Go Back   Steve's Digicams Forums > Digital SLR and Interchangeable Lens Cameras > Sigma dSLR

Reply
 
Thread Tools Search this Thread
Old Mar 22, 2005, 4:17 PM   #11
Member
 
Join Date: Aug 2004
Posts: 34
Default

Thanks for sharing, but I feel the big question isn't about if Sigma is going to launch a new SLR
or two, but rather if Foveon is going to market a bigger X3 sensor in our lifetime, or not.
code is offline   Reply With Quote
Old Apr 1, 2005, 7:23 AM   #12
Senior Member
 
Join Date: Jan 2004
Posts: 824
Default

code wrote:
Quote:
Thanks for sharing, but I feel the big question isn't about if Sigma is going to launch a new SLR
or two, but rather if Foveon is going to market a bigger X3 sensor in our lifetime, or not.
Well at least in the realm of April Fool's fantasy:
http://www.photokb.com/Uwe/Forum.asp...sal-Lens-Mount
Norm in Fujino is offline   Reply With Quote
Old May 12, 2005, 6:50 PM   #13
Junior Member
 
Join Date: May 2005
Posts: 12
Default

I like the SD30 (-;

Seriously though, if you don't plan to print larger than 6x8 (300dpi), then the Sigma SD9-10 is all you'll ever need and will provide the best color detail of any other camera in it's class. That's why I bought the SD9. I got the flash, 70mm-300mm, 28mm-70mm (with case), and two UV filters for $900 on ebay. For the occasional poster printing, many free methods for interpolation are available at:

http://www.americaswonderlands.com/d...erpolation.htm

The Irfanview software has many interpolation methods built in as well and I think the optional plugins option has more.

www.jason-page.com
jasonpage is offline   Reply With Quote
Old May 26, 2005, 1:43 AM   #14
Moderator
 
Join Date: Jun 2002
Posts: 1,139
Default

Quote:
Seriously though, if you don't plan to print larger than 6x8 (300dpi), then the Sigma SD9-10 is all you'll ever need and will provide the best color detail of any other camera in it's class. That's why I bought the SD9. I got the flash, 70mm-300mm, 28mm-70mm (with case), and two UV filters for $900 on ebay. For the occasional poster printing, many free methods for interpolation are available at:
I'm guessing that you haven't been to a PMA show lately. If you had, you would have seen stunning, gigantic poster sized prints from SD9 and SD10's. So far, to be fair, I've only seen the quality of large print improved on by the new Nikon D2X at 12 megapixels and even that was close.

Don't let the smaller native file size of 3.4 megapixels fool you into thinking that the X3 images are anywhere close to 3.4 megapixel quality. The optical resolution as measured by normal black and white resolution charts is slightly better than most six megapixel bayer processed images and the color resolution is considerably better than a six megapixel bayer processed image. In addition to the excellent resolution, the pixel level sharpness and detail are outstanding.

The optical resolution, not digital pixel count is what is important in so far as interpolation and enlargeability are concerned. All digital captures depend on interpolation to render any print size other than the one single size possible from a given print density (dpi) and a native file matrix. As most of you probably know, that size is obtained by dividing the horizontal and vertical pixel count by the desired print density. Obviously we don't print at a single size so we must either remove or add pixels to get to the desired print size.

Unlike film which is enlargement limited primarily by grain, digital is enlargement limited primarily by resolution. This is because noise levels in quality digital sensors are much lower than grain equivalencies in film. Film enlargement simply runs up against the grain limitation before resolution becomes a factor. With digital it's quite possible to enlarge far beyond what can be done with an equivalent film platform simply because noise isn't a limiting factor.

Because it is "possible" to make clean enlargement of gigantic proportions, we eventually run up against resolution barriers. Incompletely defined detail happens because sampling sites (sensor photosites or sampling "pixels") at some point are insufficient in density to properly define the boundaries of detail. These improperly defined detail objects are sometimes referred to as "marker pixels". Think of them as you would brush strokes in an oil painting representing detail such as pine needles on distant trees in a forest. A normal viewing distances with our naked eye our brains interpret these brush strokes on canvas equivalents of "marker pixels" as the detail we "expect" to see. However when we look at them up close under magnification, the illusion of detail is shattered and we are forced to realize that they are simply brush strokes in oil. Take away the magnification and step back and we can again accept them as good representations of "pine needles".

When interpolation algorithms encounter these poorly defined detail objects ("marker pixels") they faithfully reproduce them and when we enlarge a photo at some point these marker pixels are brought to the threshold of human visual acuity. At this point it's like looking at the oil painting under magnification. The illusion of detail we expect is revealed to be only marker pixels.

This gets us to the point of realizing that the wider the field of view and deeper the depth of focus or depth of field, the more detail there will be in a capture frame to resolve. So a wide angle hyperfocal landscape capture will "require" much more resolution (more sampling sites and/or better capture signal) to define than a smaller field of view such as we might find in a head and shoulders portrait. So it then follows that a head and shoulders portrait takes much less capture resolution to make a large print than a wide angle, hyperfocal landscape.

The essence then is that with digital we must consider the nature of the subject, the field of view (actual geography in the frame) and desired enlargement size when determining the sensor requirement for a given capture. It's not as simple as with film because we "can" enlarge to a much greater degree and still maintain clean and smooth, noise free prints.

Because the Foveon X3 processor used in the Sigma SD9/SD10 uses 10.3 million sampling sites to create 3.4 million display pixels, it actually gets much better pixel level information than an equivalent 3.4 million bayer camera (anything except a Sigma or a camera with a Foveon processor right now). This means that the normal process of divide the horizontal by vertical pixel count by desired print density formula we use is not adequate for determining enlargement potential. The Sigma has "better" and more accurate pixels than an equivalent "megapixel" bayer camera which is quite evident by the "optical resolution" as measured by resolution chart photos. There are three sampling sites contributing to each display pixel rather than one so that the accuracy is improved. Also there is no antialiasing filter (blur filter) which further contributes to the sharpness and enlargeability.

In short, the Sigma cameras can be used to create much larger prints than would normally be expected.

Lin

Lin Evans is offline   Reply With Quote
Old Jun 3, 2005, 1:08 AM   #15
Junior Member
 
Join Date: May 2005
Posts: 12
Default

What interpolation algorythm do you suggest? I heard that Sigma has software for the SD10 to make 10.3mp images, but what could I use for the SD9 that would be comparable?

I would like to interpolate from the RAW because that's where all the three layer information is intact. Doing it from a flattened jpg is just like doing the same with any other 3.34mp jpg. The other other softwares I know that have support for X3 is Photoshop CS1&2 and the newest version of IrfanView. I have photoshop 7 and IrfanView. IrfanView might be my only option, but I do not know if it Interpolates it as a flat image or if it uses the three layers of RGB information.

-----
http://www.pbase.com/jasonpage
jasonpage is offline   Reply With Quote
Old Jun 4, 2005, 10:13 AM   #16
Member
 
Join Date: Dec 2004
Posts: 67
Default

Jason,

almost any interpolationsoftware should be able to process X3 pictures if you convert them from X3-RAW to another lossless file format (e.g. TIFF). The three layers of X3 wouldn't get lost butwould get converted into another format without degradation in quality.

The 10.3 MP (8 bit)Sigma and Foveon claim are3.3 MP (24 bit), there's no way around that fact.The different ways of storing them (3 layers x 3.3 MP x 8 Bit or 1 flat layerx 3.3 MP x 24 Bit) neitheraffects quality norfile size. At leastif you stick to a lossless format like RAW, TIFF or PNG.

The thing that makes the Foveon chip so special is not the separate storage of the three layers but the way the rays of light are being captured. Simplifying, one could say that the Foveon chip captures each ray's colors separately but at the same point on the sensorwhereas conventional Bayer sensors capture each ray's colors separately but at slightly different points on the sensor which can result in color artifacts.

Thus, if you've already captured your beautiful Sigma X3 picture and if you've saved it as RAW you just need to convertit to a lossless file format that your interpolation software can read. The best interpolation software for blowing up your pictures is said to be "Genuine Fractals" but I've to admit I've never tried it by myself. It is available as an Add-In for Photoshop which can read Sigma RAW, too.

Interestingly, the absence of color artifacts in Foveon pictures makes them predestined for enlarging, in many cases a 3.3 MP Foveon picture inflatedto 6 MP has a better quality then a 6 MP picture from a 6 MP camera,I've seen it by myself.

Personally, I'd like to buy a SLR with a Foveon chip by myself. The only thing is that I'm not so sure if the company will be around much longer, they haven't produced anything real new for almost 2 years, only variations of their X3 3.3 MP. So I'm desperately waiting for news to show me which direction Sigma and Foveon will go.

Regards,

Fisheye


Fisheye is offline   Reply With Quote
Old Jun 4, 2005, 4:27 PM   #17
Junior Member
 
Join Date: May 2005
Posts: 24
Default

As an amateur, Idon't want to useCCD or CMOS (I hate soft pictures) after I saw SigmaSD10 picture samples.

However Sigma has to produce a new camera with;

1- More resolution

2- Better noise reduction at high ISO

3- JPEG output as an option

4- Included Flash
Aktunav is offline   Reply With Quote
Old Jun 6, 2005, 4:06 AM   #18
Junior Member
 
Join Date: May 2005
Posts: 12
Default

Aktunav wrote:
Quote:
As an amateur, I don't want to use CCD or CMOS (I hate soft pictures) after I saw Sigma SD10 picture samples.

However Sigma has to produce a new camera with;

1- More resolution
Perhaps, but you usually can't get better quality from film or digital unless you're using an expensive Zeiss lense wit h fine grain film.

Quote:
2- Better noise reduction at high ISO
What is ISO mean in the digital sense?

Quote:
3- JPEG output as an option
That's what the Sigma Pro software is for.

Quote:
4- Included Flash
Would make the unit more expensive. It would be redundant and a builtin flash offers little control. Built in flashes flash directly at the person. External flashes are designed to lightup the area sourunding the person.
[/quote]
jasonpage is offline   Reply With Quote
Old Jun 6, 2005, 4:08 AM   #19
Junior Member
 
Join Date: May 2005
Posts: 12
Default

Thanks Lin Evans and Fisheye for explainations.
jasonpage is offline   Reply With Quote
Old Jun 7, 2005, 6:29 AM   #20
Junior Member
 
Join Date: May 2005
Posts: 24
Default

Thank you for yourreply jasonpage, good to knowbutI'd like to wait anew versionSigma DSLR anyway.
Aktunav is offline   Reply With Quote
 
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -5. The time now is 7:48 PM.