comparisons of a Contax ND Raw file and a Sigma Foveon X3 file

[thril]

Anyone here seen any side by side comparisons of a raw ND file and a Foveon X3 file from the Sigma camera? I'm hearing that the little 3MP Foveon chip is blowing away the 6MP cameras on the market in terms of resolution...

will

 

[afranklin]

Will,

> I'm hearing that the little 3MP > Foveon chip is blowing away the 6MP cameras on the market in terms of > resolution...

Where are you "hearing" this? Specifically. Do you have a URL or a magazine reference?

I've done a bit of evaluating of the 3M Foveon sensor, and no, it does not "blow away" any of the 6M SLR cameras in terms of resolution...so I can't imagine where you get this from.

Austin

 

[garyh]

John,

Could be (midroll rewind). Have never used an APS camera so I don't know anything about them (other than APS itself has been less than a stellar success).

Gary

 

[thril]

>Hey Austin, I've got a friend in Houston and another in NYC who've just gotten the cameras and been doing some testing with them. They also referred me to the following url which also shows images and tests proclaiming the Foveon's superior resolution...

Please, Log in or Register to view URLs content!

Granted my "blown away" statement is a bit much, but still it's pretty impressive for a little 3MP chip...

Re:> Where are you "hearing" this? Specifically. Do you have a URL or a > magazine reference? > > I've done a bit of evaluating of the 3M Foveon sensor, and no, it does > not "blow away" any of the 6M SLR cameras in terms of resolution...so > I can't imagine where you get this from. > > Austin

 

[afranklin]

> Posted by Will Taylor on Monday, May 12, 2003 - 11:55 pm: > > >Hey Austin, I've got a friend in Houston and another in NYC who've > just gotten the cameras and been doing some testing with them. They > also referred me to the following url which also shows images and > tests proclaiming the Foveon's superior resolution... >

Please, Log in or Register to view URLs content!

Granted my "blown away" > statement is a bit much, but still it's pretty impressive for a little > 3MP chip... > > Re:> Where are you "hearing" this? Specifically. Do you have a URL or > a > magazine reference? > > I've done a bit of evaluating of the 3M > Foveon sensor, and no, it does > not "blow away" any of the 6M SLR > cameras in terms of resolution...so > I can't imagine where you get > this from. > > Austin

I've read the DPReview stuff. What, exactly, there are you referring to?

Keep one thing in mind when looking at digital images. Sharpness can be attained by processing any image, and sharpness doesn't mean reality either. We don't know what processing the SD9 is doing, so it's quite hard to tell really make a claim of anything without actually knowing what processing (sharpening or whatever) was done with what images (both in-camera and out of camera, and don't be fooled into believing raw images aren't processed, they very may well be). There are aftermarket Bayer pattern processing programs that do a better job than the in-camera processing.

What you are comparing is the results from the camera, not from the sensors, so making any claims based on the sensors is erroneous, at best.

In looking at the resolution chart tests, I'm not seeing it "blow" the D-60 away at all. Can you point me to the two images that show it even to be better? In fact, I find the D60 image:

Please, Log in or Register to view URLs content!

to show the D60 has better resolution than the SD9:

Please, Log in or Register to view URLs content!

9.jpg

Keep in mind, that lenses are important as well...and though I'm sure they mention it somewhere, I didn't see what lense they used for the D-60.

The Foveon sensor suffers from some serious artifacting anyway, which, for the most part, render it useless in low light, and bright light situations, as well has horrid color fidelity.

Austin

 

[djg]

Will,

You need to be wary of these comparisons in DPReview. For ex&le, in the Sigma vs, EOS-D60 detail comparisons, the EOS image is very visibly larger than the Sigma image, as it ought to be as it's twice the megapixels. If viewed at their relative natural size, as is obviously done here, that in and on itself, all else equal, will make the Sigma image appear sharper being smaller.

Also look at the JPEG sizes. At twice the megapixels, the EOS image is about the same file size as the Sigma. Can you say "JPEG compression factor difference" 10 times real fast?

Now I'm not implying these are deliberate, only that you need to take into consideration all the aspects of the apples you're comparing ...

DJ

 

[loom]

Dear all,

I believe we ought to be very careful about the digicam reviews on the web. I bought Canon S30 last year to shoot underwater photos (I bet there will never be a underwater housing for Contax^^). Though the photos were nice, I cannot stand its images above sea level. The images simply look unreal.

However, almost everyone, or at least every single online digital photography reviewing site praises that kind of image. Every single camera Canon produced, from low end to high end, received the finest marks on image quality. I talked this issue quite a bit to a friend of mine, who has a PhD in computer sciences, has 15+ years of photography and also a fan of Zeiss. Here is what he said:

"I do not care about whatever Steve's website says. Most of his tests were not done right and flawed. DP review's tests are much better. The common problem among those websites is that they all do not have enough photography experiences, especially with films. They only talk about digital photos in the digital way. They review digicams much like the way other people test computers. What would you expect from this kind of reviews?

The real issue here is that digicam designs are very complex. There are lots of component, signal processing, and image processing issues behind them. You need to have a deep understanding about all of the design trade-offs and component selections to do a test right. A good test should give readers more education instead of reporting superficial numbers. Otherwise it will be like a computer test which blantly tells you a 2GHz P4 CPU is faster than a 1.3GHz P3 CPU simply because the clock is faster!

For instance, to measure the digicam's resolution, a true test should be done at least with a MTF test. The line chart thing can give you false reading because the image processing brain in a digicam has been tuned to recognize the line chart to give you a good number. If the input is a sine wave then you will see how poorly the digitcam will perform."

After seeing the s&le images he bought a TVS-digital rather than the Canon S50.

By the way, you're all more than welcome to my on-line album,

Please, Log in or Register to view URLs content!

. Most of the photos are shot with G2.

Best,

Shu-Hsien

 

[fotografz]

The tests between the Sigma 9 and Canon D-60 at all mute points. The D60 was replaced by the D10 which at $1,500. is $700 less than the D-60 and $300. less than the Sigma.

A photojournalist friend brought his D-60 to my studio to compare it to my new D-10. The D-10
images were clearly superior to those of the D-60 using the same settings, same L lens and same strobe lighting. Canon said it was, and they weren't lying. My friend traded his D-60 for a D-10 the next day.

When I say we both saw the difference, it was clearly superior in every way. I've now seen D-60s selling for $800-$1,000. Sigma was a day late and a dollar short....all wrapped in a plastic body, where the new D-10 is metal.

 

[irakly]

I saw a 11x14" print from Sigma SD9, and nobody in the world will convince me that it looks like it should. It was screaming "DIGITAL", even though it was made by a pretty skillful professional photographer. Although I do not dismiss a possibility that someone is looking just for that

 

[thril]

Hey Austin, Re: your comments, > I've read the Austin stuff. What, exactly, there are you referring > to? > > Keep one thing in mind when looking at digital images. Sharpness can > be attained by processing any image, and sharpness doesn't mean > reality either. We don't know what processing the SD9 is doing, so > it's quite hard to tell really make a claim of anything without > actually knowing what processing (sharpening or whatever) was done > with what images (both in-camera and out of camera, and don't be > fooled into believing raw images aren't processed, they very may well > be). There are aftermarket Bayer pattern processing programs that do a > better job than the in-camera processing. > > What you are comparing is the results from the camera, not from the > sensors, so making any claims based on the sensors is erroneous, at > best.

I feel your comments about some of the images from DPReview showing the D60 actually beating out, or at least being comparable to, the X3. That's why I came here. There are so many factors to consider, as you say. My impressions from their review are that there was no post processing sharpening done to either of the test images from either camera, so I was impressed that such a small chip was getting as close it did to a sensor double the size. I guess the main reason I'm asking is I'm still on the fence about which digi setup to buy (if I can even find an ND these days...). The lower price point of the Sigma is attractive, but like you said, there are the lenses to consider, and since I've owned and shot with the C 645 and the N1 for years now and know about their insane color fidelity and sharpness, I'm very leery of jumping into another system such as Sigma...yuck! But money is an issue... Also, as you say, the halo problem around the highlights on the Sigma is definitely a problem. Still the Foveon technology is looking promising. I guess we'll just have to wait and see what happens with them... You mentioned in an earlier post that you've done some testing yourself with the Sigma and I imagine you have the ND. Do you have any definitive results that you've come up with one way or the other? And one more question...I'm wondering if you (or anyone else out there) have ever tried to blow a shot from the ND up to a full 13x19 Epson print size? Or what's the largest you feel comfortable printing from the ND? Thanks a million for all your input...

will

 

[thril]

Thanks for the reply DJ,

You said, > You need to be wary of these comparisons in DPReview. For ex&le, in > the Sigma vs, EOS-D60 detail comparisons, the EOS image is very > visibly larger than the Sigma image, as it ought to be as it's twice > the megapixels. If viewed at their relative natural size, as is > obviously done here, that in and on itself, all else equal, will make > the Sigma image appear sharper being smaller. The only test they did to try and equalize this issue was to use a tool in the Foveon processing software that doubles the image size through interpolation. It then yields an image which is just a bit larger than the D60's, so they shrunk it down to the same dimensions, for their final comparison. That variable alone left me wondering what's going on in the background there in Foveon's software when they blow it up to that size. Who knows what sort of algorithms they've used to boost their image sizes, and it was my guess that some sort of sharpening was involved. Hence my questions to this forum. It's always better to get info, IMO, from the professionals that are out using the gear with practical applications. I hadn't even considered the JPEG compression factor issue; good point! Thanks!

will

 

[andybennetts]

Will - the biggest print - done by a company in England for a show - from one of my own ND images, was 24x36 inches with no quality problems(shot at 100 iso in RAW).
I do a lot of prints at A3+ size on my Canon S9000 and the again quality is excellent. When you see a print from an ND, look particularly at the dynamic range which in my view is better than the other full frame 35mm slr digicams.

Andy

 

[fotografz]

Will, as I said in the post above, the D-60 is history. There are some left on dealer shelves, but who would buy one at $2,200 when the superior D-10 costs $1.500.? If money is an issue the D-10 is the answer.

 

[witwald]

> Witold, > which web site is it? > Christian

Hi Christian,

The review of the Contax TVS Digital was found at:

Please, Log in or Register to view URLs content!

The following page has the Canon S50 review:

Please, Log in or Register to view URLs content!

I hope that you find them interesting to read through.

-- Witold

 

[fastfashn]

Long exposures and digital noise...

Astrophotographers keep their CCD quantum noise down by cooling the CCDs. This would be difficult to do in a digicam. Because the noise is unpredictable it would be nearly impossible to write software to filter it out, so I believe there must be a true maximum time exposure limit for any uncooled digicam.

This also means that the hotter the electronics in a digicam run, or the warmer the day, the noisier the CCD.

Has anyone experimented with a cooled digicam vs a room temp digicam to see if there is any difference in the noise level of an exposure? I think this would be a REALLY kewl experiment.

Take a digicam down to a decently low temp, say with a dry ice pack and cooler, then shoot quickly at the max ISO of the camera in a night scene vs that same digicam at a 45C temp?

 

[djg]

The problem with cooling is that unless the whole system is designed properly you're going to have condensation problems. Interesting proposition. How much can you cool before condensation is a problem? Retorical question, of course

.

DJ

 

[rico]

Dana,

Astronomers also record successive images, and stack them. This is easy to do in a digicam and the improvement in dynamic range is theoretically unlimited. I have performed quack medical experiments with Canon RAW files on terrestrial subjects to great effect. Note that JPG cannot be used because the low-order data, starting with chroma, is wiped out at even modest compression.

For those with money and physical access to the CCD, active cooling will reduce the frames needed in a stack (to one, subject allowing).

 

[fastfashn]

I didn't know jpeg killed the chroma info off so quickly... Not every digicam will let you do RAW files, ummm... What do you think of TIFF? Is it TIFF or TGA files that digicams do? Can you define stack please?

Ok, so... Since we can't easily cool the CCD below ambient in a portable digicam and maintain a good battery life (every way I know to cool the CCD will eat batteries), do you see a limit in timed exposures that cannot be overcome, or a theoretical maximum in ISO?

Dana

"Posted by Rico Tudor on Monday, May 19, 2003 - 11:29 pm:

Dana,

Astronomers also record successive images, and stack them. This is easy to do in a digicam and the improvement in dynamic range is theoretically unlimited. I have performed quack medical experiments with Canon RAW files on terrestrial subjects to great effect. Note that JPG cannot be used because the low-order data, starting with chroma, is wiped out at even modest compression.

For those with money and physical access to the CCD, active cooling will reduce the frames needed in a stack (to one, subject allowing)."

 

[writing4me]

If you are new-ish to digital cameras, this is a nice article that explains some of the basics.

Please, Log in or Register to view URLs content!

Quoting the beginning of the article:

"So many people have been asking which digital camera to buy, how many "megapixels" do I need, and 'What's a megapixel?' You would need to read an entire book to get the real, in-depth answers to these questions, and after that you'd still be in a quandary about which camera to buy. I'm going to share some quick, simplified, rules of thumb for making your own decisions"

The article was written by: Fred Showker is Editor in Chief of "MUG" the AACUG Mac User Guide, and the User Group Network News service, and a founding Apple User Group Advisory Board (UGAB) member.

Hope this is helpful.
-Lynn L.