Ad
Ad
Ad
Pages: « 1 ... 6 7 [8] 9 »   Bottom of Page
Print
Author Topic: Canon 5D Mark III official...  (Read 27697 times)
Hans Kruse
Sr. Member
****
Offline Offline

Posts: 706



WWW
« Reply #140 on: March 04, 2012, 02:15:18 PM »
ReplyReply

Not sure what you guys are trying to prove: that more megapixels is pointless because when shot at f/22 the resolution will not be better than at f/8 at less megapixels?

The way I read it is that up to f/22 there will be resolution advantage of more megapixels. Not to mention that no self-respecting digital photographer will shoot at f/22 if he can get away with shooting with f/8 or in-between.

Well, basically DxO has measurements by which one can assess the resolution by fstop and camera sensor megapixels. They show that doubling the megapixels are negated by stopping down from f/8 to f/22 caused by diffraction. They also show that at f/22 a doubling of megapixels do increase the resolution which is in contradiction to the article linked to. So it raises some questions:

1) What level of resolution increase will an increase from the 21MP of the 1Ds mkIII or the 22MP of the 5D mkIII up to the 36MP of the D800 give when stopped down to e.g. f/16.

2) What change in the resolution tables (table 3 in the linked to article on Lula) should be done to reflect reality rather than a seamingly over conservative calculation?

I have for some time relied (maybe naively as I haven't done any real world resolution tests to check it) on the accuracy of this table and I certainly have to revise my opinion on that. I have done test shots stopping down and seen the lost of resolution on the screen on real world subjects but that is hard to translate into excact loss of resolution in terms of megapixels or lp/mm. So only recently did I check the DxO measurements in the way shown in this thread.
Logged

dseelig
Sr. Member
****
Offline Offline

Posts: 445


« Reply #141 on: March 04, 2012, 02:29:36 PM »
ReplyReply

Well t seems this is not the camera for a lot of you landscape shooters. I am more all around shooter, low light street work and I need a low ligth camera then my m9 and lighter then the mk1v . the 5d mk11 is useless as the focus in f1.4 60 th sec when you have to got iso 6400 is a joke and it simply does not mf very well. so I am happy with better low light and real af.  But you never know if canon has something else to come out later for you landscape guys
Logged
Wayne Fox
Sr. Member
****
Offline Offline

Posts: 2846



WWW
« Reply #142 on: March 04, 2012, 03:08:01 PM »
ReplyReply

Dealers have known for several days, but today it's official ... the 5D Mark 2 has just had it's price reduced to $2199 and is not being discontinued.  Same strategy as Nikon with d700.
Logged

Slobodan Blagojevic
Sr. Member
****
Offline Offline

Posts: 5890


When everybody thinks the same... nobody thinks.


WWW
« Reply #143 on: March 04, 2012, 03:14:54 PM »
ReplyReply

Well t seems this is not the camera for a lot of you landscape shooters...

I think you are right. It seems geared toward two distinct categories of users: wedding photographers (note new flash system) and videographers. In contrast, Mark II was quickly embraced by landscape photographers and raised to the status of a preeminent landscape camera. Hence this disappointment with Mark III, which was expected to raise the bar in a similar fashion Mark II raised it compared to the original 5D. Instead, it is another textbook example of "too little, too late".

I smelled the problem a couple of years ago when Canon refused to innovate in order to keep the distance it initially created vs. Nikon and others in digital quality, and instead started using bean-counters' mantra "why would we do that, when competition is still playing catch up?" (my recollection of a Chuck Westfall's interview). Well, problem with that philosophy is that competition is not only aiming to catch up, but to surpass. And they did. Translucent mirrors (Sony), high-iso quality (Nikon), dynamic range (Pentax, Sony, Nikon), megapixels, mirrorless bodies, etc. Nikon 14-24 zoom is often used by discriminating Canon users, with an adapter, of course.

"Asleep at at wheel" phrase comes to mind. Or "resting on laurels".
Logged

Slobodan

Flickr
500px
Ray
Sr. Member
****
Offline Offline

Posts: 8884


« Reply #144 on: March 04, 2012, 03:43:55 PM »
ReplyReply

Interesting to see that a Nikon 50mm lens at f/8 on D3 resolves exactly as f/16 on D3X. So basically doubling pixel count on the D3X is eliminated by stopping down from f/8 to f/16.

Not so. You've missed something very crucial. Whilst it may be true that the D3 at F8 resolves as much as the D3X at F16, this only applies at the plane of focus. Away from the plane of focus, the D3X at F16 resolves much more than the D3 at F8.

Also, I'm pretty sure that the D3X at F16 and at the plane of focus resolves significantly more than the D3 does at F16 and at the plane of focus.
Logged
Hans Kruse
Sr. Member
****
Offline Offline

Posts: 706



WWW
« Reply #145 on: March 04, 2012, 04:32:40 PM »
ReplyReply

Not so. You've missed something very crucial. Whilst it may be true that the D3 at F8 resolves as much as the D3X at F16, this only applies at the plane of focus. Away from the plane of focus, the D3X at F16 resolves much more than the D3 at F8.

Why would it be like that?

Quote
Also, I'm pretty sure that the D3X at F16 and at the plane of focus resolves significantly more than the D3 does at F16 and at the plane of focus.

Correct and that's what I posted in one of the examples using the 1Ds mkIII as an example.
Logged

Ray
Sr. Member
****
Offline Offline

Posts: 8884


« Reply #146 on: March 04, 2012, 05:36:01 PM »
ReplyReply

Why would it be like that?


Because F16 provides a much great DoF than F8 using cameras with the same size sensor. When comparing resolution advantages of same size sensors that have different pixel counts, it's only sensible to use the same lens at the same F stops. You will then find that the sensor with a substantially greater number of pixels will deliver more resolution at all F stops up to, and perhaps including, F22.

Resolution at F22 is the puzzling bit. I find it susprising that a Photozone test of the Canon 100/2.8 Macro at F22 shows the 50D producing significantly more resolution than the 350D with the same lens at F22. It's the sort of result one might expect if the 50D image had had 100% detail enhancement in ACR sharpening, which the 350D image lacked. In this sense, improvements in RAW converters can skew the results to some degree, but it's clear from tests from other sources, including my own, that we haven't yet reached the point where increased pixel count ceases to provide further resolution at all the commonly used F stops.

Logged
deejjjaaaa
Sr. Member
****
Offline Offline

Posts: 860



« Reply #147 on: March 04, 2012, 07:45:26 PM »
ReplyReply

Interesting to see that a Nikon 50mm lens at f/8 on D3 resolves exactly as f/16 on D3X. So basically doubling pixel count on the D3X is eliminated by stopping down from f/8 to f/16.

the point is that D3X will resolve more @ f/16 than D3 @ f/16...
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8884


« Reply #148 on: March 04, 2012, 10:16:07 PM »
ReplyReply

the point is that D3X will resolve more @ f/16 than D3 @ f/16...

And the point is also that we kinda know when using F16 to acheive a good DoF that we're sacrificing at least a little sharpness at the point of focus.

So it's rather nice to know that using the same size sensor with double the number of pixels allows us to get the same degree of sharpness at F16 as the older sensor provided at F8, plus the benefits of the significantly greater DoF that F16 provides.

Wow! It feels like I just upgraded all my lenses.  Grin
Logged
ErikKaffehr
Sr. Member
****
Online Online

Posts: 7424


WWW
« Reply #149 on: March 04, 2012, 10:30:42 PM »
ReplyReply

Hi,

Yes, the D3X will resolve marginally better at f/16 than the D3 at f/16, probably.

This is an old test I made using Imatest:

Another question how much of the sharpness lost at f/16 can be recovered using deconvolution. Worth looking into.

The diagram here and other tests ( http://echophoto.dnsalias.net/ekr/index.php/photoarticles/49-dof-in-digital-pictures?start=1 )I made led me to see f/16 to be the smallest aperture acceptable for normal shooting.

Best regards
Erik


the point is that D3X will resolve more @ f/16 than D3 @ f/16...
Logged

deejjjaaaa
Sr. Member
****
Offline Offline

Posts: 860



« Reply #150 on: March 04, 2012, 10:42:18 PM »
ReplyReply

Hi,

Yes, the D3X will resolve marginally better at f/16 than the D3 at f/16, probably.

This is an old test I made using Imatest:


10-15% in your picture is not "marginally"

PS: were the bars mislabeled ? higher bars should be 1.6x crop 18mp 7D vs lower bars 1.5x crop 10mp A100
« Last Edit: March 04, 2012, 10:48:54 PM by deejjjaaaa » Logged
ErikKaffehr
Sr. Member
****
Online Online

Posts: 7424


WWW
« Reply #151 on: March 04, 2012, 10:45:21 PM »
ReplyReply

Hi,

I'd say that Slobodan has a good point, as he often does. Regarding resolution, the more you have the more do you have to loose. I generally agree that we loose resolution when stopping down beyond optimal aperture. But some resolution can be regained with deconvolution sharpening. Technically I would say that advanced sharpening in CS5 with gaussian spread function and well chosen radius can recover amazing amount of detail.

Our vision is most sensitive to detail at 50% MTF (according to Norman Koren http://www.normankoren.com/Tutorials/MTF1A.html ) but resolution is limited by lower MTF (like around 20%).

The article here shows the effect of sharpening: http://echophoto.dnsalias.net/ekr/index.php/photoarticles/49-dof-in-digital-pictures?start=2

Best regards
Erik


Well, basically DxO has measurements by which one can assess the resolution by fstop and camera sensor megapixels. They show that doubling the megapixels are negated by stopping down from f/8 to f/22 caused by diffraction. They also show that at f/22 a doubling of megapixels do increase the resolution which is in contradiction to the article linked to. So it raises some questions:

1) What level of resolution increase will an increase from the 21MP of the 1Ds mkIII or the 22MP of the 5D mkIII up to the 36MP of the D800 give when stopped down to e.g. f/16.

2) What change in the resolution tables (table 3 in the linked to article on Lula) should be done to reflect reality rather than a seamingly over conservative calculation?

I have for some time relied (maybe naively as I haven't done any real world resolution tests to check it) on the accuracy of this table and I certainly have to revise my opinion on that. I have done test shots stopping down and seen the lost of resolution on the screen on real world subjects but that is hard to translate into excact loss of resolution in terms of megapixels or lp/mm. So only recently did I check the DxO measurements in the way shown in this thread.

Logged

ErikKaffehr
Sr. Member
****
Online Online

Posts: 7424


WWW
« Reply #152 on: March 04, 2012, 10:50:46 PM »
ReplyReply

Hi,

I would say that stopping down to f/16 made the 10MP A100 into a 5.6 MP camera while it made the 6 MP D7 into a 4.4 MP camera. Time to do more testing with more recent stuff.

Best regards
Erik


10-15% in your picture is not "marginally"
Logged

dreed
Sr. Member
****
Offline Offline

Posts: 1245


« Reply #153 on: March 04, 2012, 11:14:47 PM »
ReplyReply

When comparing f/8 with f/16 and f/22, remember that the resolution projected by the lens is not linear, meaning that the lens itself may resolve better or worse af f/8 compared to f/22 regardless of the diffraction problems with the sensor.

I don't believe that this issue is as easy as it just being based on the sensor pixel size and diffraction pattern.
Logged
ErikKaffehr
Sr. Member
****
Online Online

Posts: 7424


WWW
« Reply #154 on: March 04, 2012, 11:31:58 PM »
ReplyReply

Hi,

The way I see it there are advantages of increasing resolution, probably at least to 50 MP, but I also agree that the resolution advantage is easy to waste.

Here are the advantages of increasing resolution:

- Better detail at optimal aperture
- Less aliasing, so OLP filtering can be reduced or with the D800/E eliminated
- Less jaggies on thin lines

There are really only two "down sides" of increasing resolution:

- Increasing resolution will reduce DR somewhat, doubling resolution (from 25 to 50 MP) will reduce DR by 1/2 stop.
- File sizes increase

I don't think that the 36 MP the Nikon has compared to the 5DIII's 22MP does change the world, but it is certainly an advantage for the Nikon. Nikon probably also has a healthy lead in DR, due to their Sony Exmoor based sensor. Although little data has been seen for either system I'd say extrapolating from images and data regarding D7000 and D7 is a valid method.

In my view, the upgrade Canon made is a welcome improvement for Canon users, on the other hand it is first with the D800/D800E Nikon owners have access to a high resolution camera at reasonable price and superior DR at low ISO.

Sony Alpha 77 and NEX7 has put 25MP in APS-C, that correspond to 52 MP on full frame. Now, dream about a nice Alpa body with a Phase IQ248 with 148MP CMOS sensor from Sony. Wouldn't that be a nice dream? Except the price tag, of course, a real nightmare!

Best regards
Erik




I'm another who is of the opinion that more than 20-something megapixels in a 35mm full frame format is pretty much a waste. That's simply because of the small sensor format and the huge amount of magnification required to take it to large print dimensions. Lens diffraction, focus inaccuracy, and every other quality issue are magnified equally, so it's between difficult and impossible to get much more than 20-something MP of true resolution out of any 35mm format sensor. You certainly can get more pixels, but they don't necessarily translate into more information or better print quality. I'd think those who truly need more resolution than a good 20+ MP 35mm format camera can provide would be far better off spending their money on medium format. It costs considerably more, but an increased MP count in a larger format makes a difference that can actually be seen. Perhaps 30 MP would be nice, but it's on the far outside fringes of what's usable with the best prime lenses (certainly not any zoom), stopped down to their optimal apertures (say F5.6 or possibly F8), mounted on a camera that's fused to bedrock, or at least a very heavy tripod. Not many people actually shoot like that. Of course all this is just my opinion, and it along with $5 will buy a fine cup of coffee somewhere.
« Last Edit: March 04, 2012, 11:33:30 PM by ErikKaffehr » Logged

Ray
Sr. Member
****
Offline Offline

Posts: 8884


« Reply #155 on: March 05, 2012, 05:51:53 AM »
ReplyReply

I must admit I find it puzzling that anyone should hold the opinion that more than 20 odd megapixels on the full-frame 35mm format is a waste when both Canon and Nikon have already been producing cropped versions of 46mp and 36mp full-frame formats for a couple of years or more.

Are the 18mp of the Canon 7D and the 16mp of the Nikon D7000 also a waste?  Whatever the problems regarding focussing accuracy, lens limitations, adequate shutter speed, print quality etc that one imagines might afflict the 46mp and 36mp FF formats, the same problems should already be afflicting the 7D and D7000. But I haven't heard too many complaints.

Perhaps there's a failure here to understand the basic concept that a 46mp full-frame image cropped to 18mp actually would have the same quality and resolution as a 7D shot taken from the same position using the same lens at the same aperture and the same shutter speed (provided the crop is from the centre and has the same aspect ratio, and provided the pixel quality and processing of the 46mp is at least as good, and there's no reason to doubt that it would be, although there may be a reason for frame-rate to suffer.)

As I see it, for economic reasons manufacturers of DSLRs have always given us cropped formats first, and then the full expanded version much later, after sensor fabrication has becomes more economical with improving technology.

Basically, Nikon have now given us the expanded version of the D7000 in the form of the 36.3mp D800, but Canon has not progressed hardly at all beyond the expanded version of their 8mp 20D, as regards pixel density.

However, I don't wish to appear too negative. When the 5D3 is available for testing, we might get a few pleasant surprises, hopefully with regard to noise and DR performance.
Logged
stever
Sr. Member
****
Offline Offline

Posts: 1065


« Reply #156 on: March 06, 2012, 04:24:25 PM »
ReplyReply

no complaining, but i never found the small increase in resolution from the 20D to 40D or 40D to 7D to be much of an argument for up-grading.  in each case my decision was based on improvements such as better UI, improved autofocus, etc. making the cameras more useful.  for IQ, the 5D was better than any of the crop frame cameras, and the improvement in resolution and noise with the 5D2 made this upgrade a much easier decision than the crop frame cameras.  i would still prefer the 7D to have fewer pixels if it made a meaningful reduction in noise.

given the usefulness to me of lower noise combined with better weather seal, improved UI, frame rate, bracketing, etc. of the 5D3; ugrading is not going to be a terribly difficult decision - and i probably won't be able to wait until the price comes down, as it should.  as sexy as more resolution sounds, my experience with crop-frame cameras is that improbing the pixel density beyond that of the 20D will return small noticeable return in resolution and i'd still rather have lower noise
Logged
kimble
Newbie
*
Offline Offline

Posts: 1


WWW
« Reply #157 on: March 06, 2012, 06:51:59 PM »
ReplyReply

Yawn worthy indeed!
Logged
Ray
Sr. Member
****
Offline Offline

Posts: 8884


« Reply #158 on: March 06, 2012, 08:02:59 PM »
ReplyReply

no complaining, but i never found the small increase in resolution from the 20D to 40D or 40D to 7D to be much of an argument for up-grading. 

It's no susrprise you would find the increase in resolution from the 8mp 20D to the 10mp 40D of little significance, but an upgrade from the 10mp 40D to the 18mp 7D represents a worthwhile increase in resolution. The 7D also has about a 1/2 stop improvement in DR compared with the 40D, at equal print sizes.

The 40D represents a 26mp full-frame, and the 7D a 46mp full-frame.
Logged
MrSmith
Sr. Member
****
Offline Offline

Posts: 896



WWW
« Reply #159 on: March 07, 2012, 06:29:04 AM »
ReplyReply

i feel the need to photograph some brick walls or a sheet of newspaper.
i think i need to get out more. Undecided
Logged
Pages: « 1 ... 6 7 [8] 9 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad