[JJ_SO]

Due to some of those irrelevant nonetheless funny battles about Sigma lenses and cameras I had the idea of going out, find some details and take comparable images to see if there's any visible difference between those slightly outdated Sigma Merill DP cameras and up-top-date D810 as well as outdated D5100.

 

Lenses were all Sigma - I just have no more Nikon 50 mm but I could have used a 70-200/4 and set it  to 75 mm to get comparable images. I think, it wouldn't have made a huge difference. If somebody feels, "I bet it would!", please feel free to add your comparison into this thread.

 

So, the D5100 had a 50/1.4 Art in front of it. The D810 a 24-105/4 Art as well, which was set to 75 mm. Aperture was f/5 for the APS C cameras and f/8 for the FF model (5 × 1.5 = 7.5, so I think that's close enough). It's not much about pixel peeping. It's about clarity and contrast, as I see it, but of course I'm looking towards your opinions.

 

I'm completely aware that downsampling the D810 pictures although they went from RAW to JPG, was not doing justice to them. But again, it's not about resolution in absolute terms. To me it's about a believable texture. It's like lifting the mist produced by interpolation of a Bayer-pattern. I just want to be clear: I don't want to convert anyone to Sigma, I wouldn't want to give up my Nikons for such a heavily handicapped camera (slow, old fashioned display, not even a cable release socket). But for some pictures those tiny and these days cheap DP Merills still show the fat-ass, high resolving FF sensors who's boss.

 



 

That was a box I found, with a nice structure and nicely punched text on it. Now, in order of appearance some crops: D5100, D810, DP3M, no sharpening except what Aperture did by default. I adjusted exposure and colour. Large pictures can be found here

 



 



 

Another one:

 



 

Same order of appearance: D5100, D810, DP3M

 

[popo]

The differences to my eyes on the 1st test are small but there. As much as I hate the phrase normally, the Sigma output seems to have more "3D effect". The 2nd test is more obvious, especially on the bridge details and also the fine details of the tree branches.

 

I know this was "no sharpening except what Aperture did by default", but I did try sharpening the middle image to see if I can make it look closer to the Sigma. In terms of fine details, I randomly used about 40% sharpening at 0.5 pixel radius, which made the trees look much closer although still not quite as good. This didn't help the bridge detail at all. There I went into unsharp, didn't note exact settings but think it was again around 40% with 2 pixel radius and that help enhance the bridge texture, although as I didn't spend time to work around it, it introduced obvious halos in other areas. Perhaps other software would have other options to more easily allow enhancement. I know sharpening comes with the risk of making noise worse too, but working out equivalents is too much here.

 

At the pixel level, the Sigma output might even feel a little too edgy for my tastes. I can't put my finger on it. It feels like borderline oversharpening, or is it just me being used to the bayer blur?

 

Remind me, does it use an AA/low pass filter or not? That may be a contributing factor also.

[davidmanze]

Impressive!

[JJ_SO]

@popo I don't know with certainty, but I think there's no AA filter involved in front of a Foveon sensor. I also have some detail pictures (the 50 mm of the little Merill goes as close as 0.226 meter) of textile or other structures which increase the danger of Moiré, but as each Foveon pixel (in that "older sensor, the new quattros are different) contains RGB, the don't need AA. Which I think, is the main factor for this kind of sharpness - no interpolation of 22% green, 63% blue and 71% red leading to this color
per pixel. They just don't need to merge the color values of 4 pixels to one.

 

I see it the same way, sometimes it appears overly sharp. And when it comes to fine texture, I'm always amazed about seeing a kind of "relief" (I just wanted to avoid the "3D" phrase, but I know what you mean). The prints of those pictures have a special quality. I try not to get used to it too much. The disadvantages of the concept can't be forgotten. But for a normal walk with no intentions in long teles… it's nice to have altogether 2 kg to carry, with a small tripod including the bag, two cameras in and the necessary spare batteries. It takes time and needs patience which is not always a bad thing.

 

See the little plane above the clouds? Bigger version And here's a gallery with some water and ice involved. 

 



 

In that gallery there was a little canyon, trees, plants, little animals like this 

 



 

Maybe it becomes a bit understandable why I'm defending Sigma and their cameras sometimes?  ^_^

[Klaus]

yes, impressive.

 

It is a bit of mystery to me why Sigma didn't join MFT or released their own mirrorless system.

In the DSLR space they never had a chance (their cameras).

[JJ_SO]

I don't understand it either. Given the delay they usually have between product announcement and shipping, I'm not sure if I want them put more efforts to get contemporary competing systems. Their resources are limited. I mean, the feature list of CaNikonSony is quite lengthy. It's also software and the choices are only two apps: Either genuine or Iridient, none of them really user friendly for a long time.

 

I think, their cameras are victims of their lens success and they want to stay small: big factory, small office is a motto of them.

 

Actually, I didn't have their cameras remotely on my "investplan", until I saw some prints when I took my Sigma lenses to a checkup (Sigma Switzerland offers lifetime check once a year). Those prints were only A4 and a comparison between a Pro FF with Bayer sensor and the Sigma SD 1. The photog who delivered the prints was a pro and took it for travel pictures to Middle east. Mine are "only" the small DP looking a long time like overpriced point and shoot. Then I checked the prices and at the same time got mail from a friend who also fell in love with Foveon although he also doesn't want to give up his 5DMkII. Things fell into place, I paid about 400€ per unit and am currently observing the delivery date of DP0quattro, which has a (on FF) 21mm lens.

 

The DSLR is for my purposes too limited and the postprocessing takes longer due to slow software. But I'm really happy with my "little medium formats" and I like the menu and handling better than Nikon's.

[borisbg]

IMO Sigma doesn't have a final concept where is going with their camera line.I would say collaboration with some of the other camera makers. Apparently there is demand for premium compacts, or like Klaus suggested MFT, instead the line of fixed lens compacts, one camera with 3-4 primes.

[JJ_SO]

There's still the small advantage of a dust free sensor at all times, when the lens is fixed. The primes in front of the Merills are really good (ok, AiryDiscus will disagree, but anyway…). I must say, although the DP0Mquattro will cost a real good wide-angle lens, I'm tempted to get it. Once it becomes available.

 

I think you're right, the concept is hard to detect, if there's one at all. Except "we will stick on camera making as long as possible although we loose money". But I don't believe it would be easy for Sigma to get a good manufacturer, because they will keep on with their lenses and that's competition to each camera seller.

[Brightcolours]

I'm not a fan of the AA-less fake sharpness/false detail of Sigma cameras. I wonder what the quattro will be like, as it does not do every colour per pixel anymore (in an effort by Sigma to lose less light like they lose with "traditional" Foveon sensors.. The green and yellow ("red") layers have lower resolution, the detail comes from the blue top layer).

[JJ_SO]

"fake sharpness/false detail" is really one of the most stupid judgements I heard about those Sigma camera output.  At least, it's a unique and original one and you don't have to share that opinion with many others.  :lol:

 

I could understand the term "overly sharp", but what makes you say fake and false - and what is in you definition then "true and righr"? Foveon is a different type of light perception. Anyway, it's just one of BC's statements when everything he doesn't use himself can't be good, so I'd be a fool to take it seriously.   It would be so easy to find tons of disadvantages of the Sigma concept, but the one thing they have a clear advantage is what you're critizising. Just poor. It doens't add more sharpness than there is. But hey, even camera dealers who are unfamiliar with Sigma, mumble something about "in camera sharpening", yeah, sure, there must be some trick when it takes 5...10 seconds to save this 50MB RAW.

 

You're of course the big specialist in true sharpness with your Canons.  B)