 than on the D3X.

-- hide signature --

Bob

Press Correspondent
Press Correspondent Veteran Member • Posts: 3,345
Re: Tp Press Correspondent

Rick Knepper wrote:

Yes, I have a D800E. I would replace my 5D2 with a Canon version of the D800E.

And, since I have a D800E, I have the freedom to pass on any behemoth that does not meet my criteria i.e. first and foremost, a 1 series body, probably 2nd, the usual hefty 1 series price tag for hi rez.

Good strategy, provided you have (or rent) lenses for both systems. I couldn't justify the second set of lenses and had to let the Nikon go. This was my D800 mandatory cat last year. I like the way the 60/2.8 Macro melts the warm concrete:

 Press Correspondent's gear list:Press Correspondent's gear list
Olympus XZ-1 Canon EOS 5D Mark III Canon EF 24-70mm F2.8L II USM Canon EF 17-40mm f/4.0L USM Canon EF 70-200mm F4L USM +11 more
John Sheehy Forum Pro • Posts: 20,555
Re: The question is...
3

SubPrime wrote:

Yes, adding pixels does increase resolution, but if the lenses are not up to the task , you're simply resolving the flaws.

That sounds like good logic, except for the fact that along with resolving more flaws, you are resolving more details.  You never get more or better details by resolving less.  Compared side-by-side with a lower-pixel-density image with the same subject magnification, a flaw-exposing high-pixel-density looks superior.

The "flaw" in your point of view is that it totally ignores the overriding "flaw" of low resolution.

Also, an oversampled image is virtually indestructible, as far as geometric manipulations go.

jayrandomer
jayrandomer Contributing Member • Posts: 845
Only if you're using it to take pictures
1

Bobn2 wrote:

altair8800 wrote:

SubPrime wrote:

altair8800 wrote:

SubPrime wrote:

David Hull wrote:

What people seem to miss here is that the lens is what it is and the resolution of the sensor really has no influence on that. The lens has a special frequency response (in lines/mm). If the camera resolution is significantly more than that then the camera resolution won’t significantly influence the resolution of the cascaded system (sensor+lens).

Good point (in theory) but it would make the limits of the lens all the more obvious. I can only refer to the example of the Nikon 14-24G lens, which looked outstanding (corner to corner) on a 24 mpx sensor, but less so on a D800. So at this point, throwing a higher res sensor at the lens will only make the flaws more obvious.

I did not see the example , but the 24 Mp sensor would be APS-C (DX) which does not use the corners of the lens. That 24 Mp sensor pixel density would give 56 Mp FF. A 56 Mp FF sensor should look "outstanding" with that lens, except in the corners.

Dan

The camera I was referring to was the D3X, which is FF 24 mpx, so yes, it does use the corners. The lens in very good on the D800/E, but not sure I would describe it as "outstanding".

I forgot about that one. Was it only in the corners the D800/E was not as good as D3X?

It is always and everywhere better on the D800 than on the D3X.

If you're using it to look at a series of 100% crops it will probably be worse, although you will have more 100% crops.

 jayrandomer's gear list:jayrandomer's gear list
Canon EOS 550D Canon EOS 6D Canon EF 100mm f/2.8 Macro USM Canon EF 24-105mm f/4L IS USM Canon EF 100-400mm F4.5-5.6L IS II +3 more
bobn2
bobn2 Forum Pro • Posts: 58,081
Re: Only if you're using it to take pictures
2

jayrandomer wrote:

Bobn2 wrote:

altair8800 wrote:

SubPrime wrote:

altair8800 wrote:

SubPrime wrote:

David Hull wrote:

What people seem to miss here is that the lens is what it is and the resolution of the sensor really has no influence on that. The lens has a special frequency response (in lines/mm). If the camera resolution is significantly more than that then the camera resolution won’t significantly influence the resolution of the cascaded system (sensor+lens).

Good point (in theory) but it would make the limits of the lens all the more obvious. I can only refer to the example of the Nikon 14-24G lens, which looked outstanding (corner to corner) on a 24 mpx sensor, but less so on a D800. So at this point, throwing a higher res sensor at the lens will only make the flaws more obvious.

I did not see the example , but the 24 Mp sensor would be APS-C (DX) which does not use the corners of the lens. That 24 Mp sensor pixel density would give 56 Mp FF. A 56 Mp FF sensor should look "outstanding" with that lens, except in the corners.

Dan

The camera I was referring to was the D3X, which is FF 24 mpx, so yes, it does use the corners. The lens in very good on the D800/E, but not sure I would describe it as "outstanding".

I forgot about that one. Was it only in the corners the D800/E was not as good as D3X?

It is always and everywhere better on the D800 than on the D3X.

If you're using it to look at a series of 100% crops it will probably be worse, although you will have more 100% crops.

100% crops means that you're comparing different size images.

-- hide signature --

Bob

qianp2k Forum Pro • Posts: 10,350
Re: The question is...

John Sheehy wrote:

SubPrime wrote:

Yes, adding pixels does increase resolution, but if the lenses are not up to the task , you're simply resolving the flaws.

That sounds like good logic, except for the fact that along with resolving more flaws, you are resolving more details. You never get more or better details by resolving less. Compared side-by-side with a lower-pixel-density image with the same subject magnification, a flaw-exposing high-pixel-density looks superior.

The "flaw" in your point of view is that it totally ignores the overriding "flaw" of low resolution.

Also, an oversampled image is virtually indestructible, as far as geometric manipulations go.

However if large portion of those excessive pixels fall below MTF 20 resolution then it's largely irrelevant and hardly visible.  The bottleneck is still lenses.

-- hide signature --
jayrandomer
jayrandomer Contributing Member • Posts: 845
It depends on how it's done.

Bobn2 wrote:

jayrandomer wrote:

Bobn2 wrote:

altair8800 wrote:

SubPrime wrote:

altair8800 wrote:

SubPrime wrote:

David Hull wrote:

What people seem to miss here is that the lens is what it is and the resolution of the sensor really has no influence on that. The lens has a special frequency response (in lines/mm). If the camera resolution is significantly more than that then the camera resolution won’t significantly influence the resolution of the cascaded system (sensor+lens).

Good point (in theory) but it would make the limits of the lens all the more obvious. I can only refer to the example of the Nikon 14-24G lens, which looked outstanding (corner to corner) on a 24 mpx sensor, but less so on a D800. So at this point, throwing a higher res sensor at the lens will only make the flaws more obvious.

I did not see the example , but the 24 Mp sensor would be APS-C (DX) which does not use the corners of the lens. That 24 Mp sensor pixel density would give 56 Mp FF. A 56 Mp FF sensor should look "outstanding" with that lens, except in the corners.

Dan

The camera I was referring to was the D3X, which is FF 24 mpx, so yes, it does use the corners. The lens in very good on the D800/E, but not sure I would describe it as "outstanding".

I forgot about that one. Was it only in the corners the D800/E was not as good as D3X?

It is always and everywhere better on the D800 than on the D3X.

If you're using it to look at a series of 100% crops it will probably be worse, although you will have more 100% crops.

100% crops means that you're comparing different size images.

Yes, of course.  I thought the silliness of "only if you're using it take pictures" would be a tip off that it was a somewhat humorous nature of the reply.

What I meant was that in LR (or similar programs), where the tendency is to look at the image at 1:1 to see how sharp and detailed it is, the higher resolution images tend to look worse.  In an X:Y comparison, then, the lower resolution image will look sharper, which you notice, and zoomed a bit further out, which you tend not to notice.  So if you use your camera to evaluate the quality of 100% crops (which is what LR can be used to do), the lower resolution will often look better.  Of course, if you use comparable images (which LR isn't designed to do) the higher resolution camera will generally do better.

 jayrandomer's gear list:jayrandomer's gear list
Canon EOS 550D Canon EOS 6D Canon EF 100mm f/2.8 Macro USM Canon EF 24-105mm f/4L IS USM Canon EF 100-400mm F4.5-5.6L IS II +3 more
Keyboard shortcuts:
FForum MMy threads