Yahoo Groups archive

Digital BW, The Print

Index last updated: 2026-04-28 22:56 UTC

Message

Re: [Digital BW] Storage of digital images

2002-07-31 by Editor P.O.V. Image Service

Austin Franklin wrote:

>>>Surely every pixel on a digital camera (except
>>>for a few exotica) is interpolated?
>>>      
>>>
>>No.
>>    
>>
>
>YES.
>  
>

Not just "exotica" anymore --  the CCD in the Sigma is the Foveon chip.. 
It reads red green and blue in each sensor well.  So, no interpolation, 
and no need for color filters to do it.. That chip has three layers of 
substrate, so each well measures one color at each level.. It doe NOT 
use a Bayer pattern.. The chip design was originally used on spy 
satellites..

MicroSoft has put a ton of money behind this chip too, probably with the 
intention of developing WinCams (like Winmodems, where the system does 
the actual controller work). You could just plug in a cam that is almost 
all chip to a USB or FireWire port and use it as you wish..  Also, it 
could then easily be incorporated into PDAs or even cellphones using Win 
OS's etc..

There is a HUGE processing saving (and that means in time as well) 
because no interpolation is necessary.. It should avoid the vast 
majority of moiré issues as well..

>  
>
>>Normally each pixel in the output image corresponds to a physical
>>pixel
>>in the CCD.
>>    
>>
>
>Anthony, you're doing it again.  You're trying to talk authoritatively about
>something you don't seem to understand.  AND...you're responding to a
>question with an answer that, though your "answer" just so happens to be
>somewhat true, has nothing to do with the actual question.  Fact is, EVERY
>output pixel of a Bayer pattern CCD camera IS interpolated, period.
>  
>
YUP

>The CCD has SENSOR elements, not pixels.  Pixels are the result after the
>sensor information is interpolated.
>
>  
>
>>The only interpolation that occurs is for color, since each
>>pixel has a color filter over it to restrict incoming light to
>>red, blue,
>>or
>>green (with green being twice as common as the other two
>>colors, since the
>>eye is more sensitive to it).
>>    
>>
>
>Since the image sensors only give ONE value for each sensor location, and
>each sensor location is only ONE of three colors, how is the "color"
>information separated from the luminance information at the sensor level?
>Answer is, it isn't.  EVERY SINGLE output pixel (which contains RGB
>information, not chrominance and luminance information) is derived via
>interpolating the sensor data.
>
>  
>
YUP..

You only get one value from the sensor well in a Bayer pattern CCD. 
 That value basically equates to a luminance value for a particular 
color range (whatever range the filter transmits - or which the well is 
sensitive to)..  Software then takes the values fed by the sensors and 
interpolates color and luminance for 3 colors at each position.. That 
weakness is why edges can look funky or why we get many of the moires 
one sees..  Accordingly, the processing of a Bayer Pattern CCD output is 
at least as important as an accurate CCD..

With the Foveon chip, the chip itself is of vastly greater import in 
final image quality...  One advantage of the Foveon approach is that you 
need less actual wells to achieve comparable final resolution..


Keith




[Non-text portions of this message have been removed]

Attachments

Move to quarantaine

This moves the raw source file on disk only. The archive index is not changed automatically, so you still need to run a manual refresh afterward.