Go Back   Steve's Digicams Forums > Digital Cameras (Point and Shoot) > Canon

Reply
 
Thread Tools Search this Thread
Old Feb 10, 2004, 7:49 PM   #1
MPS
Junior Member
 
Join Date: Feb 2004
Posts: 1
Default Canon Pro 1: colour bit depth?

In the new Canon PowerShot Pro1 8 megapixel camera brochure I was surprised to see that the specs listed the Colour Depth as “8 bits x 3 colours.” On a camera of this calibre, I expected to see 10 or 12 bit colour, or is this a typo?

Or, is this another one of those relatively unimportant measures which seduce the ‘spec browsers’?

I would be interested if some knowledgeable person could comment on this. I’m very interested in this camera, and I would like to upgrade from my Canon G5, and so far, this is one spec that gives me reason to wonder about how this will stack up to, say, the Sony F828?

Thanks,

Martin
MPS is offline   Reply With Quote
Sponsored Links
Old Feb 12, 2004, 9:21 PM   #2
Senior Member
 
Join Date: Jul 2003
Posts: 108
Default Re: Canon Pro 1: colour bit depth?

Quote:
Originally Posted by MPS
In the new Canon PowerShot Pro1 8 megapixel camera brochure I was surprised to see that the specs listed the Colour Depth as “8 bits x 3 colours.” On a camera of this calibre, I expected to see 10 or 12 bit colour, or is this a typo?

Or, is this another one of those relatively unimportant measures which seduce the ‘spec browsers’?

I would be interested if some knowledgeable person could comment on this. I’m very interested in this camera, and I would like to upgrade from my Canon G5, and so far, this is one spec that gives me reason to wonder about how this will stack up to, say, the Sony F828?

Thanks,

Martin
If you are concern about the image quality / noise level I won't suggest you upgrading to Pro1. Everybody knows the 2/3" 8MP high density CCD generate quite a bit more noise than 2/3" 5MP or even 1/1.8" 5MP ones. In practice to get better images you might try either down-sampling 8MP photos or use a lower resolution to capture initially. The former is storage consuming and clumsy in operation. Many would choose the latter, but by checking out that the next lower resolution on Pro1 is 2272 x 1704 which is 4MP. So by "upgrading" from G5 to Pro1 you got *less* pixel practically. For me it is odd.
orion999 is offline   Reply With Quote
Old Feb 12, 2004, 11:21 PM   #3
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default Re: Canon Pro 1: colour bit depth?

Quote:
Originally Posted by MPS
In the new Canon PowerShot Pro1 8 megapixel camera brochure I was surprised to see that the specs listed the Colour Depth as “8 bits x 3 colours.” On a camera of this calibre, I expected to see 10 or 12 bit colour, or is this a typo?
If you only look at the Jpeg (or Tiff) outputs from any camera, they are all (8-bit,8-bit,8-bit) regardless of the number of pixels. What you are refering to is the output of the CCD (raw format) which is anywhere from 10 to 14-bit monochrome under their respective Color Filter Array, ie a 1/3 of each pixel (or some fraction of it) will have just one color: R, G, or B. It's only after the de-Mosaic process where each 10 to 14-bit monochrome pixel are combined to form a full (8-bit,8-bit,8-bit) color pixel!

It's just too early to tell anything about noise from 1-sample since each manufacturer can have a different image processing algorithm as well as a different flavor of 8Mp, just like the previous 5Mp. Sony picks an RGBE version, Canon choses the RGB, previously the Nikon CP5700 used a complementary CFA, but no one knows yet what they use for the CP8700, ditto with the Minolta A2! :P
NHL is offline   Reply With Quote
Old Feb 13, 2004, 9:32 AM   #4
Senior Member
 
Join Date: Dec 2002
Posts: 5,803
Default

Hey NHL, you might know something about this.

What does the camera do with the extra data above the 8-bit per pixel? Is it purely used to produce the color? Or maybe some smart person has come up with a way to help with noise? Or dynamic range? I don't know, just a thought that sprung into an inquisitive mind.

Eric
eric s is offline   Reply With Quote
Old Feb 13, 2004, 10:04 AM   #5
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default

Eric

It's all there... What you are forgetting is that the 10 to 14-bit are only mononochrome for each individual pixel, what you're getting in return is a full 24-bit number for each color pixel (albeit @ 8-bit of each). ie you're not dropping any bits, you're actually gaining 10 to 14-bit of extra information from the surrounding pixels!

The only thing you're loosing are the round-off error during the internal camera computation... of course if you take the RAW output from the camera and do the computation in the PC instead, not only you end up with less round off (higher CPU power + 32/64-bit floating point), but you can also save as 16-bit (ie 48-bit color), and also depending on the algorithm used better noise management as well. 8)
NHL 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 9:13 AM.