Canon 5Ds and 5Ds R samples at flickr

[BernardLanguillier]

The point I am making is not whether or not the files from that camera could be improved upon, if the RAWs were given. I am fully aware that this is so.

What I am saying is that just because a pig is receptive of lipstick doesn't mean that on application, it turns into Jennifer Lopez. I have seen enough pigs and Ms. Lopez to reach this reasonable conclusion.

True, this being said the 5Ds with an Otus 85mm f1.4 in studio will most probably be objectively superior to many MF combos in terms of image quality.

That may not make it a superior studio camera of course.

But Canon has nothing coming close to the Otus and they aren't going to promote Zeiss glass obviously.

Cheers,
Bernard

[John Koerner]

What I am saying is that just because a pig is receptive of lipstick doesn't mean that on application, it turns into Jennifer Lopez. I have seen enough pigs and Ms. Lopez to reach this reasonable conclusion.

Just because you make an outlandish, over-exaggerated comparison (that has nothing to do with reality) doesn't mean there aren't many applications where the 5Ds will be a superior tool to a MF camera and rival it in image quality, surpassing it with certain applications.

Jack

PS: Some people would even argue that Jennifer Lopez is herself a glorified, beautified pig ... in fact, Mariah Carey prefers the company of pigs

[RobertJ]

No amount of fine technique can magically get amazing output from what is an oversized 7D2 sensor, which is a slightly updated 7D sensor.

Listen, I agree with you that these samples contain some of the worst image quality I've ever seen, but I know for sure these are JPEGs processed by the camera, and every "Sample" I've seen dating back to even before 2004 from Canon and Nikon, and even Phase One, have been really bad!

Wait for a RAW file shot with a Zeiss, processed in Capture One, then come back and give your opinion.

BTW, I don't own any Canon gear at the moment.  Looking at DP Merrill files, then looking at these samples makes me laugh.  But just wait.  It will be a pretty damn good camera, IMO, along with the Nikon and Sony versions, no doubt.

[synn]

I have seen D800 and D810 files shot with Otii, processed in C1P and whatnot (I use Nikon's best glass on my personal D800 and process in C1P, BTW) and I still don't consider them anywhere near what I can achieve with my MF kit. This canon sensor has a pixel count advantage over the Sony in the Nikons and pretty much nothing else, so I am not expecting it to surprise me.

There are applications where I prefer the Nikon over the Credo, but that's besides the point. I didn't get the Credo to be my doitall camera and I have no illusions that it is one. In terms of absolute IQ for tripod/ studio based low ISO applications (Which are the use case scenarios for all the samples in that Flickr set), the 35mm cameras are not on par with MF. This is a working artist's opinion, not that of a professional test chart analyst. Take it for what it is worth.

If one is used to the 35mm scene and wants a step up in IQ, I am sure this thing or a D800/810 will make them more than happy. But please, don't sing that song to people who have actually used MF gear for artistic purposes. There are many scenarios where we choose something else over the MF gear to do our work, but "Superior IQ" isn't one of them.

At least, not yet.

 

[Theodoros]

True, this being said the 5Ds with an Otus 85mm f1.4 in studio will most probably be objectively superior to many MF combos in terms of image quality.


Cheers,
Bernard

How did you come to that conclusion? Can you please define "better"?

[ErikKaffehr]

Hi,

There are two things that set the Otus apart from almost other lenses I know about:

• It is almost totally free axial color aberrations at full aperture, no green/magenta fringing in OOF areas.
• It is sharp enough to produce moiré at f/1.4 in the corners on a 36 MP sensor

I have seen enough examples to feel that the above are proven facts. From that I would assume that the lens would produce pretty much perfect images with a 50 MP or even 100 MP sensor.

How Otus would stand up against a 50 MP sensor with an excellent lens is something we don't know. Personally, I would expect the Canon 5Ds to do very well if used by a competent photographer.

On the other hand, I would still think that IQ-280 with a technical camera and Rodenstock HR lenses at optimal aperture may play in a slightly different division.

Best regards
Erik

How did you come to that conclusion? Can you please define "better"?

[synn]

I'll give you a simple test that anyone can do without knowing what a nyquist or a Bayer or whatever is.

Shoot your 35mm high res camera with whatever lens you want and process in the raw converter of your choice. Export as 16 bit TIFF to photoshop. Now do the same with an MF camera of similar MP with a competent lens (say, the Schneider Kreuznach P1 lenses). Blow both images up to 200% (without using fancy plugins. We are not evaluating the efficiency of a scaling algorithm). You'll see that the MF image holds up MUCH better with only a hint of quality loss. The 35mm file completely breaks apart in comparison.

Whatever the scientific reason behind it is, MF files have a certain fatness to them that the 35mm cameras are yet to match. This is something MP counts or DxO scores won't tell you.

[JoeKitchen]

I'll give you a simple test that anyone can do without knowing what a nyquist or a Bayer or whatever is.

Shoot your 35mm high res camera with whatever lens you want and process in the raw converter of your choice. Export as 16 bit TIFF to photoshop. Now do the same with an MF camera of similar MP with a competent lens (say, the Schneider Kreuznach P1 lenses). Blow both images up to 200% (without using fancy plugins. We are not evaluating the efficiency of a scaling algorithm). You'll see that the MF image holds up MUCH better with only a hint of quality loss. The 35mm file completely breaks apart in comparison.

Whatever the scientific reason behind it is, MF files have a certain fatness to them that the 35mm cameras are yet to match. This is something MP counts or DxO scores won't tell you.

I am trying to stay out of this as much as possible, but I have to agree with Synn here.  I often examine my files at 200% on location to find the critical focus point.  At 100%, the files are too "nice" looking to find it, at 200% they are still very nice, but it is easier to find where the camera is exactly focused at. 

Any DSLR files do not hold up at 200% and look horrible. 

[Theodoros]

Hi,

There are two things that set the Otus apart from almost other lenses I know about:

• It is almost totally free axial color aberrations at full aperture, no green/magenta fringing in OOF areas.
• It is sharp enough to produce moiré at f/1.4 in the corners on a 36 MP sensor

I have seen enough examples to feel that the above are proven facts. From that I would assume that the lens would produce pretty much perfect images with a 50 MP or even 100 MP sensor.

How Otus would stand up against a 50 MP sensor with an excellent lens is something we don't know. Personally, I would expect the Canon 5Ds to do very well if used by a competent photographer.

On the other hand, I would still think that IQ-280 with a technical camera and Rodenstock HR lenses at optimal aperture may play in a slightly different division.

Best regards
Erik

Yeah, I know Otus is a great lens, so are some MF lenses too... My question is (was) different though... A good lens certainly improves things... but this is with any sensor! And there are many very good lenses out there for MF cameras.

[Theodoros]

I'll give you a simple test that anyone can do without knowing what a nyquist or a Bayer or whatever is.

Shoot your 35mm high res camera with whatever lens you want and process in the raw converter of your choice. Export as 16 bit TIFF to photoshop. Now do the same with an MF camera of similar MP with a competent lens (say, the Schneider Kreuznach P1 lenses). Blow both images up to 200% (without using fancy plugins. We are not evaluating the efficiency of a scaling algorithm). You'll see that the MF image holds up MUCH better with only a hint of quality loss. The 35mm file completely breaks apart in comparison.

Whatever the scientific reason behind it is, MF files have a certain fatness to them that the 35mm cameras are yet to match. This is something MP counts or DxO scores won't tell you.

Agree... Nikon's 16mp sensor is about the only one that comes close to an MF sensor when blown to 200%.... Things are even more impressive if one prints an image at ...36ppi!!! ...with D800/E/810 ...just forget it!

[Theodoros]

I now use two MFDBs, one of 9μm pixel size and another of 6.8μm size... The second one has some clear advantages (at single shot mode) when it comes down to moire, but that's about the only thing I would criticise if compared to the "fat" pixel other back... On anything else the image of the (same image area) lower pixel count sensor is in the lead. No surprise, the same happens with my Nikon 16mp sensor when compared to any other FF sensor I've ever tried... it is simply in a class of its own for FF performance.

EDIT: Especially now that I use my (7 + TC) Zeiss lenses of the Contax645 on my Nikons via the JAS adapter (which is AF and aperture compatible)... the 16mp sensor behaves more close to the performance of an MFDB than I could ever imagine an FF sensor would.

[ErikKaffehr]

Hi,

I just made that test of yours.

• Left one is Distagon 40/4 FLE on Hassellblad 555/ELD with P45+ back, processed in C1v8
• Center is Sony Alpha 99 (24MP) 24-70/2.8 at f/8 and 30 mm, processed in LR5
• Right One Distagon 40/4 FLE on Hassellblad 555/ELD with P45+ back, processed in LR5

.

Same day, same time, camera repositioned for the DSLR shoot about two meter to the right for better composition. All enlarged 200% using bicubic in Photoshop CS.

I would love to do this comparison with sensors of similar resolution, but I used what I have and use on subjects I normally shoot.

Raw files are here:

P45+ 39MP: http://echophoto.dnsalias.net/ekr/Articles/Shoots/Aliasing/20140708-CF045336.iiq
Alpha 99 (24 MP): http://echophoto.dnsalias.net/ekr/Articles/Shoots/Aliasing/20140708-_DSC4865.dng

Best regards
Erik

I'll give you a simple test that anyone can do without knowing what a nyquist or a Bayer or whatever is.

Shoot your 35mm high res camera with whatever lens you want and process in the raw converter of your choice. Export as 16 bit TIFF to photoshop. Now do the same with an MF camera of similar MP with a competent lens (say, the Schneider Kreuznach P1 lenses). Blow both images up to 200% (without using fancy plugins. We are not evaluating the efficiency of a scaling algorithm). You'll see that the MF image holds up MUCH better with only a hint of quality loss. The 35mm file completely breaks apart in comparison.

Whatever the scientific reason behind it is, MF files have a certain fatness to them that the 35mm cameras are yet to match. This is something MP counts or DxO scores won't tell you.

[Dshelly]

I shoot in studio every week using primarily a 5D3 and I am very happy with this camera. To date, it is the best dslr I've owned. Occasionally, a client will request a MFD camera, but not often. For what I do, which is advertising campaigns for the film studios and tv networks – the 5D3 is more than adequate for the 27" x 40" printed posters. Often times the artwork is repurposed for billboards, bus sides, and sometimes building sides and other collateral materials. My camera stills holds up on billboards because the viewing distance is typically over 60' and beyond, so whatever artifacts and blemishes are on the image are rarely perceived by the naked eye.

Based on what I'm seeing in theses sample photos, I feel that the 5Ds may provide a satisfying bump in resolution to make the posters that much better, in terms of detail - though, it remains to be seen how good this camera will perform. I don't expect it to provide the heightened details you'll find in medium format images, in terms of getting a full body shot of a person and being able to zoom in and find exceptional detail in the eye lashes, eye brows and the iris.

[torger]

I am trying to stay out of this as much as possible, but I have to agree with Synn here.  I often examine my files at 200% on location to find the critical focus point.  At 100%, the files are too "nice" looking to find it, at 200% they are still very nice, but it is easier to find where the camera is exactly focused at. 

Any DSLR files do not hold up at 200% and look horrible. 

That's interesting, I find it to be quite the opposite and for a simple reason -- MFD images due to lack of OLPF are, depending on subject and sharpness, often riddled with aliasing and I don't like that to become visible. The more you enlarge the more visible it becomes. However the idea of MF is to have so many megapixels that you don't need to enlarge very much of course and then aliasing is usually less visible unless you havemoire. Anyway that's also why I would get the 5DS rather than the 5DS R, to get better image quality.

Photographers tend to think sharper pixels = better, that's why the R version is there in the first place due to popular demand, but from a real image quality aspect it makes no sense, especially if you must scale up. When sensors was really noisy and sharpening software was very limited it might have been an advantage with no OLPF under some conditions, but that time has passed.

I do agree though that without OLPF it's easier to find the exact focus point, which might be an advantage in live view focus peaking too, but that's the only real advantage I see. However if the manufacturer actually implements USM sharpening in the live view that advantage is nullified.

[torger]

Hi,

There are two things that set the Otus apart from almost other lenses I know about:

• It is almost totally free axial color aberrations at full aperture, no green/magenta fringing in OOF areas.
• It is sharp enough to produce moiré at f/1.4 in the corners on a 36 MP sensor

I assume it actually has round aperture too? I don't like edgy out of focus aperture rings, and I can't understand why high end MF lenses keep those cheap-looking aperture shapes. I can understand it on my Copal shutters as it's really old design, and it doesn't matter much as you rarely have out of focus areas, but on my Hassy HC lens... come on!

The Zeiss Otus is manual focus though which to me makes it a dud for portraiture so I don't find that lens that interesting anyway from practical reasons. What use is that great resolving power at f/1.4 if I won't be able to place focus anyway? It's useful in landscape, but then you stop down and at least f/8 there are many competitors up to the task.

Anyway, I don't see any reason for fanboyism, some things are better on MFD systems and some things are better with DSLR systems. For me it's very simple, I have seven evenly spaced primes corner-to-corner sharp flexible movements on them all (thanks to view camera design), and actually lighter weight than a DSLR system would be, and f/16 is a decent and for my style useful workaround for the aliasing issue.

[ErikKaffehr]

Hi Anders,

You are right. I assume the Otuses have circular apertures. I own no Otus, just seen test images.

All my Hasselblad lenses have some colour fringing (axial chromatic aberration) on OOF areas, and I have seen it almost all lenses. Even some Leica-S samples I have seen have a lot of it. The "rheingold1" image in this sample is a very obvious example of that: http://s-newsletter.leica-camera.com/s2-raw/

I am quite interested in the Otus 85/1.4. Once I have a camera I can put it on I might buy it, waiting for an upper end Sony (say A9) with 50MP+ and electronic first shutter curtain. The reason I want a wide aperture lens is that I have some interest in short DoF shooting, now that I can focus accurately using live view, but I don't want a lot of green/magenta fringing on the out of focus areas.

Just as an explanation. My APS-C camera is a Sony Alpha 77. That camera has 3.77 micron pixel pitch. At that pixel size most of my lenses are bit soft, but the Alpha 77 has widely better resolution (lp/mm) than either the Alpha 99 or the P45+, also the Alpha 77 has little aliasing at f/8. So from that that I believe that full frame with 3.5-4 micron pitch would be pretty decent.

I also don't think that 3.8 micron pixels outresolve most lenses. The Planar 80/2.8 I have can resolve 150-180 lp/mm at f/8 on Adox 120CMS film in the sweet spot. I have not tested my other lenses on film, but MTF measurements I made indicate that they perform at about the same level.

The cameras I use are Hasselblad 555/ELD with 40/4, 50/4, 80/2.8, 100/3.5, 120/4 and 180/4 lenses and a P45+ back. All lenses outresolve that back, as indicated by aliasing. To this comes the Sony Alpha 99 and the Sony Alpha 77 which I mostly use with a couple Sony's better zooms.

Best regards
Regard

I assume it actually has round aperture too? I don't like edgy out of focus aperture rings, and I can't understand why high end MF lenses keep those cheap-looking aperture shapes. I can understand it on my Copal shutters as it's really old design, and it doesn't matter much as you rarely have out of focus areas, but on my Hassy HC lens... come on!

[torger]

I am quite interested in the Otus 85/1.4. Once I have a camera I can put it on I might buy it, waiting for an upper end Sony (say A9) with 50MP+ and electronic first shutter curtain.

As I pointed out in my edit above I think a huuuuge disadvantage of the Otus is that it's manual focus. To me it's a show stopper. Sure you can focus on live view and shoot great landscapes but then I would use f/8-f/11 on a full-frame DSLR, and I'm sure lenses like the Sigma Art is up for the task then?

When it comes to shooting wide open portraiture I don't like manual focus. There's the focus indicator in the viewfinder so it's not as difficult as on a legacy 100% manual systems, but I'm not really fast or precise enough getting there. I guess it can come with training though, can be a fun challenge to become good at, similar to that I've enjoyed to become a master at ground glass composition and focusing with my Linhof, there I use a 20x loupe and several seconds to get to the "focus point" though.

I would not suggest a Zeiss Otus to someone that wants to see how capable a DSLR system is especially a Canon system, taking away the autofocus from a recent higher end Canon is taking away a lot of it's really strong points. The light sensitivity of the autofocus system is just incredible.

I think there's still limitations in the lens lineups for high res DSLR, there are gaps here and there. For Canon the weak TS-E 45 is one such gap. In 4-5 years it will be better, today I still see high res DSLR as an "early adopter" thing.

[ErikKaffehr]

Hi Anders,

I have some interest shooting subjects with thin DoF like say strains of grass with very soft background. I feel that a short tele is best for what I am interested in shooting, so the Sigma 50/1.4 Art is no option, a Sigma 85/1.4 Art would be quite interesting. I found out that the Sony 85/1.4 ZA is quite OK, and I may buy that lens. It is very different from the Zeiss Planar T* 1.4/85. Would we have an 85/2.0 with true apochromatic performance, I would prefer that, I think.

Best regards
Erik

As I pointed out in my edit above I think a huuuuge disadvantage of the Otus is that it's manual focus. To me it's a show stopper. Sure you can focus on live view and shoot great landscapes but then I would use f/8-f/11 on a full-frame DSLR, and I'm sure lenses like the Sigma Art is up for the task then?

When it comes to shooting wide open portraiture I don't like manual focus. There's the focus indicator in the viewfinder so it's not as difficult as on a legacy 100% manual systems, but I'm not really fast or precise enough getting there. I guess it can come with training though, can be a fun challenge to become good at, similar to that I've enjoyed to become a master at ground glass composition and focusing with my Linhof, there I use a 20x loupe and several seconds to get to the "focus point" though.

I would not suggest a Zeiss Otus to someone that wants to see how capable a DSLR system is especially a Canon system, taking away the autofocus from a recent higher end Canon is taking away a lot of it's really strong points. The light sensitivity of the autofocus system is just incredible.

I think there's still limitations in the lens lineups for high res DSLR, there are gaps here and there. For Canon the weak TS-E 45 is one such gap. In 4-5 years it will be better, today I still see high res DSLR as an "early adopter" thing.

[chrismuc]

Hi Erik,
I suggest to wait for a Sigma 85f1.4 Art ... or go for the sharp and extraordinary well corrected ZE 135f2 Apo-Sonnar (which also has a long focus tube up to 1:4).
:-)
Christoph

[Ken R]

As I pointed out in my edit above I think a huuuuge disadvantage of the Otus is that it's manual focus. To me it's a show stopper. Sure you can focus on live view and shoot great landscapes but then I would use f/8-f/11 on a full-frame DSLR, and I'm sure lenses like the Sigma Art is up for the task then?

When it comes to shooting wide open portraiture I don't like manual focus. There's the focus indicator in the viewfinder so it's not as difficult as on a legacy 100% manual systems, but I'm not really fast or precise enough getting there. I guess it can come with training though, can be a fun challenge to become good at, similar to that I've enjoyed to become a master at ground glass composition and focusing with my Linhof, there I use a 20x loupe and several seconds to get to the "focus point" though.

I would not suggest a Zeiss Otus to someone that wants to see how capable a DSLR system is especially a Canon system, taking away the autofocus from a recent higher end Canon is taking away a lot of it's really strong points. The light sensitivity of the autofocus system is just incredible.

I think there's still limitations in the lens lineups for high res DSLR, there are gaps here and there. For Canon the weak TS-E 45 is one such gap. In 4-5 years it will be better, today I still see high res DSLR as an "early adopter" thing.

I agree. I wouldnt choose a portrait lens for its ultimate sharpness capability but one that I would like to work with and has the look I want. Unless I am using a camera optimized to be used in Manual Focus I would much rather use a lens that has really nice AF. Of course if you want to shoot test charts and/or landscapes/still life/products on a tripod then Manual Focus is fine. But for working with people AF is key, at least for me. That said, the newer cameras with great EVFs and built in manual focusing assists in the EVF are making manual focus a viable option again. The Fuji X-T1 and the Sony A7 series qualify. All modern 35mm DSLRs do not. The Leica S (with microprism screen) is the only DSLR I have used that does.

I can't wait for the 5DS/R to be thoroughly tested. I am not expecting much improvement in Dynamic Range (some, but not a lot, as has been suggested) but a bit of Improvement in Color. Putting more pixels on a scene (with a bayer sensor) means significantly better color sampling. Any resolution improvements are welcome of course but I am sure they will be more lens dependent, like with any camera.

Anyway, Im overall curious on the workability and look of the raw files. This is something that I liked about the D810. To me it the raw files are a nice improvement over the D800/D810 both in Color and even some DR. They are nicer no question.

Of course no DSLR will replace my IQ160 back. The Phase is just great to work with in a huge range of conditions and situations (with both the Arca and the H1) and the files are unreal and most importantly WAY different than any CMOS camera. I like something different.