Some new SD1 samples!!!
i dont' know what to think..
fromwhat i see now these images looks more like oversharpened uprezisedfiles, not fine details, poor clariness, lots lots of artifacts like halos.
in my opinion they will struggle even against a nikon d3x and a good glass.
they are not comparrable to any mf format camera, and in my opinion the new 30 million camera will beat them for resolution.
probably the firmware will fix something, but i doubt they will change things completely
this is a photo i took yesterday in france. it's a 100 crop, i oversharpened it to match the level of sharpness of sd1. I didn't do any pp for color and contrast just sharpening in lightroom. i cropped the same position. persoanlly i struggle to see the difference in IQ
im sure the sd1 in good ahands will beat any apsc sensor out there, we must wait for the new sony 24 million, but doubt the difference are so big, and i doubt this camera can deliver 30 million pixel quality.
it's a superb 16 million aps csensor in daylight probably.
well i see a lot of artifacts and a lot of oversharpening, not really great fine details.
i have sd14 and i know the per pixel sharpness at 4 million pixel. is much much better than this.
personally considering all the problems linked to foeveoon, and that seems also in the sd1, i ll take my k5 nigh and day, simply beacuase i can shoot everywhere and in everry ligh tcondition.
in myy opinion thi camera is overpriced even at 2000 dollar.
Well, I don't think these tester will change the setting too much. I tend to believe that they may shoot with default settings mostly and adjust a little bit. It seems like Sigma is facing a new challenge to take care of this new baby chips.
i opened them and i m not impressed. any bayerish camera with a good glass can do that. thinking this camera can reach 45 million is cheating, even at 28 is cheating. this camera is on par just with the actual 24 mm full frame.
in my opinion all the photo show a lot of sharpening and artifacts due to this.
i need more sample to a complete evaluation but i'm thinking the foveon magic end up at 4 million pixel.
Well, I believe that there may be some bayer camera can do this, but not that much. When considering open at 100% to view the image, the number will be reducing to single number. Just a few photos from FF+great lens satisfy me with the same result under full image view.
Just wait and see for the final firmware.
Nevertheless, I agree that SD1 should not cost this much. A terrible price tag.
IMO, these photos in the article are too small to show the merit of X3. Right click on the image and open the image in new window. These are pretty impressive shots.
this unscientifc comparision clearly. simply put, if the sd1 was so superior to a bayerish 16 apsc sensor, that would be showed in any samples, even the worst.
the difference are in my opinion pretty low, and k5 is 1000 camera rememebr.
i have analyzed all the samples taken, apart the one with the 70 macro and rocks, the other don't shotw nothing more than what u can achieve with any bayer camera.
stanislaw stitchanow wrote:
with that attitude you can also post a flower pic and make a comparison.
The photos for my trip start here: http://kofronphotography.com/blog/the-american-southwest-post-2.html
I didn't embed full sized images that could be clicked on, however...
Ironically, the best parts hinting what the sensor is truly capable of are the central parts of the photos shot with the 8-16mm. It' a really good lens! But even the best primes like 50/1.4 and 85/1.4 can't make pictures sharp corner to corner.
I am sure colours, noise and DR will be better with the new firmware, but if the pictures simply aren't sharp there's nothing can be done!
Yes, but also green and magenta traces on ISO200 - in the sky and in the shadow parts.
It is also very clear that the camera is lens limited, this is not that unexpected but in one sense it is just a shift of 'problem' from sensor limited tot lens limited.
Still, very nice to see more examples and with a great variety of lenses.
If the price had remained on the original indication the camera would have been a big hit.
You may want to look into calibrating your monitor. I pulled the images and looked at the RGB values in the skies and there is no magenta cast whatsoever. In the shadow areas of #19 there is some faint green & magenta in the shadow area of the hill.
But there is noise and banding in the sky of the second shot.
The subjects chosen are hardly exciting...I'd rather see some head and shoulder portrait shots than boring shots of a barb wire fence in the middle of nowhere.
I've read about 5-ish posts, and there's a lot more below the level I'm at, but I decided to just assess these sample images without any further influence. I agree that the most 'Foveon-esque' image at 100% in the bunch from the SD1 is the aforementioned colourful rock close-up using the 70mm macro. The super-wide zoom and the primes shots aren't bad, if you can crop out the outer 20% margin for unacceptable sharpness and probably-beyond-fixable CA (yes, I know that it can be fully removed with this degree of occurrence, but doing so makes it even harder still to preserve any 'Foveon look').
Boy-oh-boy, do the optics play a huge role in the image quality.
I hope all my EXes will be up to muster, or more realistically, just not made to look ridiculous. This sensor eats glass for lunch!
An interesting thing is that if you crop out the outer edges, and then do an even-fractional downsample for the remaining part of the image - it begins to look like a Foveon shot again for many of the shots. Actually, it often looks beautiful, when you do that.
Funnily enough, the DP2s shots look pretty good, quality wise (excluding composition). When trying to pick the best SD1 shot early on, I accidentally included the DP1x shot, then had to remove it when I came to that realization.
I think I'm beginning to really appreciate my existing Foveon!
I would love to see some low ISO shots. Plus I think some of those SD1 shots went beyond the cameras's aperture capabilities in terms of emphasizing circle of confusion for a super-dense sensor. And how were these processed? The JPG EXIF says Photoshop Elements for Windows v.7, but the bottom of the web page says Sigma Photo Pro. Please tell me these were at least exported lossless before processing them next in Elements. Even then, you can get side effects when processing things like noise and sharpness more than once or using successive passes. To me, I think most appear to have too much sharpening and possibly weird noise-removal artifacts applied? There are posterization blotches (sometimes even streaks) in the areas of super-subtle colouration changes, and that's one of the beautiful things Sigma handled better than anyone else in the past. Maybe that's the result of Post-Processing the shots twice (it said there was dust-removalicious steps or translations to that effect).
The only way these SD1 shots really look like Foveon images is when they have the outer lens effects cropped and the remainder evenly downsampled.
Now I saw some other SD1 samples by a different photographer (someone in the U.S. - there was a link in this forum about a week ago) at three magnifications: a reduced full-frame reference shot, a 100% crop and a 500% crop, and they didn't look bad at all. In fact, they looked like what I was expecting from this camera. The images were of a coin and a hand, if I recall correctly.
In summary, my only two points are that:
1) hopefully the firmware and SPP (or Elements?) will improve by launch, and,
2) if they can't, it looks like I'm gonna be able to get myself one of these for CAD$800 much faster than I hoped for, so I can crop and downsample them and use it as a faster, weather-sealed Sigma body, and not have to upres images anymore!
Anyway, that's my gut response before reading any further posts that might cloud my perception. Hope I wasn't nuts and letting a knee-jerk reaction rule my rationality. Apologies in advance if that was the case. Enthusiasm rapidly depleted takes sense right along with it.
Sad for me as I want that lens one day... which is why I looked for and saw the decentering before I read your message. Its the Sigma lens lucky-dip.... put your $$$ down and hope you get a good one. That level, for me, is unaceptable, and I have seen even worse out of that lens. What am I supposed to do... buy the lens and have to send it back and say its unaceptable? Sheesh... it must be acceptable for Sigma or they would not have let it out the door.
Oddly enough all the other 8-16 shots look rock solid... its only that one shot, but its not related to aperture... so whats going on eludes me.
This looks much much softer than a real (sorry: old) Foveon shot.
I still think the problem is the lens or the shooter, not the camera.
BTW what is a 181% crop? Did you uprez the file to 181% of its original size? Any reason behind the number?
This is a 181% crop of the petrified forest shot. Does anyone see any "Bayer-ish" here. Maybe a direct comparison with a D7000 or 7D would be helpful.
Click Original to view full size.
To me it looks like Foveon all the way down.
If this was made with firmware 0.91 - what's fw 1.00 going to look like?
Capture all the light and colour!
. Does anyone see any "Bayer-ish" here.
Actually, yes. The red channel of that image is not as sharp as the green. Here is a 100% crop of that image with the two channels side by side.
View the original in my gallery and it's clear the red channel is not as detailed as the green - which is exactly "Bayer-ish"
Even though I don't own one, I've admired the SD1 for years now, but these photos just don't have the same pop. The colors aren't there, and they seem harsh, almost over sharpened, not smooth and rich.
There are some DP shots but most of them are SD1. Also, a large portion of the SD1 shots are ISO200
I don't see good colours and good resolution...
Really poor IQ. The concept of SD1 is strange for me. My panoramic camera costs 5 times less and produces 150 Mpx.
My film SLR costs 100 $ and produces 20 Mpx.
If it would cost 800$ I would buy SD1, because I really don't like how Bayer sensor's images does look sometimes.