Why is the SuperCCD better?

[Seachicken]

I have been comparing the output of the S2 to other 6MP cameras using images from the various review sites. It does seem to be the case that the the 12MP images from the S2 show more detail than the images from other 6MP cameras.

Is this really the case? I have read reports of the differences vanishing when comparing RAW files.

If it is really true, can anyone explain why?
It seems to me that by interpolating extra pixels to go from 6->12 MP no new information is being added to the image, old information is just being re-used.

I have read that the idea is that by rotating the sensor through 45 degrees you gain resolution in vertcal and horizontal directions which due to properties of the human visual system look sharper.

BUT if I rotate the sample S2 images that appear to show more detail through 45 degrees, they still show more detail.

Andy.

[Laren]

From what I understand, the method the S2 uses to make 12 megapixel files is not the same as interpolation.

So the actual resolution of a 12 megapixel file from the S2 is higher than a 12 megapixel file produced by interpolating a 6 megapixel file.

In Popular Photography they said that the resolution of the S2 is actualy closer to a 9 megapixel CCD using conventional methods.
This is somehow achieved becaused the sensors are closer together in the S2 and there is some overlapping.

I would get a copy of the new Popular photography magazine for more (first hand) information.

Besides the higher resolution this camera seems to have a more acurate exposure and produces pictures that are more pleasing right out of the camera than others do.

The people at Popular Photography seemed to be fairly impressed by it.

Larry

[Seachicken]

Thanks,

I'm currenly in Japan, so I don't have access to Popular Photography.. but theres quite a lot of stuff on the web for me to read.

The S2's SuperCCD is clearly designed to go up to 12MB by 'interpolating' in the missing pixels. So it's almost certainly the case that interpolating a normal 6MP image up to 12MP will be less good.

In my original question I was puzzling over the fact that extra detail was present.. but both types of sensor have the same 6 million pixels of information... and adding false dots isn't going to add any new.. just blurr what you already have.

Perhaps the answer is that horizontal and vertical features are just more common in images, and the superCCD's emphasis on these means that most stuff gets imaged in more detail.

For large prints, I will usually upsample an image before printing. I don't like to see the dots, and somehow the image looks better after the process. The SuperCDD does the upsampling for you, and does so in a high-quality fashion.

Another advantage of the sensor is the photo-diodes are larger, will collect more light and therefore be less noisy.

The accurate exposure of the S2 is a definate plus ;) .. though it seems that the D100 community may have circumvented this problem by creating a set of tone curves that can be uploaded into the camera. I don't think the S2 has this feature (?) which would be nice.. the default colors coming from the camera look splendid, rather like Fuji film color.. but it would be really cool if you could load in a more neutral set, or an 'Ektachrome' set... I guess this can all be done by post processing the RAW images.

Andy.