Still waiting for high MP canon while Nikon is coming out with new 800

[timcz]

I have to confess as a landscape shooter I wouldn't mind a few extra mp, purely so when I crop down into a panorama format, I still have a decent mp count to work with. Currently I stitch a lot, but some scenes (seascapes in particular) this can be difficult. I have printed 30" prints off my old nikon d70 back in the day which worked well enough - but I guess the less upscaling you have to do the better.

I do love my 5dmk3 though, but still often shoot my 1ds3 as I feel at iso 100 the files can be pushed a little more than the 5d in the shadows.

 

[ksagomonyants]

Guys, a comment and a question at the same time. For those who crop or want to print large images and doesn't want to get d800 (for whatever reason), would it make sense to buy software like Perfect Resize/BlowUp/Photozoom, etc.? Has anybody had good experience with either software? Thanks.

 

[Synkka]

I really don't think there is a strong demand for a higher mega pixel body I think it is more that if more mega pixels are on offer they would take it.
I do a bit of varied shooting, with wildlife I probably see 1 out of every 20 users is a Nikon user, never any d800s and never any high end telephotos.
With weddings maybe 1 in 5 and Nikon users tend to have a d800 and a d600.
Landscape I do see a more even split of Nikon users but this is also where I see the least photographers.

So overall I really don't think there is any application for a high megapixel body outside of landscapes and I also don't think landscape photography is as popular as some people think.

Also in regards to large printing, just how close do you view prints ? For big prints I normally place them so you can stand back and enjoy them and from a distance it's tough to tell high megapixel prints apart.

 

[privatebydesign]

I asked, why can not Canon introduce a high mega pixel camera as Nikon and Sony. Not if any brand are out selling another. So why?

And about signal noise and if the cameras file size are compared at the same file size, there are not much to discusse, they are equal regarding higher iso. So my question is, why can not Canon introduce a high megapixel camera year 2014 ?

Similar questions were asked in Nikon forums prior to them using Sony's 36MP sensor and Nikon folks used all of the same arguments here that Canon guys do about why they didn't need it.

I'm a firm believer in the fact that Canon have been working down a different path for sensor R&D in the belief that their current design pattern for full frame sensors was good enough and that Sony's advances with the 36MP sensor caught Canon by surprise.

But, Nikon users had a mere 12MP in their FF cameras, the D3/D3S and the 5D MkII competition D700, until the $7,000 D3X (the 1Ds MkIII competitor).

There is a huge difference in usable resolution between 12MP and 21MP files, there is not such a huge leap between 24MP (the current 5D MkIII) and the 36MP Sony sensor.

 

[RLPhoto]

I've waited a long time for Nikon to recapture the lead on Nikon and their 36 mp camera. And now I read that Nikon will introduce an upgrade to the 800e in June. Why am I not jumping ship? I do have a lot of canon glass and have been a local customer. I own the mk III and 6d and I love the images .BUTT, COME ON ALREADY cCanon. Get your stuff together!

Technically, The new D800s would finally be that d700 replacement, Alas with Small RAW and Decent FPS options. I'd imagine this would please the majority of d700 users who complained about the 36 megapixels files.

I stayed with canon because - Lens selection > Sensors.

 

[Loswr]

I'm a firm believer in the fact that Canon have been working down a different path for sensor R&D in the belief that their current design pattern for full frame sensors was good enough and that Sony's advances with the 36MP sensor caught Canon by surprise.

Yes, Canon was caught totally by surprise. I mean, it's not as if big, research based companies have Competitive Intelligence departments to ferret out what their competitors are doing. Surely not. Well, ok, mine does...but they must be the only ones.

Canon was so surprised by the D800 that they had no choice but to go ahead with the 5DIII launch, even knowing that a 36 MP body in the same price class would absolutely trounce the 5DIII in sales. There Canon was, thinking an awesome general purpose camera combining great IQ, class-leading AF, and a decently fast frame rate would be popular, but then, like a 1-2 punch, after the D800, they get hit by Sony's 36 MP mirrorless with great ergonomics, excellent battery life, and a stellar native lens collection.

Obviously, Canon is feeling the pressure...they have predicted a drop in dSLR sales this year. Nikon's great sensors mean they they'll do much better than Canon this year. At least, I think that's what they predicted...I mean, 6 is a bigger number than 1, so Nikon's prediction of a 6% loss in dSLR sales is better than Canon's predicted 1% loss. Right?

 

[100]

There is a huge difference in usable resolution between 12MP and 21MP files, there is not such a huge leap between 24MP (the current 5D MkIII) and the 36MP Sony sensor.

Could you explain that?
I mean, when I look at the math I don’t see “a huge difference in resolution”.
1Ds mk III => 5616 × 3744 versus D3S => 4256 × 2832
D800 => 7360 x 4912 versus 5D mk III => 5760 × 3840

1Ds mk III versus D3s => 3744/2823 = 1.32
D800 versus 5D mk III => 4912/3840 = 1.28

It’s 32% more resolution versus 28% more resolution.
However, you wrote usable resolution so maybe the usability diminishes with the increase in mp and if so I like to understand why that is.

 

[Loswr]

There is a huge difference in usable resolution between 12MP and 21MP files, there is not such a huge leap between 24MP (the current 5D MkIII) and the 36MP Sony sensor.

Could you explain that?
I mean, when I look at the math I don’t see “a huge difference in resolution”.
1Ds mk III => 5616 × 3744 versus D3S => 4256 × 2832
D800 => 7360 x 4912 versus 5D mk III => 5760 × 3840

1Ds mk III versus D3s => 3744/2823 = 1.32
D800 versus 5D mk III => 4912/3840 = 1.28

It’s 32% more resolution versus 28% more resolution.
However, you wrote usable resolution so maybe the usability diminishes with the increase in mp and if so I like to understand why that is.

Why would you consider just the vertical/linear resolution?

5DII vs D700: (21 – 12) / 12 = 75% increase
D800 vs 5DIII: (36 – 22) / 22 = 64% increase

That's sensor resolution. The 'usable' part comes when you consider that there's a lens attached to the camera.

From DxOMark:
D700 with Nikon 24-70/2.8G – 9 P-Mpix
5DII with Canon 24-70/2.8L – 12 P-Mpix

D800 with Nikon 24-70/2.8G – 15 P-Mpix
5DIII with Canon 24-70/2.8L – 14 P-Mpix

So, with a very popular professional wedding/event lens from each brand, the 21 MP 5DII delivers 33% more usable resolution than the 12 MP D700, and the 36 MP D800 delivers only 7% more usable resolution than the 22 MP 5DIII. Those values are for the original version of the Canon 24-70/2.8, if you look at the new MkII version...

5DIII with Canon 24-70/2.8L II – 18 P-Mpix

So when you factor in the current 24-70/2.8 lenses, the D800 actually delivers 17% less usable resolution than the 5DIII, despite it's 64% higher sensor resolution.

If you instead compare the D800E + 24-70/2.8G, it fares better...but at 21 P-Mpix is still only delivering a 17% usable resolution increase over the 5DIII, nowhere near 64%.

Obviously, other lenses can be chosen with different results. But the key point is that what really matters is the system resolution (camera + lens). Given the resolution of most lenses available for both brands, going from the low teens to the twenties in MP count makes a lot more difference in output resolution than going from the twenties to the thirties.

 

[100]

Why would you consider just the vertical/linear resolution?

5DII vs D700: (21 – 12) / 12 = 75% increase
D800 vs 5DIII: (36 – 22) / 22 = 64% increase

Spatial resolution (how closely lines can be resolved) seems a good way to measure resolution if you look at test charts for comparisons, that’s why I used it in the calculation.
I’m ok with pixel resolution though, the difference between 75% and 64% increase isn’t that big either.

The 'usable' part comes when you consider that there's a lens attached to the camera.

From DxOMark:
D700 with Nikon 24-70/2.8G – 9 P-Mpix
5DII with Canon 24-70/2.8L – 12 P-Mpix

D800 with Nikon 24-70/2.8G – 15 P-Mpix
5DIII with Canon 24-70/2.8L – 14 P-Mpix

So, with a very popular professional wedding/event lens from each brand, the 21 MP 5DII delivers 33% more usable resolution than the 12 MP D700, and the 36 MP D800 delivers only 7% more usable resolution than the 22 MP 5DIII. Those values are for the original version of the Canon 24-70/2.8, if you look at the new MkII version...

5DIII with Canon 24-70/2.8L II – 18 P-Mpix

So when you factor in the current 24-70/2.8 lenses, the D800 actually delivers 17% less usable resolution than the 5DIII, despite it's 64% higher sensor resolution.

If you instead compare the D800E + 24-70/2.8G, it fares better...but at 21 P-Mpix is still only delivering a 17% usable resolution increase over the 5DIII, nowhere near 64%.

Obviously, other lenses can be chosen with different results. But the key point is that what really matters is the system resolution (camera + lens). Given the resolution of most lenses available for both brands, going from the low teens to the twenties in MP count makes a lot more difference in output resolution than going from the twenties to the thirties.

I agree it’s the total system that makes the picture, so if we want to compare resolution it’s best to use the same glass.
The Sigma 35mm F1.4 DG HSM A on a D3s => 10P-Mpix
The Sigma 35mm F1.4 DG HSM A on a 1Ds MkIII => 17P-Mpix

The Sigma 35mm F1.4 DG HSM A on a 5D MkIII => 19P-Mpix
The Sigma 35mm F1.4 DG HSM A on a D800 => 23P-Mpix

70% versus 21%, that confirms the gap in usable resolution between 12mp and 21mp is larger than between 22mp and 36mp if we assume DxO’s P-Mpix value are a good indication for the difference in (spatial) resolution.
A 1mp difference between the 1Ds MkIII and the 5D MkIII results in a 2P-Mpix points difference, so a grain of salt is need with this P-Mpix value.

My question stays the same, I like to understand why that is and what it will mean if we get sensors with even more megapixels?

 

[unfocused]

I'm a firm believer in the fact that Canon have been working down a different path for sensor R&D in the belief that their current design pattern for full frame sensors was good enough and that Sony's advances with the 36MP sensor caught Canon by surprise.

Yes, Canon was caught totally by surprise. I mean, it's not as if big, research based companies have Competitive Intelligence departments to ferret out what their competitors are doing...

Canon was so surprised by the D800 that they had no choice but to go ahead with the 5DIII launch, even knowing that a 36 MP body in the same price class would absolutely trounce the 5DIII in sales. There Canon was, thinking an awesome general purpose camera combining great IQ, class-leading AF, and a decently fast frame rate would be popular, but then, like a 1-2 punch, after the D800, they get hit by Sony's 36 MP mirrorless with great ergonomics, excellent battery life, and a stellar native lens collection.

Obviously, Canon is feeling the pressure...so Nikon's prediction of a 6% loss in dSLR sales is better than Canon's predicted 1% loss. Right?

Neuro, you missed a very important point.

Not only was Canon forced to go ahead with the 5DIII launch but in order to beat the D800 in sales, they had to charge 20% more for the 5DIII. Canon made such a terrible mistake with the 5DIII that it's ranked as the seventh highest seller on the Amazon list, even though it sells for twice what the second most expensive camera in the top 20 sells for which...wait for it...happens to be the Canon 6D.

Funny how Canon has fallen so far behind that the latest flagship releases from Sony (A7s) and Nikon (D4s) have substantially less resolution than even the 5DIII.

Oh and maybe Sony just shaved $200 off the price of the A7 because it was selling so well.

 

[Loswr]

A 1mp difference between the 1Ds MkIII and the 5D MkIII results in a 2P-Mpix points difference, so a grain of salt is need with this P-Mpix value.

Perhaps. But consider the differences that can be observed in P-Mpix values for the same lens compared on the 36 MP D800 vs. the 36 MP D800E (e.g., a 6 P-Mpix = 40% difference for the 24-70/2.8G, which isn't the sharpest of lenses). Given that, it would not be at all surprising that a slight change in AA filter design between camera generations could result in a sharpness difference beyond what you'd predict from the change in MP count.

My question stays the same, I like to understand why that is and what it will mean if we get sensors with even more megapixels?

It is that way because in most cases resolution is limited by the lens more than by the sensor. If we get higher MP sensors without a corresponding increase in lens resolution, we'll see diminishing marginal returns in output resolution as MP counts increase.

DxOMark's P-Mpix values can illustrate this, in part. Comparing the calculated P-Mpix value with the real sensor MP count gives an approximation of resolution lost 'to the system' (lens, AA filter, etc.). With the very sharp 35A, the 5DIII loses 14% of its theoretically possible resolution, the D800 loses 36%. Sharp as the lens is, you're still seeing the diminishing return with increasing MP counts.

Keep in mind that what really matters isn't the MP count, but pixel density (though MP count works when discussing same-sized sensors). An 18 MP APS-C sensor is far more limited by lens resolution than an 18 MP FF sensor. That means even if/when FF sensors get into the MP count range of medium format sensors, the output resolution of MF will still be much higher.

 

[privatebydesign]

For usable resolution it isn't even just about the system (sensor, AA filter, lens, firmware etc), that just gives you a potential resolution in best case scenario situations that most of us practically never shoot in, and require 10x Live View manual focus, optimal aperture for lens sharpness, minimum or base iso (if we take the time and trouble to work out our sensors true base iso), very good light with high contrast, very firm support, cable release etc etc.

Now that list sounds like a landscape and architectural shooters M.O., and I believe they are the only people who will get any remote chance of worthwhile resolution increases when going over 30MP in the 135 format etc. Having said that I do a reasonable amount of higher end real estate work that is regularly used in quality print advertising as well as posters and billboards and I haven't found 21MP to be a serious limitation. Peolpe who use AF, BIF, sports shooters, action shooters, most wedding shooters etc will get practically nothing from a 30+MP sensor over current models.

After some extensive testing I did with the 1Ds MkIII and the 7D (sure things are better now but they are better for both formats) in ideal shooting situations set up to maximise the difference between the 21MP sensor and the effective 46MP sensor that showed minimal resolution differences, I concluded the numbers mean very little, in real world shooting situations where I was using AF and not optimal iso and aperture settings the differences disappeared completely. In truth AF had a far bigger impact on resolution than a 21 or 46 MP sensor.

If I shot landscape 100% of the time I'd use TS-E's on an A7R where the methodology of realising the potential resolution wouldn't impact my shooting, but I don't, I am a generalist and need AF, a range of ISO's, apertures, and shutter speeds, zoom lenses, etc etc.

In conclusion, I am not saying, and never have, that a higher MP sensor doesn't resolve "more", it does, but the rule of diminishing returns kicks in for virtually all real world shooting scenarios and makes the difference so small as to be imperceptible most of the time. Having something because it has a higher number is a game I stopped playing long ago, I look very closely at what it can actually do for me, so far the disadvantages of higher MP sensors in 135 format have not convinced me that Canon don't know far more about this than us and they hit a sweet spot with the 5D MkIII. I am a long time 1Ds MkIII user, if they come out with a true replacement in the 35-45MP range I am not interested, if they come out with a 1DX MkII with a 24MP sensor I died and went to camera heaven.

 

[privatebydesign]

Of course you're welcome to tell me that bigger isn't better

Bigger is only better if the smaller thing you are enlarging more is better than the native big thing.

You are not arguing bigger is better, you are saying smaller (pixels) are as good as bigger ones, that has been demonstrated repeatedly to not be the case.

 

[GMCPhotographics]

There is a huge difference in usable resolution between 12MP and 21MP files, there is not such a huge leap between 24MP (the current 5D MkIII) and the 36MP Sony sensor.

Could you explain that?
I mean, when I look at the math I don’t see “a huge difference in resolution”.
1Ds mk III => 5616 × 3744 versus D3S => 4256 × 2832
D800 => 7360 x 4912 versus 5D mk III => 5760 × 3840

1Ds mk III versus D3s => 3744/2823 = 1.32
D800 versus 5D mk III => 4912/3840 = 1.28

It’s 32% more resolution versus 28% more resolution.
However, you wrote usable resolution so maybe the usability diminishes with the increase in mp and if so I like to understand why that is.

Why would you consider just the vertical/linear resolution?

Whether it is vertical or horizontal doesn't matter - the number of dots in each direction defines how large your "print" can be and given both sensors have the same geometry, the comparison between verticals should be the same as the comparison between horizontals. The total number of pixels is not really that relevant.

24.5"x16.4" vs 19.2"x12.8" (@300dpi)

Of course you're welcome to tell me that bigger isn't better

Well....who would have guessed it? 36mp and the thing holding the resolution back is the lenses attached...oh actually....we've all been saying that all along. 36mp of the sharpest 22mp optical / lens resolution can muster.

As to printability....the 300dpi rule is only a guide line. I have an A1+ print on my wall from a 5D classic and it's pin sharp and looks great.

 

[Loswr]

Of course you're welcome to tell me that bigger isn't better

More real resolution is better. More megapixels are not necessarily better.

I use Zeiss microscopes at work (I'm in drug discovery, BTW), some of the cameras on them have 1.3 MP sensors with a CFA but no microlenses, and the camera can move the sensor with piezo motors in full-pixel or sub-pixel increments. Color can be interpolated, or with full-pixel moves each pixel can see R/G/B with no interpolation. Sub-pixel moves shift the position of the photodiode within the area of each pixel (microlenses would obviate that benefit), so the 1.3 MP sensor can sample a 2x2 sub-pixel array for a 5 MP image or a 3x3 array for a 12 MP image, and the 5 and 12 MP images are real resolution. Zeiss also provides information for their objective lenses, indicating the digital resolution necessary to fully capture the available optical resolving power of the lens. For some lenses, the 12 MP image is needed. For other lenses, even a VGA camera (0.3 MP) can fully capture the spatial resolution delivered by the lens, so a 1 MP image is more than adequate (even allowing for oversampling according Nyquist). Those lenses for which 1 MP or 5 MP are sufficient to fully capture their optical resolution can still be used with a 5 or 12 MP capture, but using a higher resolution than the lens can deliver doesn't capture any additional information. That's referred to as 'empty resolution', all it adds is 'overhead' – needlessly larger files, longer acquisition time, etc.

So in fact, 'bigger' (meaning more MP) isn't always better.

 

[LightandMotion]

Guys, a comment and a question at the same time. For those who crop or want to print large images and doesn't want to get d800 (for whatever reason), would it make sense to buy software like Perfect Resize/BlowUp/Photozoom, etc.? Has anybody had good experience with either software? Thanks.

For all my prints (typically 1m / 40" conventional or 1.5m / 60" for panoramas) I use Perfect Resize whether the images are from a 1dx or A7R. I've had libraries and and restaurants order prints for wall type murals and have printed to 5m using Perfect Resize with excellent results. Even a 36MP A7R image will only print to about 25" at 300 dpi, so anything above this will need software to resize. I've found the genuine fractal algorithm of Perfect Resize superior to other software and to PS own bicubic smoother.

 

[privatebydesign]

Of course you're welcome to tell me that bigger isn't better

Bigger is only better if the smaller thing you are enlarging more is better than the native big thing.

You are not arguing bigger is better, you are saying smaller (pixels) are as good as bigger ones, that has been demonstrated repeatedly to not be the case.

So the 1Ds and original 5D have better IQ than the current 1DX, 5D3 and 6D?

So a 70D pixel enlarges as well as a 1DX pixel?

If you look at same generation sensors on a per pixel basis bigger pixels have always performed better than smaller ones.

 

[dslrdummy]

I've waited a long time for Nikon to recapture the lead on Nikon and their 36 mp camera. And now I read that Nikon will introduce an upgrade to the 800e in June. Why am I not jumping ship? I do have a lot of canon glass and have been a local customer. I own the mk III and 6d and I love the images .BUTT, COME ON ALREADY cCanon. Get your stuff together!

the D800 update is marginal.. so what?

if you complainers really need more MP and more details.. buy MF... i have.

I wish I could afford MF,,I should of done it years ago..But I their is economics and loyalty involved here,,,especially when I have 5 canon lens, a 5dmk III and a 6d ,not to mention other canon accessories.

And when you want to print to 40x60, its an issue..36 Mp would give that edge i needed. I should have jumped to nikon when the 800e came out,,but I didn't because I was convinced Canon would not sit idly bye trying to figure out how to recapture the glory days of the 5D Mk II.. That was a great camera at the time!!

And before that the 5D, the first "affordable" full frame dslr.

 

[GaryJ]

I'd rather get a PhaseOne and be done with it 8)

 

[tinkertinker]

hello world.

RIGHT OR WRONG?
i think that´s not the question.

many of you (like me) want a highmegapixel camera from canon!
as a studio and architecture photographer i would appreciate
more megapixel than 21...(21MP is quite enough but i would welcome
about30-40MP), because i (hope to) know how to use it, and I have the fantatic TSe lenses.
but, i hope i won´t be more than 50MP because that would
be way too much for the small sensor and best lenses used with perfect aperture.

@neuro: i really love to read your posts but sometimes
so please.... not every photographer is doing weddings, sports or events!!
and sure does the 5diii outsell the d800, because its an allroundcamera!!
and to be honest the company canon is a little bigger than nikon...
cheers.