Sigma DP1 for those considering a DSLR

[Groucho]

Umm, have you checked to see how a normal bayer array sensor works? Each of the 6 megapixels captured by a K100D (or any other 6 megapixel camera) is also "black and white" (monochromatic, actually). So what do you want a mere 6 million photo-receptors or 14 million photo receptors? Do you want 6 million different monochromatic pixels or a mere 4.7 million full color pixels? There is just no way to make an apples to apples comparison.

You're preaching to the choir here. I'm not debating the potential superiority over a Bayer sensor.

However, like I said, I think you need to measure the total number of pixels that come out of the camera when all is said and done. The Sigma makes a 4.7mp image. A 6mp Bayer-sensor DSLR makes a 6mp picture. Both have their own "trickery" to produce the final result.

 

[MarkBarbieri]

However, like I said, I think you need to measure the total number of pixels that come out of the camera when all is said and done.

So if the Sigma did an in-camera interpolation to build a 14 megapixel image, would that make it a "14 megapixel" camera?

 

[Anewman]

So if the Sigma did an in-camera interpolation to build a 14 megapixel image, what would make it a "14 megapixel" camera?

It is a start

 

[webshark3]

I'll try to find the web article for the water bag....

 

[YEKCIM]

I've read this and re-read this many times trying to visualize it and correct me if I'm wrong, but this will not work.

If the camera and lens are in the bag and you screw the filter over the lens from the outside of the bag, then the part that you want to cut away is now between the lens and the filter. You'd have to cut away the glass on the filter to get to the plastic that needs to be cut away.

Now I do think that it will work if you attach the lens hood from the outside of the bag as the lens hood has no glass to cover over the part of the bag you want to cut away. Then you can attatch the filter to hold the bag to the lens. Unless of course the filter doesn't have glass on it, then its just a plain old ring that can screw onto a lens.

How about using a step ring instead of a filter?

~YEKCIM

 

[webshark3]

So my directions were a little off. it's been years.

Here's the links:
http://gizmodo.com/gadgets/digital-...eap-splashproofing-for-your-camera-189371.php

Step by Step:
http://www.instructables.com/id/ERZGHB60KVEP286425/

 

[Groucho]

So if the Sigma did an in-camera interpolation to build a 14 megapixel image, would that make it a "14 megapixel" camera?

Certainly, though not a very good one. There goes most of the extra sharpness that you gain from their sensor design. I believe this is sort of what the Fuji S3 does to get a 12mp image out of two 6mp sensors.

The key difference here is that each group of 3 of the Sigma's sensors is recording the same pixel. It records one third of a full pixel.

Yes, the Bayer filters are monochromatic as well, but each sensor is a unique pixel. The color is then estimated back - but you're still recording six million unique pixels.

 

[webshark3]

Not bad for a 3.4mp x3 sensor

http://www.dpreview.com/reviews/sigmasd10/page17.asp

"What these three comparisons have proven is that the SD10 image [3mp] is (for all intents and purposes) at least as detailed and containing as much resolution as a six megapixel image from a Bayer pattern sensor. That said it can also be argued that you can produce an 'X3 like' image from a six megapixel camera by simply downsampling the image."

Certainly, though not a very good one. There goes most of the extra sharpness that you gain from their sensor design. I believe this is sort of what the Fuji S3 does to get a 12mp image out of two 6mp sensors.

The key difference here is that each group of 3 of the Sigma's sensors is recording the same pixel. It records one third of a full pixel.

Yes, the Bayer filters are monochromatic as well, but each sensor is a unique pixel. The color is then estimated back - but you're still recording six million unique pixels.

 

[Anewman]

Not bad for a 3.4mp x3 sensor

http://www.dpreview.com/reviews/sigmasd10/page17.asp

"What these three comparisons have proven is that the SD10 image [3mp] is (for all intents and purposes) at least as detailed and containing as much resolution as a six megapixel image from a Bayer pattern sensor. That said it can also be argued that you can produce an 'X3 like' image from a six megapixel camera by simply downsampling the image."

So what that proves is that a 10.2mp(3.4x3) Foveon sensor competes with a 6mp Bayer sensor.

Cool, now if we can find a comparison that is not from 2003...

 

[webshark3]

So what that proves is that a 10.2mp(3.4x3) Foveon sensor competes with a 6mp Bayer sensor.

Cool, now if we can find a comparison that is not from 2003...

From 2003 and just as valid today as it was then. Cameras don't stop functioning because newer models came out. I'll check my 300D, but I'm pretty sure it hasn't "expired" since the XT and XTi came out. lol

The SD14 comprehensive review is not out yet (will be soon). And what the old review shows is that a 3mp camera (as everyone has pointed out in the thread based on the final actual resolution) retains as much detail as a 6mp Bayer Sensor.

It is speculated and shown in side by side comparisons that the SD14 (~5mp if you want to view it that way ) will have the same resolving power as a 10-12mp Bayer Sensor.

It all comes down to throwing out the whole "my camera's better than yours because it has more pixels" debate. It doesn't hold up anymore than the GHz game that was played by Intel and AMD.

Let's say the Sigma SD10 has 10.2 million pixels which can hold up to the resolution of a 6mp Bayer Layer Sensor while maintaining full and accurate color. See why it's hard to do an apples to apples comparison? To give Canon justice, their SLRs produce incredible color IMO.

 

[GrumpyOne]

Alright, I'm confused. Isn't this essentially the same as multiple exposures?

 

[Anewman]

From 2003 and just as valid today as it was then. Cameras don't stop functioning because newer models came out. I'll check my 300D, but I'm pretty sure it hasn't "expired" since the XT and XTi came out. lol

Fact is that sensors have improved over the last 4 years, sorry if I offended your 300D by stating the truth.

It is speculated and shown in side by side comparisons that the SD14 (~5mp if you want to view it that way ) will have the same resolving power as a 10-12mp Bayer Sensor.

Yes is it speculated, that is why I posted my desire for a CURRENT COMPARISON. Sounds like you agree with me that it would be nice to have one.

It all comes down to throwing out the whole "my camera's better than yours because it has more pixels" debate. It doesn't hold up anymore than the GHz game that was played by Intel and AMD.

but isnt that what you are doing by claiming 14mp???

If a SD10 is 10.2mp isn't it a disappointment that it only resolves as much detail as a 6mp Bayer sensor camera?
If this new camera is 14mp why do you keep making it sound like coming close to a 10mp bayer sensor camera is such a big accomplishment?

And that GHz war is still going strong even if it does not hold up in your opinion.

Let's say the Sigma SD10 has 10.2 million pixels which can hold up to the resolution of a 6mp Bayer Layer Sensor while maintaining full and accurate color. See why it's hard to do an apples to apples comparison? To give Canon justice, their SLRs produce incredible color IMO.

Again if it has 10.2 million pixels, why not compare its resolution to a D80 or 30D? isn't that apples to apples?

 

[webshark3]

Alright, I'm confused. Isn't this essentially the same as multiple exposures?

All light can be separated in 3 primary colors. Each of the 3 layers in the Foveon sensor picks up one of the 3 (red,green, blue). Each layer of the sensor "absorbs" the color it's responsible for.

In the Bayer Layer, It's a single Layer sensor, so each pixel is responsible for all three colors. There are no dedicated pixels.

So Foveon has less pixels in each layer, but more pixels in depth, and dedicated pixel for each color.

Bayer has more pixels for it's layer, but no depth, so each pixel has triple duty.

In Theory, the 3 layer should be a better design. In practice, it takes a lot of x3 pixels to do what a Bayer does (3.4mpx3 Foveon ~ 6mp Bayer).

Let me know how much more I've confused things.

 

[webshark3]

Fact is that sensors have improved over the last 4 years, sorry if I offended your 300D by stating the truth.

But lenses haven't. I'll take my 300D with my good lenses over a XTI with kit lens any day.

Yes is it speculated, that is why I posted my desire for a CURRENT COMPARISON. Sounds like you agree with me that it would be nice to have one.

YEP


but isnt that what you are doing by claiming 14mp???

If a SD10 is 10.2mp isn't it a disappointment that it only resolves as much detail as a 6mp Bayer sensor camera?
If this new camera is 14mp why do you keep making it sound like coming close to a 10mp bayer sensor camera is such a big accomplishment?

And that GHz war is still going strong even if it does not hold up in your opinion.

So you bought a new 5GHz+ Computer? Or was it a lower GHz dual core?



Again if it has 10.2 million pixels, why not compare its resolution to a D80 or 30D? isn't that apples to apples?

Can't. It's not apples to apples, but the results speak for themselves.

 

[Anewman]

Can't. It's not apples to apples, but the results speak for themselves.

Yes and the results that you posted say that a "10.2mp" Foveon sensor competes with a 6mp Bayer sensor.

 

[Groucho]

The SD14 comprehensive review is not out yet (will be soon). And what the old review shows is that a 3mp camera (as everyone has pointed out in the thread based on the final actual resolution) retains as much detail as a 6mp Bayer Sensor.

Actually, we're talking about a 4.7mp Foveon being similar in detail to a 6mp Bayer. 4.7 is a lot more than 3.

Speaking of Bayer, anyone feeling a need for aspirin?

It is speculated and shown in side by side comparisons that the SD14 (~5mp if you want to view it that way ) will have the same resolving power as a 10-12mp Bayer Sensor.

Like I mentioned early, you're only getting 4.7 megapixels of resolution. That's IT. Each sensor is recording different colors, not different actual image data, unlike the Bayer, in which a 6mp camera actually grabs 6mp worth of image data.

It all comes down to throwing out the whole "my camera's better than yours because it has more pixels" debate. It doesn't hold up anymore than the GHz game that was played by Intel and AMD.

To be fair, Intel was the one who played that game (AMD changed the rules), and I think most DSLR owner with a little bit of sense will realize that mp are certainly not the top criteria in a camera's desirability. Again, how many XTi owners wouldn't be happy going to a 30D despite losing 2mp?

For better of for worse, though, I fear that the Foveon sensor will go the way of three-strip Technicolor. It is likely that CCD and CMOS sensors will continue to improve and chip away at any advantages of the Foveon, and their advantages in cost, complexity, and other features (like noise control), to say nothing of ubiquity, will ultimately lead to Foveon becoming completely obsolete, remembered only by gray-haired photographers, chewing the cud in a retirement home somewhere many years from now.

 

[GrumpyOne]

Let me know how much more I've confused things.

I would but I'm afraid I'll get an explanation.

I think it's probably one of those things where I'll need to see more info in order to make sense out of it. It's not a big deal since I'm not getting a camera with that sensor, it would be kind of like buying a VHS and wanting to know how a Beta-Max worked. Interesting but not necessarily practical.

 

[webshark3]

Groucho,

We can continue the technical discussions over at dpreview.com. If you think I'm "supportive" of the Foveon concept, just go to the Sigma forum where they they it's the greatest thing since sliced bread. LOL

I really posted at disboards becuase of these facts (regrardless of resolution):
1) Many Disneyers want the most from there photos, and want their DSLRs.
2) Many Disneyers was a P&S camera to "just record the moment."
3) A third class of Disneyer (me) is looking for a P&S sized device that could produce that 20x30 masterpiece hanging on my livingroom wall. You just never know when the moment will arise.

The Sigma DP1 is the first option the industry has given us. You're probably right that the concept may be short-lived, but fortunately, even if Sigma folded overnight, my DP1 would still function. I wouldn't worry about it since i'm not buying into a SLR system. People are too hung up about cameras being obsolute once the next model comes out. I have to laugh as it's usually the same people who didn't know how photos could be any better with their cameras when they were new...

 

[webshark3]

No Problem. All you need to know is that the Sigma DP1 will be the first point-and-shoot compact camera on the market that uses the same sensor as the large (non pocket friendly) Digital SLRs. It will be quite interesting.

I would but I'm afraid I'll get an explanation.

I think it's probably one of those things where I'll need to see more info in order to make sense out of it. It's not a big deal since I'm not getting a camera with that sensor, it would be kind of like buying a VHS and wanting to know how a Beta-Max worked. Interesting but not necessarily practical.

 

[MarkBarbieri]

Like I mentioned early, you're only getting 4.7 megapixels of resolution. That's IT. Each sensor is recording different colors, not different actual image data, unlike the Bayer, in which a 6mp camera actually grabs 6mp worth of image data.

That depends. Let's say you're taking a picture of a beautiful red rose. Just about all of the data is in the red color channel. With the Foveon, you're getting your full resolution throughout the photo. With the bayer camera, you're only getting 1/4 of your resolution in the red areas because the green and blue channels aren't seeing anything. So the resolving power of a bayer array sensor varies depending on the color composition of the subject.

[With a bayer array] you're still recording six million unique pixels.

Even this statement isn't as simple as it sounds. At the A/D level, it's true that you are recording data from six million different positions on the X/Y grid of the sensor. However, you have to have an anti-aliasing filter (also known as a blur filter) in front of the sensor. The filter blurs the image so that the samed focused points of light are spread across the different colored sensors in the bayer array. So the effective resolution even of a gray image is still reduced below nominal resolution.

Once again, it is not very meaningful to compare the megapixel numbers of full color and bayer array sensors. I would assume that when people talk about pixel numbers, they are really trying to determine resolving power. Just like with the analogy of computer speed and CPU clock rate, the relationship between sensor pixel or photosite counts and photo resolution is not straightforward.