The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Canon 5Ds/5DsR

tjv

Active member
DP Review:

"Canon tells us that at a pixel level, noise levels should be very similar to the EOS 7D Mark II and slightly better than what we'd expect from a 5D Mark II (note: not a 5D Mark III)."

and

"As far as dynamic range is concerned, we're told that the new 5DS and 5DS R should give the same performance as the current EOS 5D Mark III."

Seems like it's pretty much an upscaled 7DII sensor, right? Personally, I call that disappointing.

We use 5D III's at work, and I generally like them but hate the pattern noise, which I can see at base ISO without any big edits. Regardless, like Nikon – which I own and use – I wonder what AF lenses are seriously up to the task of 50MPX, particularly at the wide end.

This seems like an advancement of sorts, but certainly not revolutionary like the D800 was. For Canon shooters who need more resolution and are happy with the IQ of their MKII or III, then I guess this is the ticket. For people buying into full frame from scratch and wanting "the best", I guess it's a harder sell. In my opinion, one would be better to choose what they feel more comfortable holding and using than buying on spec anyway. In my case, I'd prefer the D810, but others are different and difference makes the world go around!
 

tjv

Active member
Really,

Doubling the MP's is not raising the bar. Phase One and HB are shaking in their boots!

I guess you need 100MP and DR of 20EV to make a statement
Lets wait until the real world samples come out, eh?
I'm no fanboy. but I'll put money on this Canon being no competition against the H-50c / IQ250/150 / Credo 50 / 645Z.

Anyway, :worthless:
 

torger

Active member
Yes the DR thing was a bit disappointing, although not that surprising, we've seen this over and over again for a few years now. For studio work with controlled light I don't think it will be an issue though, but as said above it isn't a "now I'm switching back to Canon!" type of camera.

With no ability to compete on the DR I'm a bit surprised that they waited this long with releasing a high MP camera.

I don't think we will see any noise pattern issue, just checked out the 7DII and it has no such issues, which means that the DR that is there can be used to its full potential. Just need that raw file...
 
Last edited:

gazwas

Active member
Seems like the makings of a great camera but considering all but the resolution has stood still (actually slightly less tested DR than my 1Ds3) and Canon has decided to reduce its video capability while all other manufacturers are boosting their cameras (Olympus OM-D E-M5 II) its not a camera I'll be rushing to pre-order. With an overly long predicted June delivery date estimate, this gives Sony 5 months to come out with their version so I'll sit this one out and see what develops. The £3200 price point was a pleasant surprise though as I thought it would be more.
 

torger

Active member
Wow those sample images do look very impressive, especially the 5DsR ones. I do however see some weird colours in the specular highlights along the body of the Hippo in the water. Is that just a lens thing?
No, that's aliasing. False colors, a result from the demosaicing. AA filter has always been a good idea to have, but somehow it has in the photography world been distorted to be a "feature" rather than the bug it is.

Moire will be rare, but false colors "christmas tree colors" is there as soon as you have a sharp lens and sharp detailed transitions between dark and bright.

I see this all the time with my medium format camera, I can live with it though especially since I shoot f/16 often (where diffraction evens it out a bit), but if there would be a choice like it is in this case I would go for the AA-filtered version.

1-2% in extra resolving power is just not worth the false color issues.

Aliasing is also not going to make lens corrections or image rotations perform better. No problem on my medium format camera as I have distortion free tech lenses lenses and always shoot from a tripod, ie I don't do any image transformations, but with a DSLR with retrofocus lenses and hand-held shooting I would think about taking image transformation performance into account, and then again an AA-filtered input is better.
 

gazwas

Active member
I see this all the time with my medium format camera, I can live with it though especially since I shoot f/16 often (where diffraction evens it out a bit), but if there would be a choice like it is in this case I would go for the AA-filtered version.
Yes, shooting architecture I never liked to sharpen my Phase files too much as I always thought they suffered quite badly from aliasing.

As a 50Mpix H4D shooter what do you thing of those files - they look very impressive to me and no issues with moire or aliasing in the 5DsR skyline shot. At first glance and considering they are only early sample files they compare very similar to how my Phase files looked. With prime lenses and better processing this looks to improve further I'd imagine.
 

torger

Active member
Yes, shooting architecture I never liked to sharpen my Phase files too much as I always thought they suffered quite badly from aliasing.

As a 50Mpix H4D shooter what do you thing of those files - they look very impressive to me and no issues with moire or aliasing in the 5DsR skyline shot. At first glance and considering they are only early sample files they compare very similar to how my Phase files looked. With prime lenses and better processing this looks to improve further I'd imagine.
I think it looks promising, and Canon has quite many sharp lenses so I think it will be fine. I have only looked at them briefly, and in-camera jpegs tend to render a good bit softer than if you have a raw workflow. Canon (and Nikon and I guess any other Japanese camera maker) has never really worked hard to make the sample shots overly impressive, it's more like random test shots.

When it comes to color aliasing it depends also on the demosaicer a lot, the demosaicers that are best at extracting detail from AA-filtered sensors may make false colors more visible on AA-less sensors, maybe Canon has some work to do there. Personally I would still pick the 5Ds over the R, even more so now when I've started to work with prints, where micro contrast is of less importance and aliasing is more worrying.

I have the TS-E 24 II, but for Canon to be a real alternative to my tech system they need to upgrade the 45 and 90 too. At some point I'd like to do a side-by-side test with the TS-E 24 II + Canon 5DS and my H4D-50 + SK35. I think results will be similar, with a little sharpness edge to the H4D-50, and less distortion of course. DR will probably be about the same.

It shall be interesting to see if there is any truth in the rumours that they have focused the CFAs on better color separation. I don't have the best eyes to see such a thing, but others have. If color is great and they have the noise just as well-behaved as on the 7DII (which is better than the 5Dmk3!), then I think it will be a great camera.

High MP DLSRs is not super-cheap from a lens perspective though, so just-below-top-notch MFD (like I have) or the Pentax 645z are quite real alternatives. The flexibility of a DSLR system and lens lineup is undeniable though of course.
 

gazwas

Active member
Personally I would still pick the 5Ds over the R, even more so now when I've started to work with prints, where micro contrast is of less importance and aliasing is more worrying.
Very interesting as I thought all MFD shooters would jump at the R version as many seem to love and have switched to the AA filter free Nikon D810 or Sony A7r which in theory should suffer more from aliasing having less pixels.

I shoot architecture and product photography so it seems I may need both versions, R for product and regular S for architecture?
 

torger

Active member
Very interesting as I thought all MFD shooters would jump at the R version as many seem to love and have switched to the AA filter free Nikon D810 or Sony A7r which in theory should suffer more from aliasing having less pixels.

I shoot architecture and product photography so it seems I may need both versions, R for product and regular S for architecture?
Well, I'm a software engineer professionally and serious enthusiast landscape photographer. I've done quite some contributions to raw software, and when you get to see things on the "signal level" the taste gets a bit different I guess. What is detail worth if it's false? You can get a similar crisp look by adding a little noise. I've spent quite some time to look at different demosaicers and false color reduction etc, and my taste has then been turned towards shoot with more diffraction to smoothen things up, and AA filter yes please if it's available.

I can't deny that there is a certain type of satisfaction of zooming to 100% and see that special crispiness, but really, that's artificial image quality. An image that instead is a little bit smoother is very robust to scaling and transformation (printing in huge sizes, lens corrections, rotations), and with proper sharpening workflow the detail loss is say 1-2% (Bart at Lula can explain in detail). AA filters in cameras are not particularly strong, just enough to remove the worst artifacts.

One more thing is that for a demosaicer to hide color aliasing issues it seems like it must lower it's color resolution in some way. This haven't been that much investigated but I would not be surprised if you actually could get slightly better color resolution from the AA filtered sensor.

I don't believe that there is any global advantage of not having an AA filter, that is if you zoom out or look at a distance so you don't see the pixel peep level then you won't see a difference. This is one of those religion type of things though. Before the D800 when there was a bigger gap between MFD and 135 the "no AA filter = feature" view grew strong, but I think it has more to do with that it was a differentiator than it actually provides a real advantage (except for pixel crisp).

So I think you should only need one, the 5Ds... but yes it will be against the mainstream taste. There's nothing wrong in liking that pixel peep crispness and accept that the cost is false colors and details and occassional moire, but I think if someone made a well-presented case with A/B comparisons of the false color issues popping out here and there (also in "organic" subjects) I think more would be skeptical about the R.

There have been some D800 vs D800E comparisons, but I don't think anyone with a deeper knowledge of aliasing artifacts, so they only looked for moire patterns which are rare (moving water can be a pain though, but fortunately shutter speeds on the tech cam usually smoothens out the water.. hehe). If I get the chance sometime maybe I'll do a 5Ds vs 5Ds R comparison.

It should be said though that microlenses help (since more of the pixel area is sampled), so the 5Ds R will have less moire and false color issues than my H4D-50, I just don't know how large the difference will be.

I think it will probably be like this; there will be more aliasing with the R and you can show it, also in organic subjects, but it will be at a level that some say "it's not too bad, I'll accept that", but on the other hand the sharpness difference will be very small too after proper post-processing. So it becomes very much a matter of taste, what would you rather get, drop 1-2% in resolution but get more accurate details, or the other way around? Since pixel peep sharpness stands very strong I think the R will win in most cases.
 

torger

Active member
Concerning the DR, the most likely thing seems to be that it's like the 7DII on the pixel level. It means a ~12 stop camera without any banding or other noise pattern issues. That's not bad in an absolute sense, I can work with that in landscape no problem.

But it does means that Canon is sort of last place in the DR race, and that's unfortunate, I'd like to see tighter competition now when Sony has a monopoly on high MP sensors with high DR. The lesser DR also means less pressure on the MF companies, which I also think is unfortunate as I think they need even more pressure to come up with some better business model that end customers would benefit from...
 

turtle

New member
The lack of improvement in DR is my only beef, although if files are free of nasty banding when shadows are pulled that's at least something. Looks like a great camera, but regarding DR I can only sigh in exasperation at Canon's total lack of meaningful progress.

For those implying that extra DR is only being requested by people who cannot expose properly, or who are pushing their files too far, I guess you don't do many 'on the ground' cityscapes at night (hello bright street lights and deep shadows in unlit areas) or do much contre jour shooting where flash is not remotely feasible. Sure, I can live with Canon DR 98% of my shooting, but when shooting London at night the lack of DR on my 5D III started to drive me nuts while my friend sailed along with his D600. I ended up having to use Nik Dfine to control banding in a fair percentage of my shots. You cannot always do exposure blending with moving (lit) water and other moving elements in the scene. If I can get 14 stops out of my£400 APS-C compact I feel I have every right to feel peeved that Canon falls two stops short of that with a full-frame camera launching at £3000. Its not like we need the additional DR all the time, its just that when you need it, you need it. Besides, we are only asking that Canon give us what pretty well every camera under the sun now has: more than 12 stops. Heck, the original RX100 compact with a sensor the size of an ant has more DR! So.. please stop implying we don't know how to use a camera or post process. Its just that your needs are not ours.
 

gazwas

Active member
Before all us Canon shooters get too down in the dumps and go jump off a bridge over the DR and how rubbish the 5Ds will be, I've just had a look at what else was announced today and I feel its worth mentioning the new EF 11-24mm F4L. I already own the 17mm TS-E and while not perfect I have always considered it Canon's crowning glory and a fantastic piece of glass but this new 11-24mm just looks jaw dropping!! :bugeyes:
 
V

Vivek

Guest
Before all us Canon shooters get too down in the dumps and go jump off a bridge over the DR and how rubbish the 5Ds will be, I've just had a look at what else was announced today and I feel its worth mentioning the new EF 11-24mm F4L. I already own the 17mm TS-E and while not perfect I have always considered it Canon's crowning glory and a fantastic piece of glass but this new 11-24mm just looks jaw dropping!! :bugeyes:
A lot of Sony users would be interested in the lens despite it looking HUGE even on the 5Ds body.

I think the rear filter slot is a very nice touch. :)
 

turtle

New member
No doubt this monster lens is being launched because they know they have to produce a lens for landscapers that can have a fair stab at 50MP. The lens looks incredible and judging by the performance of recent wide angle L zooms, I expect performance to be superb.

I certainly don't think the 5Ds will be any more rubbish than the 5D III I own, but I did dare to hope we would have a bit more of, you know, that stuff which cannot be mentioned!
 

Uaiomex

Member
Personally, I see no use for a high megapixel camera if you don't print large. When I buy a HMP camera this year, I will make sure it is the best option under my budget possibilities.
A good portrait studio would be better set with a 645Z for the tonality. A landscape photographer too but under a budget, FF is a must. Nikon, Canon and Sony (soon Pentax) have second best choices.
Which is best for me? As I see it today and being invested in Canon L glass, the Sony still looks the way to go. I will keep my 6D for everything else that demands speed or for low-light conditions. In short, for all photography without a tripod.

I'm sure at Metabones they're going now for the the fourth case of Veuve Clicquot. Cheers!
 

kdphotography

Well-known member
It's a step up for Canon, but far from revolutionary. I was hoping for a home run. (A little too late to the party, Canon)

I'm still hoping that the "s" in the 5Ds model is a sign that a 1D series body with commensurate improvements is in the making. But somehow why is it I feel like Canon and Nikon have switched places (from about ten years ago) where Nikon had nothing with its feeble attempts to "sell" pros that DX sized sensors were all you needed while they cleaned up in the P&S market making money and only Canon had a FF DSLR?....

I have all this Canon L glass and I'm not about to jump ship. So I wait. I don't care what Sony/Nikon are doing. Maybe I'll pick up the 5Dsr, maybe not. Maybe I'll just wait until Guy buys one and sells it after a few weeks....:ROTFL:

Fingers-crossed for a really clean 6400---not that sketchy thin 3200 file that I get with the 1DsIII.

my .02. So I wait.....
 
Last edited:

Uaiomex

Member
Torger, I love the way you explain things. Certainly this is not a black or white universe and of course photography neither.
Regards and thanks
Eduardo

Well, I'm a software engineer professionally and serious enthusiast landscape photographer. I've done quite some contributions to raw software, and when you get to see things on the "signal level" the taste gets a bit different I guess. What is detail worth if it's false? You can get a similar crisp look by adding a little noise. I've spent quite some time to look at different demosaicers and false color reduction etc, and my taste has then been turned towards shoot with more diffraction to smoothen things up, and AA filter yes please if it's available.

I can't deny that there is a certain type of satisfaction of zooming to 100% and see that special crispiness, but really, that's artificial image quality. An image that instead is a little bit smoother is very robust to scaling and transformation (printing in huge sizes, lens corrections, rotations), and with proper sharpening workflow the detail loss is say 1-2% (Bart at Lula can explain in detail). AA filters in cameras are not particularly strong, just enough to remove the worst artifacts.

One more thing is that for a demosaicer to hide color aliasing issues it seems like it must lower it's color resolution in some way. This haven't been that much investigated but I would not be surprised if you actually could get slightly better color resolution from the AA filtered sensor.

I don't believe that there is any global advantage of not having an AA filter, that is if you zoom out or look at a distance so you don't see the pixel peep level then you won't see a difference. This is one of those religion type of things though. Before the D800 when there was a bigger gap between MFD and 135 the "no AA filter = feature" view grew strong, but I think it has more to do with that it was a differentiator than it actually provides a real advantage (except for pixel crisp).

So I think you should only need one, the 5Ds... but yes it will be against the mainstream taste. There's nothing wrong in liking that pixel peep crispness and accept that the cost is false colors and details and occassional moire, but I think if someone made a well-presented case with A/B comparisons of the false color issues popping out here and there (also in "organic" subjects) I think more would be skeptical about the R.

There have been some D800 vs D800E comparisons, but I don't think anyone with a deeper knowledge of aliasing artifacts, so they only looked for moire patterns which are rare (moving water can be a pain though, but fortunately shutter speeds on the tech cam usually smoothens out the water.. hehe). If I get the chance sometime maybe I'll do a 5Ds vs 5Ds R comparison.

It should be said though that microlenses help (since more of the pixel area is sampled), so the 5Ds R will have less moire and false color issues than my H4D-50, I just don't know how large the difference will be.

I think it will probably be like this; there will be more aliasing with the R and you can show it, also in organic subjects, but it will be at a level that some say "it's not too bad, I'll accept that", but on the other hand the sharpness difference will be very small too after proper post-processing. So it becomes very much a matter of taste, what would you rather get, drop 1-2% in resolution but get more accurate details, or the other way around? Since pixel peep sharpness stands very strong I think the R will win in most cases.
 
Top