Sigma DP1 priced, dated, and manhandled

Looks like the Sigma DP1 has a date and price: according to their site, it'll be out tomorrow for ¥89820, or about $850 US. Those that just can't wait to find out more about this sucker should check out Pop Photo's very early first impressions. Although they don't have a lot of conclusions to make about it in their first 24 hours of testing, there's certainly a sense of excitement in the air about a smallish point and shoot equipped with an over the top 14 megapixel FOVEON X2 CMOS pulled right out of Sigma's SD14 DSLR.

[Thanks, Tano]

Read - Sigma's price n' date
Read - Checking out the DP1

[shaliron @ Mar 2nd 2008 6:28AM]

This camera represents the opposite of the current camera market. Instead of stuffing megapixels upon megapixels in a tiny sensor with gimmicky features, this camera strives for image quality front and foremost.

May this signal a new range of cameras in the marketplace for the enthusiast.

[z @ Mar 2nd 2008 6:44AM]

the only depressing thing about this camera is the f4.0 lens...

[Yem @ Mar 2nd 2008 8:00AM]

and no stabilization to offset that.

[Yem @ Mar 2nd 2008 8:00AM]

and no stabilization to offset that.

[peayq @ Mar 2nd 2008 10:50AM]

and no stabilization to offset that

this camera has a 16mm lens (28mm in 35mm-format equivalent), which means you should be safe from camera shake at speeds down to 1/10-1/20. stabilization is for sucky ultrazooms and is by no means a compensation for lens speeds.

[dustin @ Mar 2nd 2008 11:28AM]

You can't have any larger of an aperture than that with this sensor. You'd have color shifting.

[Z @ Mar 2nd 2008 11:34AM]

thanks for the color shifting info, it finally does make sense, thought it just kills off the camera to me (and I guess to many other semi-pro or pro willing to have a small form-factor cam)

[gnarayan @ Mar 2nd 2008 7:00AM]

and that price tag...

[contreras @ Mar 2nd 2008 7:29AM]

I think the goal here is quality with excellent portability. It will influence the portable cameras market, in a great way.

[z @ Mar 2nd 2008 10:24AM]

I don't know, but it doesn't look so much better apart from that false 14MPx resolution than a Ricoh GRII 10MPx and it's 28mm F2.4 (!) or even the Lumix LX2 with it's 28-105mm F2.8... for a price that's a bit steep

[Adam @ Mar 2nd 2008 9:09PM]

I'll stick with my Leica (d-lux 3) 10mp is plenty, raw, manual control, and great video in wide screen.

[Superprime @ Mar 2nd 2008 7:57AM]

Eh check out the shots. The image quality is lackluster

[Irfan @ Mar 2nd 2008 10:02AM]

Do you realize that they show a compressed scene first (about 10-20% of true image size, so you cant really make a quality judgement from that besides color and exposure), followed by 100% crops of specific details in that picture? Standard reviewing technique for every camera review website. They are zoomed in to 100% pixel size to show how detail and noise compares to others. showing the entire 14MP picture on your screen would require a ton of bandwith for them, and lots of scrolling for you. they dont have the same scenes with other cameras to do a side by side comparison, but id bet its pretty good.

[Superprime @ Mar 2nd 2008 11:07AM]

Well the 100% parts should be enough. I agree it's not a profi image quality test but it does give a general impression and that impression is bad

[Chris Paris @ Mar 2nd 2008 8:28AM]

I agree, the sample shots posted at Pop Photo are horrible -- what I am assuming are either zoom shots, or cropped closeups of higher res shots, are grainy, pixelated and dim. Many of the shots literally look like they were taken with a 1 MP camera phone. What is up? It just doesn't jive with the text of their review. Could a bad photographer be the culprit? Or do we need to know the context of the sample photos? Whatever it is, those shots are in complete opposition to the reviews and expectations, as they are simply terrible, terrible photos.

[Seung-Hwan @ Mar 2nd 2008 8:49AM]

I seriously hope you're kidding.
Please learn to read.

[Leoedin @ Mar 2nd 2008 10:39AM]

Did you even read the fudging descriptions? 100% crops at ISO 800 - they're not perfect, but the certainly look rather good, I'd say.

[Piero @ Mar 2nd 2008 8:57AM]

The sample shots look horrible because they used the "14 Mpixels" conversion from the raw engine... Which is an extrapolation of the 4Mpixels picture that the sensor REALLY outputs. The 100% samples of the 4 Mpix files (some can be found on the net) and the other hand look amazing (the sharpness is astounding). But when you think about it, for 4Mpix pictures extrapolated @ 14 Mpixels, these samples loo quite good !

[Z @ Mar 2nd 2008 10:28AM]

and for a great 4MPx cam, with a great output you might as well get a used Nikon 4500... with a great zoom and the swivel body...

[Zzephyr @ Mar 2nd 2008 9:54AM]

As I recall, Foveon sensors were considered to be a gimmick a few years ago; upon detailed review they were judged to be lacking performance. Perhaps they have improved significantly in recent times?

[z @ Mar 2nd 2008 1:06PM]

an interesting wiki on the foveon
http://en.wikipedia.org/wiki/Foveon_X3_sensor

[David @ Mar 2nd 2008 1:39PM]

"...Foveon sensors were considered to be a gimmick a few years ago; upon detailed review they were judged to be lacking performance. Perhaps they have improved significantly in recent times?"

Not so much.

1. Foveons use a stacked-pixel structure, relying on the difference in absorption depth to separate out colors. Problem is that although blue absorbs quickly near the surface, green can be absorbed in a range from the surface to deeper in the silicon. Thus, the blue channel also detects a lot of green. Both blue and green similarly detect a lot of red. This overlap in sensitivity makes it more different to separate out color accurately and error can also lead to noise.

2. Once the top channel of a pixel is saturated (near clipping), you need to stop the exposure to prevent clipping. However, the deeper channels are less sensitive and tend not to be near clipping. So, the top channel, which saturates quickly because it detects all 3 colors of light, becomes a bottleneck that keeps the deeper channels underexposed. This is a fundamental inefficiency that leads to noise.

3. A Foveon sensor with, say, 5Mpixels is able to yield both luminance and color resolution that are the same, proportional to this 5Mpixel value. So, 5Mpixels x 3 channels, of 15 million optical elements, give you "5Mpixels" worth of both luminance and color. On the other hand, if you put those same 15 million optical elements to use to make a 15Mpixel Bayer sensor, you end up with the luminance detail equivalent to about 80% of the pixel count or the equivalent to what a 12Mpixels full-color-pixel sensor would have produced. What suffers is color resolution. However, real-world images have very little color detail. The detail is mostly in the luminance. So, Bayer sensors give you more luminance and less color data compared with Foveon sensors that put out the same quantity of data. This is a better real-world match, giving you more of what's important and less of what's not. From an image-quality per unit of data point of view, Foveon does not win.

You can see this by taking, say, a 10-12Mpixel Bayer image such as from a Canon 40D or 450D and properly downsampling to the output resolution of the DP1. Done correctly, you end up with the same "pixel sharp" images that the DP1 is producing ... and that's after having thrown away data by downsampling.

4. The pixel-sharp look of Foveon images comes from the lack of an anti-aliasing optical filter in front of the sensor, not from the sensor itself. Now, it's true that the Foveon can better get away with this compared with Bayer sensors. Any time you do away with this filter, you will end up with artifacts in your images. In the case of a full-color-pixel sensor like the Foveon, these artifacts occur as false patterns in texture, monochrome Moire patterns, jaggies in diagonal lines, etc. In a Bayer on the other hand, you can get color Moire patterns (with a rainbow look) that are more noticeable to many. Thus, Bayer cameras can't get away with dropping the AA filter. Still, Foveon cameras are capturing their own artifacts from dropping the AA filter and really should not be doing so. The images look good at first glance, but the images are less accurate than with an AA filter in place.

David

[o0o @ Mar 2nd 2008 1:25PM]

seriously, by looking at those sample shots, i think my 3 years old g6 still can kick dp1's flat butt...

[phez @ Mar 2nd 2008 2:43PM]

As much as I adore the photo elitists jumping on the sigma defense force - this is still a fixed-lens point-and-shoot that's going to cost you well over $850.

No matter which way you spin it, you're mentally insane to buy this over a $200 canon.

[max andrews @ Mar 2nd 2008 2:51PM]

I said it back when this thing was announced and I'll say it again:
this camera is pointless and a waste of money.
First of all, it's NOT a 14 megapixel camera. It's a four megapixel camera. If you want 14MP it upsamples and looks like crap, as you can see from the sample photos. As long as you keep it at 4MP your images will be sharp and have great color, but have fun printing anything larger than an 8x10. C'mon 4MP?! what is this 2002?

Next up, the lens. First of all, it's a prime wide angle. Why not a more mild wide angle such as a 35mm, or standard 50mm? That would reduce color fringing issues and allow a much larger aperture. Especially with being capped at ISO800 and the poor luminance noise at that level, this camera is very limited in low light situations, and useless indoors.

NO MANUAL FOCUS RING. You get a bullshit dial when all they had to do was put a ring on the lens. Sorry, but for $850 you're gonna have to give me legit manual focus.

No built-in optical viewfinder. You have to dish out an extra $140 for a clip-on rangefinder. So Now you've dropped $1000 on this thing, and it's much less portable.

Shitty video mode. Sorry, but if this thing is supposed to be a compact where's the decent video mode? Almost all consumer compacts can manage at least VGA video these days, and some can manage 720P. This gives you a camera-phone level 320x240, which is bullshit. I might as well draw a flipbook.

Non standard RAW files. You HAVE to use sigma's raw converter, since the X3F raw format isn't supported anywhere else. Yeah, have fun with that! This means an whole extra step of converting and saving all your images if you want to do HDR processing. Plus, sigma's raw conversion engine is probably not the best out there, but you're stuck with it.

[zak1516 @ Mar 3rd 2008 1:40PM]

you should really read up on foveon processors i have a sigma sd14 and its just as good on quality as my olympus e410 except for the colors are better on the sigma, i will say that it is not equivalent to 14 mp but im sure your blowing every photo up to huge dimensions

[L. Kraven @ Mar 2nd 2008 4:11PM]

Holy hell, that's a lot of misinformation and lack of research.

As for the video, I'll grant you that the resolution is rather low, but here's a link to some example video:

http://www.rytterfalk.com/2008/03/01/sigma-dp1-movie-clip/

One thing you may notice right away is that for 320x200, that looks worlds better than what you get from most 640x480 point and shoot video. Second thing is the background is blurred. Try doing that with your typical point and shoot.

Regarding the "proprietary" raw format... which manufacturer doesn't have one? They're all proprietary. About the only major piece of software that won't decode the X3F is Apple's Aperture. Lightroom, Photoshop, Raw Developer ALL decode X3F files.

I see a lot of knee-jerk reactions to this camera. OMG! It's $1000!! For a point and shoot! If pixel count is all that matters, you're absolutely right. But here are a list of things this camera does that NO other camera does:

1. Pocketable point and shoot with a DSLR sized sensor/quality.
2. High quality 28mm prime lens (a good prime for a a DSLR will cost almost as much or more than this entire camera.)
3. Point and shoot with decent DOF control.

Did I mention it's pocketable? I have a Canon 5D which I love... but it is huge compared to this camera. I have a Canon G9, which I love, but the quality is not comparable, based on what I have seen.

This camera is NOT for everybody. In fact, I would not recommend this camera to ANYONE who is asking, "What kind of camera should I get?" The people who are interested in this camera already know how and why it meets their needs.

As for the pop photo shots, yeah, they're horrible. I have seen much better, and here's a gallery of shots out of camera you can try to compare to any point and shoot out there:

http://www.sigma-dp1.com/jp/sample-photo/

I think the people trying to compare this to a Rebel, or a D40 or whatever, are barking up the wrong tree. Those are great cameras, great kits, but they are not pocketable. I TAKE my 5D places. I plan to BRING my DP1 places. The difference is when I have the 5D, I am going out to shoot. When I have the DP1, I am out doing something and have the opportunity to get great shots while I am doing it.

I wouldn't compare this with a $200-$400 point and shoot either. It's kind of somewhere in between. If you don't get it, or think it's ridiculous, you probably shouldn't buy it. But I know plenty of people who are really really excited about this camera.

[max andrews @ Mar 2nd 2008 6:29PM]

@ L. Kraven,
You start off by saying my post was full of misinformation and lack of research, but you fail to prove anything incorrect about it. It specifically says in the review article that the DP-1 X3F files are not compatible with any software except the sigma conversion software, so you will need to rely on this for raw conversion for the time being.

And I could overlook a few of what I consider to be faults if they were not lying about the megapixel count. It is not a 14 megapixel camera, that is a straight up lie. It is a 4.6 megapixel camera. In the SP-1 sample gallery you linked to, full of hand selected images that are supposed to make the camera look good (a few of them still are still ho-hum), you'll see that the actual resolution is 1760x2640, which is 4.6 megapixels. You can get more than twice the res for half the price with a high-end compact. The G9 will produce better image quality if you downres from 12 megapixels to match the 4.6 of the SP-1, and even so I'd much prefer a 12 megapixel image to make a decent print after some color work.

So call it what it is. It's an $850 4.6 megapixel compact camera with a fixed lens, very similar in size to the G9. I personally think that's a bad deal, as if I really care about images I can use I'm going to bring my DSLR. If I want a camera with great quality but above all a compact form factor, I'd stick with the G9 for far less coin. Just stop calling this thing a 14 megapixel camera. It's not. If this was truly a 14 megapixel shooter I'd have no hesitation recommending it and maybe even picking one up for myself. But it's a thinly veiled marketing campaign for what is as of now a very expensive camera with few features and substandard resolution.

[Robert in Texas @ Mar 2nd 2008 10:14PM]

It is a 14 MP camera just as much as a bayer sensor camera is 14 MP. The Foveon has 14 MP (actually 13.8) divided by three stacked colored pixels or 4.6 million red pixels, 4.6 million blue pixels, and 4.6 million green pixels. A 14 MP bayer sensor will have 3.4 million red pixels, 3.4 million blue pixels, and 6.9 million green pixels because of the bayer array layout.

[max andrews @ Mar 2nd 2008 11:09PM]

@ Robert,
not quite. You are correct about the bayer sensor color detail being the same. But a bayer sensor has full luminance detail at every pixel. So with a 14MP bayer sensor you get 14 full MP of B&W detail and then the interpolated color detail only, which tends to be done quite while in the higher and cameras. With the foveon, you only get 4.6MP of luminance data. So if you want 14MP, the entire image is interpolated, not just the color, so it ends up looking a lot worse than a 14MP shot from a bayer camera. Still, you're splitting hairs. The full-res images out of this camera are 4.6MP, end of story. Anything larger is made-up detail in both the color AND luminance.

[Jani Tarvainen @ Mar 2nd 2008 3:25PM]

test

[Jani Tarvainen @ Mar 2nd 2008 3:26PM]

test

[cpt @ Mar 2nd 2008 3:50PM]

Here is one sample night shot, and I honestly don't think my Canon G7 is ever gonna produce a night shot with such image quality:
http://www.slrclub.com/bbs/vx2.php?id=sigma_forum&no=49674

Another interesting comparison is DP1 vs 5D vs GX20(which is essentially K20D):
http://www.slrclub.com/bbs/vx2.php?id=sigma_forum&no=49188

IMO the DP1 is not for everyone. You need to know exactly what you're doing to get the best out of it. And the best out of it could be truely amazing.

[Ypoknons @ Mar 2nd 2008 6:33PM]

For people who haven't noticed yet:
BIG, FOVEON SENSOR

That's what makes it different from other compacts.

[bebop @ Mar 3rd 2008 2:04AM]

Pity they didn't just leave it at BIG SENSOR, i.e. a standard APS-C sensor from Sony that most everyone else uses in their dSLRs. Lower cost, image quality guaranteed to be at least on par with dSLRs... and they might just have managed to bring the damn thing to market back before everyone lost interest and went out and bought a D40.

[Sandy Fleischmann @ Mar 3rd 2008 8:54AM]

see sample photos here http://www.sigma-dp1.com/sample-photo/index.html and an informative brochure is linked for download on the same website http://www.sigma-dp1.com/main.html which will explain too on the Foveon technology. Good discussion forum is at www.dpreview.com here http://forums.dpreview.com/forums/forum.asp?forum=1027

[tallgrave @ Mar 3rd 2008 10:52AM]

Hello.
I am tano.
The comment here was read interesting!

In Japan, DP1 of initial production missed and went to the person of some.
And, it doesn't reach my hand.
In a word, it regrets it though I want to test a real machine.

Early adopter has already updated the preview of DP1 in various places.
It was an image that finished being satisfactory as long as I saw with web.
The design of FOVEON and the lens is at least good.
I think that the distant future work becomes unnecessary when this level continues though I am doing the work of the image data processing.