Advantages of Foveon sensor Locked

Started Jun 23, 2013 | Discussions thread
This thread is locked.
bobn2 Forum Pro • Posts: 50,604
Re: Advantages of Foveon sensor

Rudi wrote:

Bobn2 wrote:

Rudi wrote:

Talking about frequency we got to talk about bandwidth too and in this respect I suggest the lecture of Dick Lion´s discours on the Foveon sensor. With the Foveon technology we get exact voltage values for each single color of the color spectrum in each single pixel position.

No, you don't. Each 'voltage' has nothing to do with the 'color spectrum' at all. It is a measure of the integration of luminous energy weighted by one of three different frequency response functions. That produces three 'stimulus' values, not any measure of spectral value. The issue here is that in the Foveon sensor those functions are very different from what they need to be to accurately calculate the three stimulus values needed fro standard RGB colour models.

Those values (for orange color for example) are clearly measured and are the results of simply adding / substracting the three values comming from sensor output lines.

They are not 'clearly measured' - what is measured is three stimuli not at all close to human vision, which require not just adding/'substracting' but complex calculations to get anything close to the CIE stimuli.

With CFA Sensors you need a lot of reference values and have to do interpolation to get the the closest value.

What you are confusing is the work needed to be done by the raw decoder in the XY space of the sensor ('spatial space') and that done in the colour space. The Foveon design indeed eliminates interpolation in spatial space, but it needs an enormous amount of work in colour space, and still suffers from the metamerisms intrinsic in having a completely different colour space.

and here, I think, Sigma has done a pretty good job in firmware design and SPP developement. The Foveon sensor technology is much more complex then CFA sensor technology and that is why it can´t be compared with each other on a 1:1 basis.

I'm not sure what you mean by that. In the end, someone choosing one against the other would choose what was available, for instance the 15MP Foveon against the current crop of 24MP APS-C Bayer sensors. I think they'd be found to produce very similar output, the Bayers would have better colour accuracy (though many here prefer Foveon colour), the Bayer would be cheaper and in the end the Bayer would produce 42Mb per output file while the Foveon would produce 67.5 (and require more complex processing for that). So, when someone came to compare, most of the plus features go with the Bayer.

At least the higher amount of data ( 42MB for a 24MP Bayer vs. 67,5MB for a 15MP Foveon image) proves that a Foveon sensor is giving more data content to play with. If we could get this amount of information from a CFA Sensor the world would look a little different. But the problem is in the RGB Color model not in the sensors which are capable of delivering more information in an image file.

But the Foveon data is not very useful information, it relates to a colour space which is not close to human vision. It detects colours that for us don't exist and fails to catch ones that do exist. Saying that the problem is the RGB colour model is frankly ridiculous, it is that model which is designed to approximate human colour vision. The Foveon colour model might be better for aliens, but not for human beings.

-- hide signature --


Post (hide subjects) Posted by
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow