Camera settings and raw

[Helen B]

I see a light gray font on a dark gray background.

About half way down he uses the red font for a list of things - my gods does it jar the eyes! (shockingly more than I thought it would).

But the information on that site is generally very good

A few inaccuracies...

There's no standard format for RAW... each camera can be different.

The light-sensitive points which cover the surface of the sensor are not pixels... they are photo-sites. In a typical color camera using a Bayer mask, four photo-sites are clustered together (2 green, 1 red, and 1 blue) to create 1 pixel. A monochrome camera has no Bayer mask, so no need to cluster photo-sites, but it can still cluster to use "binning" modes.

RAW files can be (and usually are) compressed. RAW does not use any "lossy" form of compression so the compression is not very high (unlike JPEG, which deliberately uses lossy algorithms.)

Tim,

You have mentioned this idea that a pixel is made of four photosites before and I have tried to correct you before. I'll try again. A camera with a Bayer matrix over the sensor makes one pixel for every photosite (sensel) at its native resolution.

 

[480sparky39 38]

Tim,

You have mentioned this idea that a pixel is made of four photosites before and I have tried to correct you before. I'll try again. A camera with a Bayer matrix over the sensor makes one pixel for every photosite (sensel) at its native resolution.

Correct me if I'm wrong, but doesn't the data from an individual photosite get 'used' 4 times to create 4 separate pixels?

 

[Helen B]

Yes, when de-Bayering the data from four sensels is used to create four pixels, not one as Tim seems to suggest.

 

[480sparky39 38]

So it's kinda like this:

________________________________________________________

________________________________________________________

________________________________________________________

________________________________________________________


The data from photosite #6 is used in creating 4 individual image pixels, one using data from 1,2,5,6; another using 2,3,6,7; a third using 5,6,9, 10, and a fourth using data from photosites 6,7,10 and 11.

 

[TMC]

geesh, i didn't realize this thread was gonna turn into a school lesson on the bayer pixel of the foveon photosite of the 4th or 5th point on a camera sensor. i didn't realize how stupid i was...

 

[Helen B]

There's no single way of de-Bayering, there are different methods.

 

[Benco]

There's no single way of de-Bayering, there are different methods.

Which thankfully I don't have to worry about....I hope.