New Sigma 46 megapixel flagship announced

[BernardLanguillier]

http://www.dpreview.com/news/1009/10092129sigmasd1.asp

Cheers,
Bernard

[Bart_van_der_Wolf]

http://www.dpreview.com/news/1009/10092129sigmasd1.asp

Here we go again, 4800 x 3200 pixels is 15.4MP output. Sure, they are based on co-located RGB-ish sensels, but it is silly marketing spin to suggest R/G/B are separate output pixels.

I'm glad the technology is still alive, it'll keep the other companies alert, but I could do without the spin. Next thing they'll claim higher resolution than mainstream sensor designs, while leaving out the fact that they don't use an AA-filter, making it an apples/oranges type of comparison.

Cheers,
Bart

[Dustbak]

Yes, it is a pity that they revert to this. It is quite a step up from the SD14 though! A pity this sensor technology is only available in the Sigma package. Also a pity they use their own proprietary mount, it would have made sense to provide the body in Nikon or Canon mount too.

[BernardLanguillier]

Here we go again, 4800 x 3200 pixels is 15.4MP output. Sure, they are based on co-located RGB-ish sensels, but it is silly marketing spin to suggest R/G/B are separate output pixels.

I know, but it is still 3 times more than the SD14, which is a more ambitious jump than we have seen from any other brand in the past.

If they have managed to keep noise under control, this might be the DSLR with the highest detail of all.

The lack of Nikon mount is a show stopper for me, but I expect to see many enthousiast buy one in Japan.

Cheers,
Bernard

[deejjjaaaa]

Yes, it is a pity that they revert to this. It is quite a step up from the SD14 though! A pity this sensor technology is only available in the Sigma package. Also a pity they use their own proprietary mount, it would have made sense to provide the body in Nikon or Canon mount too.

it is K-mount (unfortunately the camera has IR filter in front of the mirror) w/ reverse engineered Canon electronic interface... so people have Canon lenses running on Sigma bodies w/ some modifications.

[deejjjaaaa]

The lack of Nikon mount is a show stopper for me

there is a kit to allow Nikon lenses mounted... no AF of course - http://www.sigmacumlaude.com/ - look for their kits on ebay

[JLK]

It makes nearly no sense to bring Sigma cameras out in anything but a Sigma mount. It's the razor/razorblade theory, and they make a good deal off of lenses. Assuming that this camera is up to snuff, it will demand some very good EX glass. And handholding will be difficult with this pixel pitch (I noticed an increase in motion blur moving from an SD10 to the SD14/15---and this camera is a 50% increase in the linear resolution). I've noticed that Sigma has really increased their OS lens lineup too. And the 8-16 mm wideangle is a stellar performer.

With regards to comparison, I'd love to see an end to it. I print 12 x 18" from the SD14/15 (and SD10 frankly), and have never had an issue with 'not enough' pixels. My pals with their APS-C Nikons and Canons love the details and sharpness of these. This camera will be like an 'apple to orange' comparison to the current crop of APS-C sensored cameras. It ought to be much, much better. If you have good glass on it. If you have it steady... etc.

It's really great to see Sigma make such a leap only two years (a little less) after acquiring Foveon. And if the new sensor is improved with regards to noise and exposure behavior, this'll be a very nice camera to own (price is rumored to be 'comparable' to a 7D, according to the US manager of Sigma). If they sell enough of these, I'd expect that we'd see a similar effort in a full frame or 1.2 crop camera down the line. And if they can jam this thing into their compact body (along with the dual True II processors), that'd be something to have in the pocket.

[welder]

Spin indeed. But, marketing-wise, they are in a bit of a dilema with the sensor. While the output size is 1/3 the advertised sensor size, a Foveon sensor and no AA filter does produce results that uprez nicely. It's been a few years but I think I recall seeing that 6mp output from a Foveon sensor could be scaled up to match the output of a 10mp bayer sensor (at least at low ISOs). (I did a couple of my own tests with files downloaded from the web and found that was a reasonable approximation.) If that still holds true, this new 15mp Foveon sensor could reasonably be exptected to upsize to the equivalent of 24mp-ish files.

Here we go again, 4800 x 3200 pixels is 15.4MP output. Sure, they are based on co-located RGB-ish sensels, but it is silly marketing spin to suggest R/G/B are separate output pixels.

I'm glad the technology is still alive, it'll keep the other companies alert, but I could do without the spin. Next thing they'll claim higher resolution than mainstream sensor designs, while leaving out the fact that they don't use an AA-filter, making it an apples/oranges type of comparison.

Cheers,
Bart

[Ben Rubinstein]

Is it still a 1.7X crop?

[Bart_van_der_Wolf]

Is it still a 1.7X crop?

Hi Ben,

With a 24×16mm APS-C sensor it's a 1.5x crop versus full 36x24mm sensor.

Cheers,
Bart

[feppe]

Here we go again, 4800 x 3200 pixels is 15.4MP output. Sure, they are based on co-located RGB-ish sensels, but it is silly marketing spin to suggest R/G/B are separate output pixels.

I'm glad the technology is still alive, it'll keep the other companies alert, but I could do without the spin. Next thing they'll claim higher resolution than mainstream sensor designs, while leaving out the fact that they don't use an AA-filter, making it an apples/oranges type of comparison.

It's certainly marketing speak, but so is equaling a 15 megapixel Bayer Array to 15 effective megapixels - it's just a convention we're used to. To make an equivalent comparison between BA and Foveon sensors, the effective MP count of the Sigma is somewhere between 15 and 46. I'll leave that for those more technically knowledgeable

In any case, this camera looks stellar on paper. MP count on-par with Bayer Array cameras, foveon sensor with arguably sharper and higher IQ at effective megapixels (caveat above, pending actual images, etc.), magnesium body, weather sealed, cross AF sensors. No video and Foveon hasn't been good at ISO as far as I know. We'll see if it delivers as a camera and real-life IQ, and how long it takes for Sigma to release it - they've taken literally years in the past between announcement and release.

This would make for a compelling product for me when I'm upgrading from 550D in a year or two - if it wasn't for the Canon lens and flash investments I have.

[JeffKohn]

What I remember from reviews is that a Foveon sensor would give you resolution equivalent to a Bayer sensor with about 1.75x the megapixels, not 3x. I recall the 3.4mp sensor being on par with 6mp Bayer sensors in reviews. Which is still impressive, since that puts the resolution of this camera above the full-frame models from Canon and Nikon.

But resolution is only one aspect of sensor performance, and from what I recall past Foveon sensors have not exactly excelled in other areas such as dynamic range, color rendition, high-ISO performance, and long exposures. Then there's the lack of live-view, low-res LCD, etc. It will be interesting to see what price-point they target. Based on the limited specs released so far, it looks like a D90-class camera body (excepting the sensor).

[bradleygibson]

It's certainly marketing speak, but so is equaling a 15 megapixel Bayer Array to 15 effective megapixels

They're not quite the same.  Correctly specified, a D3X has 24 million sensels, which it uses to deliver a 24 million pixel image file (by using interpolation).  But the SD1 doesn't use its 46 million sensels to deliver a 46 million pixel image file--that is the expectation that we are used to (rightly or wrongly) that the SD1 is flouting.

[BJL]

Is it still a 1.7X crop?

As noted it is now "full APS-C sensor size" ... finally fixing the embarrassment that Sigma's lenses have till now suffered a wide-angle FOV crop on their own DSLRs compared to the sme lenses on other "APS-C" DSLR's. And marketing hype aside, previous experience suggests that this camera should have better resolution than any DSLR in up to 35mm format: maybe about "30MP Bayer equivalent."

So if your dominant priority is high resolution, and you mostly care about good light situations (I expect the high ISO performance to continue to be worse than all competing sensors) and you are not bothered by the color accuracy problems that some have complained of, and which seem inherent to the Foveon approach, it could be an interesting option. I have become skeptical about the Foveon's execution of X3 technology after initial enthusiasm for the basic concept, but this new sensor looks like a worthy comeback effort.

P. S. My subject line is my preferred way of ducking the pixel counting debate.

[DarkPenguin]

30mp should be right as most of the bayer resolution comes from the green pixels which comprise half of the bayer array.

What's the word on the new body?  Sigma bodies were garbage.  This one looks like it might be okay.

As noted it is now "full APS-C sensor size" ... finally fixing the embarrassment that Sigma's lenses have till now suffered a wide-angle FOV crop on their own DSLRs compared to the sme lenses on other "APS-C" DSLR's. And marketing hype aside, previous experience suggests that this camera should have better resolution than any DSLR in up to 35mm format: maybe about "30MP Bayer equivalent."

So if your dominant priority is high resolution, and you mostly care about good light situations (I expect the high ISO performance to continue to be worse than all competing sensors) and you are not bothered by the color accuracy problems that some have complained of, and which seem inherent to the Foveon approach, it could be an interesting option. I have become skeptical about the Foveon's execution of X3 technology after initial enthusiasm for the basic concept, but this new sensor looks like a worthy comeback effort.

P. S. My subject line is my preferred way of ducking the pixel counting debate.

[dkoons]

  I ran across a site claiming it was to out in Feb for 1500 pounds, what ever $$ that will be in the states  I don't know, but,it made me pause

[feppe]

  I ran across a site claiming it was to out in Feb for 1500 pounds, what ever $$ that will be in the states  I don't know, but,it made me pause

Yep, that'd be a pretty good price if it delivers on the specs.

[dkoons]

  It will take that time to save the money and watch for the reviews

[Eruditass]

Here we go again, 4800 x 3200 pixels is 15.4MP output. Sure, they are based on co-located RGB-ish sensels, but it is silly marketing spin to suggest R/G/B are separate output pixels.

I'm glad the technology is still alive, it'll keep the other companies alert, but I could do without the spin. Next thing they'll claim higher resolution than mainstream sensor designs, while leaving out the fact that they don't use an AA-filter, making it an apples/oranges type of comparison.

Cheers,
Bart

Yeah, and bayers really have 4x the "pixels", heh!

Some claim the previous 4MP(?) sensor was equivalent to 10MP in bayer w.r.t. resolution and quality.  Since you don't really want to advertise 4MP, and 10MP is arbitrary, I guess that's why they chose to do the physical 3x.  However, I think the only people that will buy this camera already know that and marketing is somewhat pointless.  I suppose it will get people to investigate it that aren't familiar with it.

[Rob C]

"This would make for a compelling product for me when I'm upgrading from 550D in a year or two - if it wasn't for the Canon lens and flash investments I have. "


Feppe, isn't that a somewhat optimistic view on the purchase of any photographic equipment nowadays?

It's raining here, so I see everything much more clearly now than in high summer.

Rob C