Go Back   Steve's Digicams Forums > Digicam Help > What Camera Should I Buy?

Reply
 
Thread Tools Search this Thread
Old Sep 27, 2006, 5:51 PM   #11
Senior Member
 
Idan's Avatar
 
Join Date: Jan 2006
Posts: 267
Default

It is written in foveon website that Since silicon absorbs different wavelenghts of light at different depths, each layer recors a different color. because the layers are stacked together. all three colors are captured.

So first I tought that if the light passes thourgh the blue layer, the other spectrum will not be absorb at the other layers, and thelight that was filtered though the first layer (the blue) will then encounter the green layer with only the blue light spectrum, but as far as I can understand from the images on foveon website and the text, it is not like that, all the spectrum absorbs in each location of a pixel. did I get it right ??

Check the link HERE
Idan is offline   Reply With Quote
Old Sep 27, 2006, 5:57 PM   #12
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

It's absorbed at each physical pixel location. But, it's absorbed at a different depth within the same pixel location, and you don't have filters to help separate the colors like you do with a Bayer based sensor using multiple locations and interpolating. There is no "continuous spectrum readout" involved. They try to get what they can from only 3 readouts per location (one readout per layer, with separate layers for red, green and blue).

So, from a color accuracy perspective, the Foveon technology can leave something to be desired, based on comments I've seen (both from people like Dave Coffin writing software to process images from sensors about color separation between layers, and from a number of Sigma owners). I can recall reading through a number of threads where Sigma owners were trying all sorts of techniques trying to get accurate colors from some of these models with very little success before Sigma refined the software more.

There are pros and cons to both approaches.

JimC is offline   Reply With Quote
Old Sep 27, 2006, 6:00 PM   #13
Senior Member
 
Idan's Avatar
 
Join Date: Jan 2006
Posts: 267
Default

rjseeneyyou may be right in that the differences are not like black and white differences for sure, professional photographers use Nikon and Canon and provides magnificent photos, but if you check the reviews you'll see that reviewers when there are doing deep photography analysis are looking and comparing the colors (are they real life true to scene colors) and sharpness (like how pro lenses contribute to the sharpness of the image), horizontal and vertical accuracy with consistent lab testing, other things are more robust functions and improvemnt like high iso performance and noise comparison,shots per seconds etc..etc.. that are, at least for me, less important.
Idan is offline   Reply With Quote
Old Sep 27, 2006, 6:06 PM   #14
Senior Member
 
Idan's Avatar
 
Join Date: Jan 2006
Posts: 267
Default

Quote:
There is no "continuous spectrum readout" involved. They try to get one they can from only 3 readouts per location (one per layer for red, green and blue).
I just can't fully understand the meaning of 'continuous spectrum readout'. can you link me a page with in depth information or you can try to explain to me again with better details. :?

Check this LINKand HEREit looks that theysay that Foveon captures light like Film does !


Idan is offline   Reply With Quote
Old Sep 27, 2006, 6:18 PM   #15
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

Nobody does it. So, no, I can't give you a link. It wouldn't be practical. Instead, you do one readout per color for red, green and blue (and in the Foveon's case, it's based on layers more sensitive to each color due to the way light is absorbed. It's not an unlimited number of readouts or layers (my continuous spectrum comments).

But, from my point of view, Sigma's marketing implies something differernt (and I'll admit that it's pretty slick marketing).

In the case of a Bayer Pattern CCD, it's based on reading separate photosite locations that have filters that make them sensitive to only one color each instead of layers at the same location.

Is there a difference? Sure. Is the difference as great as Sigma's marketing implies? Not in my view, and the lack of separation between layers can cause some additional hurdles from an image processing perspective that can impact color accuracy trying to do only 3 readouts (one each for red, green and blue) per pixel location.


JimC is offline   Reply With Quote
Old Sep 27, 2006, 6:20 PM   #16
Senior Member
 
Idan's Avatar
 
Join Date: Jan 2006
Posts: 267
Default

As I mentioned before, I need to ask how does Film reacts to light, is it not the same as foveon, arranged in layers ??

See this Link
Idan is offline   Reply With Quote
Old Sep 27, 2006, 6:34 PM   #17
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

You're missing my point. With film emulsion, you are passing light through it to print the images, and you get the full color spectrum the emulsion is capable of recording (or, you're just reading the light passed through with a typical sensor to capture the colors).

If you do a readout from a sensor with 3 layers, you are trying to do only 3 readouts (one per layer) per pixel location to represent a full spectrum of color, when the color is not that well separated between layers. Yet, you're trying to dump it into "containers" for red, green and blue during the image processing process, based on only 3 readouts for the pixel location (one per layer).

It looks great on paper, and there are some advantages to having all 3 colors at the same physical location (eliminating some of demosaic artificats you can get with a Bayer sensor).

But, you don't have the advantage of color filters to give more separation between red, green and blue like you do with a Bayer Pattern sensor design. I'd have to dig around for them. But, I've seen a lot of complaints about color accuracy from Sigma owners in the past on other forums. Being able to capture all 3 colors at a single location is an advantage. But, color accuracy can leave something to be desired, due to the way they're accomplishing it, because of lack of separation between layers.

IOW, don't get too caught up in the marketing hype. ;-) Instead, I'd wait for reviews to see how well it performs.

JimC is offline   Reply With Quote
Old Sep 27, 2006, 6:37 PM   #18
Senior Member
 
Idan's Avatar
 
Join Date: Jan 2006
Posts: 267
Default

Is there any preview on the way.. ??
Idan is offline   Reply With Quote
Old Sep 27, 2006, 6:50 PM   #19
Administrator
 
Join Date: Jun 2003
Location: Savannah, GA (USA)
Posts: 22,378
Default

Not that I'm aware of.

I think Sigma surprised most reviewers with this one. It wasn't until very recently that information on the new Sigma models started leaking out, and Steve didn't even have a press release from Sigma on the SD14 until right before they wanted to go public with it.


With the number of new cameras coming out this time of year, you may have to be patient to see any in depth reviews from popular review sites.

It may end up being a super camera, and Foveon's had a lot of time to try and refine the sensor technology (and Sigma's had more time to refine the software).

But, I try not to buy into the marketing hype until seeing results from new products. I'll give them credit though. The marketing material looks very slick. But, so did the marketing material for the SD9.

A lot of people really like their Sigmas, too. Again, there are pros and cons to any technology.

JimC is offline   Reply With Quote
Old Sep 28, 2006, 12:05 PM   #20
Senior Member
 
BenjaminXYZ's Avatar
 
Join Date: Jul 2006
Posts: 788
Default

Lets perform some calculations here concerning the image sensor:

Sensor dimension;

New Foveon X3 14.1 million pixels CMOSsensor: 20.69 x 13.79 mm = 235.3151 *mm

Since the 14.1 M.P. photon-detectors is to be divided into equal amounts of RGB pixels. Therefore: 14100000/3 =4700000 [4.7 M.P. effective pixels]

235.3151 *mm/4700000 =.00005006704 mm (Area of each photo-sitesin 4.7 M.P. terms, on235.3151 *mm Foveon X3 image sensor) [That is also true for the each & individual RGB photo-sites in the X3 sensor] - [Since they are all staked up one upon another]

***Picture this: There are 14.1 million .00005006704 mm photo-sites in the new X3 sensor, since every original .00005006704 mm pixels have 3layers of .00005006704 mm pixel area!***



Lets take a look at the pixel pitch of the 6.3 M.P. SONY CCD: (As in the Nikon D70s presently)

Sensor dimension: 23.7 x 15.5 mm = 367.35 *mm

Since this is a BAYER design CCD, there is no pixel divisions. (All the photo-detectors are located on the same plane)



367.35 *mm/6300000 = .00005830952 mm (Area of each & individual photo-detectors in the 367.35 *mm CCD)

Summary:

Based on my calculations, the old 1.5 crop factor 6.3 M.P. SONY CCD still have larger photo-sites than the new 4.7 M.P. Foveon X3 CMOS sensor with the 1.74 crop factor (Focal Length Multiplier).

My earlier guess was wrong; that the new Foveon X3 CMOS sensor will be having the larger photo-sites.

That is it, that's all my concerned for now.














BenjaminXYZ 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 8:57 AM.