If you do a lot of hiking or touring on foot, I can easily see the appeal of a pocket rocket that can have 5D Mk2 quality++ with a small tabletop tripod or TrekPod. Slogging through a very humid jungle, I would rather have a 2 liter water bottle than a 70-200 F2.8.

There just seems to be too many compromises here for too much money to make it worth it though on first glance.

From the samples someone posted earlier, the DP2M vs the NEX-7 + Sigma 30mm F2.8... there doesn't seem to be much advantage to the DP2M and if anything, seemed to resolve a bit less detail in the medium range print. Add that on top of ISO, dynamic range, post process workflow, pre-processed RAW files, camera responsiveness, and cost...

FlyPenFly wrote:

From the samples someone posted earlier, the DP2M vs the NEX-7 + Sigma 30mm F2.8... there doesn't seem to be much advantage to the DP2M and if anything, seemed to resolve a bit less detail in the medium range print. Add that on top of ISO, dynamic range, post process workflow, pre-processed RAW files, camera responsiveness, and cost...


Where were those DP2M/NEX-7 samples posted? I must have missed those.

cputeq wrote:

2) That same mountain photo looks like it's had a round of unsharp masking done and then a very fine level of tonal contrast, which gives it the "surreal" look. I can't really find many sharpening halos (unless you count the mountains, which looks like CA+Halos) - definitely interesting.


This is pretty much what I was thinking. I was searching these images for telltale halos from a high pass overlay layer, as that's the feeling I have viewing these... and that the strong/deep blacks remind me of some film images. And of older, dynamic range restricted sensors (compared to some recent higher range ones), where the manufacturer intentionally used a strong tone curve to hide noisy shadows by keeping them quite dark. IIRC, Canon did this with the first 1D to hide the sensor's banding tendency. When users complained about it, IIRC, one of the first fixes was a new camera tone curve that darkened the image even more. But to my eyes this is very much the look of transparency film. Dark, somewhat sooty shadow tones. Contrast this with D800-like images where we're opening shadows that maybe would better be left alone.

Regarding color quality - there is likely a conflict going on here between the sensor and software's ability and our individual perceptions of what colours should be. It's possible we've been so inundated with Velvia-like and Kodak Gold-like images over the years that images with low saturation in certain areas feel unnatural, even if they might be closer to reality than we'd like to believe. These images kind of remind me of Kodachrome. Everyone seems to think of it as some wonder film with great colours. And it could be, but I had my share of bland looking images from it as well, where I expected higher colour saturation but didn't see it in the final piece of film. This is what I've seen in a number of DP2M landscapes with distant vistas, where I would expect more vibrant greens in foliage but instead it looks like -50 saturation.

Regarding the edge outlining crop on the previous page: I get similar effects when I do selective luminance adjustments for specific colours in LR. I call it my 'poor man's polarizer' because often I will reduce the blue luminance a bit to darken skies... but the consequence is often a one pixel white line where the sky meets foreground.

FlyPenFly wrote:
If you do a lot of hiking or touring on foot, I can easily see the appeal of a pocket rocket that can have 5D Mk2 quality++ with a small tabletop tripod or TrekPod. Slogging through a very humid jungle, I would rather have a 2 liter water bottle than a 70-200 F2.8.

There just seems to be too many compromises here for too much money to make it worth it though on first glance.

From the samples someone posted earlier, the DP2M vs the NEX-7 + Sigma 30mm F2.8... there doesn't seem to be much advantage to the DP2M and if anything, seemed to resolve a bit less detail in the medium range print. Add that on top of ISO, dynamic range, post process workflow, pre-processed RAW files, camera responsiveness, and cost...


I'd like to see some samples done with fine landscape detail at farther away distances. That seems to be where the DP2M performs a bit differently than most cameras. Things like the tips of trees, individual leaves etc just seem to still be rendered in fine detail while they have gone a bit soft on most cameras.

I had the Sigma 30mm for my NEX7 and it was quite a surprisingly lens for $199. It was a little slow to focus, but it was darn sharp and while a little slow, had a rather pleasant bokeh quality as well. Honestly, in some ways I felt it was better than the Zeiss 24mm I also had for less than 1/5th the price, and a much nicer size. The Zeiss 24 did have a bit of a signature Zeiss color that was different though.

I sadly have long since sold the whole NEX kit, but certainly will look forward to seeing how the Sigma compares with the little RX100 on one side and the heavyweight D800e on the other. Really more so to the RX100 though as I for me at least, the D800e is just a camera I use far differently and that is going to have obvious advantages in terms of DR, high ISO performance, lens section etc, provided one is willing to carry it for non-professional outings.

RX100 though should provide a bit interested, in that its got a pretty decent lens and 20megs of resolution, which when shot in RAW and at low ISO, can deliver some impressive IQ, from a little pocket package that is much cheaper than the DP, focuses much faster than the DP, has better high ISO than the DP, has a better LCD than the DP, and better battery life than the DP.

All and all, on paper even a $650 pocket camera seems to trump the DP in most areas, but I'm hoping at least to see a huge IQ gap between them, under the shooting that I do, which in turn would also mean I would of seen a smaller, though still pronounced gap if the test was with the NEX7, EM-5, X100 etc that I no longer own, because the RX100 in my eyes could come surprisingly close to matching those in the right condition.

FlyPenFly wrote:
IFrom the samples someone posted earlier, the DP2M vs the NEX-7 + Sigma 30mm F2.8... there doesn't seem to be much advantage to the DP2M and if anything, seemed to resolve a bit less detail in the medium range print. Add that on top of ISO, dynamic range, post process workflow, pre-processed RAW files, camera responsiveness, and cost...

My impression of the samples was that there wasn't an overwhelming resolving power advantage in either direction --- the greater per-pixel sharpness of the DP2M made up for its lower pixel count, but didn't reveal detail that the NEX-7 fundamentally couldn't. I did prefer the starting sharpness of the DP2M, and think the NEX-7 image would take a fair amount of PP work to reach the same level (which would exacerbate artifacts/blockiness/"digital-ness" in the image). One strong difference that I noticed was the complete absence of color filter artifacts (that I have always taken for granted in the past) in the DP2M shot: in areas with contrasty detail/texture, the NEX-7 sprinkles multicolored pixels over the image (typical of any Bayer-filtered camera with a sufficiently sharp lens), while the DP2M maintains perfect color uniformity without spurious artifacts. This also makes a difference in the black-and-white versions of the photos, where the NEX-7 fine details sometimes dissolve into "checkerboard" artifacts, while the DP2M maintains its full resolution.

I agree that, in all other respects, the DP2M appears to be a pretty lousy camera compared to everything else available. I hope Sigma ships mine soon.

rscheffler wrote:
cputeq wrote:

2) That same mountain photo looks like it's had a round of unsharp masking done and then a very fine level of tonal contrast, which gives it the "surreal" look. I can't really find many sharpening halos (unless you count the mountains, which looks like CA+Halos) - definitely interesting.


This is pretty much what I was thinking. I was searching these images for telltale halos from a high pass overlay layer, as that's the feeling I have viewing these... and that the strong/deep blacks remind me of some film images. And of older, dynamic range restricted sensors (compared to some recent higher range ones), where the manufacturer intentionally used a strong tone curve to hide noisy shadows by keeping them quite dark. IIRC, Canon did this with the first 1D to hide the sensor's banding tendency. When users complained about it, IIRC, one of the first fixes was a new camera tone curve that darkened the image even more. But to my eyes this is very much the look of transparency film. Dark, somewhat sooty shadow tones. Contrast this with D800-like images where we're opening shadows that maybe would better be left alone.

Regarding color quality - there is likely a conflict going on here between the sensor and software's ability and our individual perceptions of what colours should be. It's possible we've been so inundated with Velvia-like and Kodak Gold-like images over the years that images with low saturation in certain areas feel unnatural, even if they might be closer to reality than we'd like to believe. These images kind of remind me of Kodachrome. Everyone seems to think of it as some wonder film with great colours. And it could be, but I had my share of bland looking images from it as well, where I expected higher colour saturation but didn't see it in the final piece of film. This is what I've seen in a number of DP2M landscapes with distant vistas, where I would expect more vibrant greens in foliage but instead it looks like -50 saturation.

Regarding the edge outlining crop on the previous page: I get similar effects when I do selective luminance adjustments for specific colours in LR. I call it my 'poor man's polarizer' because often I will reduce the blue luminance a bit to darken skies... but the consequence is often a one pixel white line where the sky meets foreground.

+1 great observation, particularly with regard to the shadows and comparison to Chrome.

Well I think what you pay for in lacking color artifacts in a man made scene is the weirdness we're seeing in the files thus posted so far of natural scenes. For not adding sharpening, the outlining effect seems like just a form of aggressive halo effects.

I'm curious why Sigma chose to sharpen the files in software for what is supposed to be naturally sharp output. Then again, I'm curious as to why Sigma would launch with beta firmware.

They sharpen because they have to.

Blue (or the very pale white color Foveon calls "blue" on the sensor) is captured as sharp as the lens can image. Green (which is about 0.5 red + 0.9 green + 0.5 blue) is also "quite sharp". Red on the other hand is about as sharp as a very sharp image that you've run a 1px gaussian blur on - even if the lens throwing the image is almost perfect.

To get the channels to overlap without serious fringing, you have to sharpen the red channel quite a lot, the green channel a little, and then do a secondary pass that blurs chroma information with about 0.7px radius.

AFTER all this is done you can start the chroma noise-reduction schemes that have to be present already at ISO100 due to the error-diffusion between the "color" channels.

The camera does this very well, other raw converters do it... well, let's call it "less well". I've full understanding for that, writing raw-conversion routines for the Foveon is a bloody nightmare. I know - I consider myself to be pretty proficient in this, and I've failed completely.
..........

Contrary to what most Foveon enthusiasts would like to think (and proclaim), the output from the sensor in a Foveon camera isn't even close to a input>>output affair. There's a huge amount of after-processing necessary to make the images even acceptable. This is also why the people most experienced with the problems inherent in the sensor type also are (so far) the best at taking care of (raw convert) the images the cameras produce.

Blue (or the very pale white color Foveon calls "blue" on the sensor) is captured as sharp as the lens can image. Green (which is about 0.5 red + 0.9 green + 0.5 blue) is also "quite sharp". Red on the other hand is about as sharp as a very sharp image that you've run a 1px gaussian blur on - even if the lens throwing the image is almost perfect.

So, is the image technically any better than from sensor with Bayer mask?

I hope you like yours as much as I do.

The dp2m is light years ahead of a Nex 5n with a Contax G 28 in terms of resolution. The images have much more detail, it is not only about being sharper. I owned a Fuji GSW690 (6x9cm medium format film camera) many years ago. I would compare the dp2m output to it. It feels like a medium format pocket camera.
BTW it is not a big deal to shoot handheld. The lens is extremely sharp at 2.8 and the bokeh is very nice. The auto-focus is very precise.

An interesting 1:1 comparison would be between the DSC-RX1 and dp2m. The DSC-RX1 sample images on dpreview published today look smooth but not very sharp. The same scene shot with a dp2m would be definitely sharper. But the published DSC-RX1 samples where just jpgs.

It is not an all rounder. But it does what is supposed to do, to output medium format like images in a pocket format. A great camera when you are willing to live with the ISO limitations film cameras had for decades. The lens is as good as it gets. Better than my Leica 50 Summicron (non Asph).

It is a stock photographers dream camera.



mpmendenhall wrote:
FlyPenFly wrote:
IFrom the samples someone posted earlier, the DP2M vs the NEX-7 + Sigma 30mm F2.8... there doesn't seem to be much advantage to the DP2M and if anything, seemed to resolve a bit less detail in the medium range print. Add that on top of ISO, dynamic range, post process workflow, pre-processed RAW files, camera responsiveness, and cost...

My impression of the samples was that there wasn't an overwhelming resolving power advantage in either direction --- the greater per-pixel sharpness of the DP2M made up for its lower pixel count, but didn't reveal detail that the NEX-7 fundamentally couldn't. I did prefer the starting sharpness of the DP2M, and think the NEX-7 image would take a fair amount of PP work to reach the same level (which would exacerbate artifacts/blockiness/"digital-ness" in the image). One strong difference that I noticed was the complete absence of color filter artifacts (that I have always taken for granted in the past) in the DP2M shot: in areas with contrasty detail/texture, the NEX-7 sprinkles multicolored pixels over the image (typical of any Bayer-filtered camera with a sufficiently sharp lens), while the DP2M maintains perfect color uniformity without spurious artifacts. This also makes a difference in the black-and-white versions of the photos, where the NEX-7 fine details sometimes dissolve into "checkerboard" artifacts, while the DP2M maintains its full resolution.

I agree that, in all other respects, the DP2M appears to be a pretty lousy camera compared to everything else available. I hope Sigma ships mine soon.

Ultimately, it seems so far it's a great idea with terrible execution.

Sigma just doesn't seem to have the expertise and resources required to pull this off effectively. They can't seem to be able to build just a good camera let alone one with a highly specialty sensor. Unfortunately there seems to be too many issues especially with the images themselves. They've had YEARS to fix this since the original DP series. Now that we're taking a serious look at it, the images can sometimes be nice if you don't look too closely. From Suede's explanation, I can understand now why the files look overcooked coming straight out of the camera. They either have no choice or haven't figured out a better way.

I think a much more interesting camera will be the next Ricoh GXR with a 16mpx AA-less micro-lens enhanced camera.

FlyPenFly wrote:
Ultimately, it seems so far it's a great idea with terrible execution.

Sigma just doesn't seem to have the expertise and resources required to pull this off effectively. They can't seem to be able to build just a good camera let alone one with a highly specialty sensor. Unfortunately there seems to be too many issues especially with the images themselves. They've had YEARS to fix this since the original DP series. Now that we're taking a serious look at it, the images can sometimes be nice if you don't look too closely. From Suede's explanation, I can understand now why the files look overcooked coming straight out of the camera. They either have no choice or haven't figured out a better way.




Hmmm, to me it was specifically looking closely at the images that made me want to buy one. They look fine and good at a small scale, as do most any camera, but it was only when opening them up 100% that I really saw something unique and special that made me open my wallet.

Why are people otherwise buying said cameras if not for the 100% IQ it offers ? They are out of stock nearly everywhere it seems right now, and I'm pretty sure no one is buying it for its AF, LCD, Battery life etc.

Probably half the shots I see from the camera are posted along with a 100% crop, if not posting the original image, simply because thats where the "wow" factor falls to most viewers, or I guess I should say viewers who have purchased or plan to purchase the camera.

No one is forcing you to buy the camera, and I know we all have different taste, but I simply don't think it can be generalized that the files are not nice if you look too closely.

If that was really the case, I just can't see why people would be posting 100% crops and raving about them as they do.

Pretty spot on all inclusive wrap-up/review here that covers all the good and bad points of the camera, and puts them into what I think its a pretty good overall context

http://soundimageplus.blogspot.co.uk/2012/10/sigma-dp2-merrill-so-what-do-i-think-of.html

millsart wrote:
FlyPenFly wrote:
Ultimately, it seems so far it's a great idea with terrible execution.

Sigma just doesn't seem to have the expertise and resources required to pull this off effectively. They can't seem to be able to build just a good camera let alone one with a highly specialty sensor. Unfortunately there seems to be too many issues especially with the images themselves. They've had YEARS to fix this since the original DP series. Now that we're taking a serious look at it, the images can sometimes be nice if you don't look too closely. From Suede's explanation, I can understand now why the files look overcooked coming straight out of the camera. They either have no choice or haven't figured out a better way.




Hmmm, to me it was specifically looking closely at the images that made me want to buy one. They look fine and good at a small scale, as do most any camera, but it was only when opening them up 100% that I really saw something unique and special that made me open my wallet.

Why are people otherwise buying said cameras if not for the 100% IQ it offers ? They are out of stock nearly everywhere it seems right now, and I'm pretty sure no one is buying it for its AF, LCD, Battery life etc.

Probably half the shots I see from the camera are posted along with a 100% crop, if not posting the original image, simply because thats where the "wow" factor falls to most viewers, or I guess I should say viewers who have purchased or plan to purchase the camera.

No one is forcing you to buy the camera, and I know we all have different taste, but I simply don't think it can be generalized that the files are not nice if you look too closely.

If that was really the case, I just can't see why people would be posting 100% crops and raving about them as they do.



You're very right Millsart.
You can't actually say much about it if you didn't see the 100% opened original files in SP.
I have nothing seen like it.
BTW SP is no Lightroom but it serves me well, there is not to do a lot on these files anyway. Just export them as TIFF and do the final adjustmens in ACR or Lightroom.

Everybody can download SP for free and find an original file on the net and see for himself. The only objection one can have is: this is to dam sharp. Besides all the quirks one can find, short battery life etc, this camera delivers very high quality files for a very low price. For a lot of things I do this is an ideal camera, and of course one wants always more, but I am sure this will get a follow up, because I think these two DP's are going to sell realy well.

In deed it is all well said in the soundimage blog.

Michiel

Wow. The detail on my MacBook Pro Retina is stunning. I don't think I've seen anything like that out of any other camera. It certainly has a unique "crispiness" to it that other cameras don't achieve.

Japanese again, but had some comparisons between the D800e, 5D mk3, RX100 and the DP2M.

http://type-g.at.webry.info/201208/article_4.html

Not sure the lens used for the DSLR's or the technique so take it with a good sized grain of salt.

Still though, nothing seems to render like the DP2M, in terms of fine detail or edge sharpness. Just take a look at the image with the trees (which maybe is a corner I'm betting)

http://www.flickr.com/photos/type-g/7942167612/lightbox/

Distant tree's have leaves with the DP and to a varying extent, don't with the other camera's, RX100 the most pronounced lack of fine detail of course.


Basically this reminds me of the difference in looking at trees with my naked eye before vs after I got Lasik done. I was used to seeing trees just sort of as green shapes, and while I of course knew what leaves looked like when viewed up close, after being 20/20 it was such a revelation to actually see the trees in my yard were made up of lots of different leaves, not just a green mass.

Here you go, links to the full scene original image from each camera. D800e was using the 24-120 f4, 5Dmk3 was using the 24-105L.

Its pretty interesting to open them all up in PS and play around with them a bit. The amount of you can uprez the DP2M shots and still retain good detail is pretty phenomenal.

Nikon D800e
http://www.flickr.com/photos/type-g/7885578878/sizes/o/in/photostream/

DP2M
http://www.flickr.com/photos/type-g/7885580906/sizes/o/in/photostream/

RX100
http://www.flickr.com/photos/type-g/7885697460/sizes/o/in/photostream/

5D mkIII
http://www.flickr.com/photos/type-g/7885583196/sizes/o/in/photostream/

Millsart,
Thanks much
very helpful

Looks like the 800e was paired with the Nikon 24-120/4, the 5D3 paired with the Canon 24-105/4. Not exactly world-class lenses, but impressive nonetheless.

I'm actually rather impressed with the 24-105L, granted its shot at f8, but still, pretty good performance in relation to its price and versatile range. I sold all my Canon gear and switched to Nikon right when that lens was coming out so never got to use one or paid it much attention but I think I'd own one for my DSLR's if I was still shooting Canon