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

Reply
 
Thread Tools Search this Thread
Old Jan 30, 2003, 12:04 AM   #21
Senior Member
 
wsandman1's Avatar
 
Join Date: Jun 2002
Posts: 318
Default RE: 12 Bit Color

Most digital cameras, including the D100 produce 12 bit color in Raw mode. When a raw image is opened in photoshop, the image is converted into a 16 bit mode image. If you operate the camera in jpeg mode the image will be captured in 8 bit per color mode. What does this mean? 8 x 3 (RBG) = 24 bit or 16 million colors. 12 x 3 = 36 bit or billions of colors. The bit depth assignment is for each color. The advantage of using a raw image is that you will have greater range of adjustments (brightness, contrast, color)than with jpeg or TIF. If you use photoshop, you'll notice that many of the tools and adjustment options don't work until you convert the image to 8 bit mode. But those that do will have a greater range of adjustment. I got all this from attending the Nikon photo school. Someone asked the instructor "why would anyone use RAW images instead of TIFF or JPEG?" Part of the answer was the ability to have the image processed in 12 bit mode instead of 8 bit.
wsandman1 is offline   Reply With Quote
Old Jan 30, 2003, 8:06 AM   #22
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default

We are all saying the same thing, let me try to explain the "disconnect":

1. A monitor color resolution is independent (from 16 to true-color) so let keep this aside.

2. A CCD is monochrome and capture a 1/3 of its CCD in 12-bit Red, a 1/3 in 12-bit Green, and a 1/3 in 12-bit Blue from its filter arrays. Theses array are arranged in a 'Bayer' pattern (ie with color voids in between, unlike the Foveon sensor) the most common for RGB. The camera then re-construct the individual pixels of the picture which are now in full color (ie no color void!), so instead of three individual 12-bit monochrome pixels red, green, and blue recorded by the CCD, we now have three individual 24-bit full 16,777,216 colors pixels (8-bit x 8-bit x 8-bit combination) on the resulting image file that is finally displayed on a monitor. The round off could occur during the computation or during the compression (Do you get the full pixels count in jpeg anyway?). But I believe 24-bit is rather like a standard for sRGB that most PC used.
A Nikon camera actually capture images in complementary color with a fourth channel (but not more resolution) from a different filter array arrangement so the reconstruction and transformation are slighly different!

3. Photoshop get the full 12-bit Red, 12-bit Green, 12-bit Blue (Ye, Cy, Mg, and G for the 5700) in the raw mode from the camera. Of course with a higher horsepower Pentium4 and 32-bit floating-point arithmetic everyone will get better results without round-off or susequent jpeg compression than the cost-conscious/battery-miser processor internal to the camera after the conversion! (ie off-camera vs in-camera processing)

Everyone still on-board?


... then like always we have a few exceptions ie 36(3x12)-bit color jpeg:
http://www.kodak.com/global/en/profe...4.5.12.3&lc=en
Quote:
High-quality JPEG files can be stored in-camera for later color correction or manipulation, saving precious storage card space. All of the original image data is preserved.
NHL is offline   Reply With Quote
Old Jan 30, 2003, 5:29 PM   #23
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default

Here we go:

http://forums.dpreview.com/forums/re...essage=3996121
Quote:
Most of the time colors are expressed as an 8-bit RGB value like (86, 121, 55 - a grassy yellow green) That's because most devices (and the human eye) cannot easily distinguish between two successive values (e.g. that color and (87, 121, 55) would look almost identical). So why do people work in 16-bit color? Two words: rounding errors. Apply many edits and the errors may accumulate to a visible difference. This is also why the camera samples at 12 bits instead of just 8 (and calculates in 16 bits or more.)
Also check the "color space" and sRGB vs AdobeRGB...
Quote:
This relationship is also why if you view an AdobeRGB image in a viewer that assumes everything is sRGB, then it will look dull. sRGB (216, 0, 0) is a dull red.
NHL is offline   Reply With Quote
Old Jan 30, 2003, 7:31 PM   #24
Senior Member
 
Join Date: Oct 2002
Posts: 173
Default comment

Whoa....this rounding stuff I have to think about

I appreciate your research NHL. Good stuff.

The Nikon 5700 does have a 12 bit per channel mode in raw file format

Jpeg is not a lossless file format. The compression does reduce the quality of the photo. It may still look great, dont get me wrong and the space saving may be well worth it. I shot Jpeg photos and I thought they were quite good. I switched to Tiff because I wanted to do large prints and have all the data I could get.

Tiff and raw are the only lossless file formats available with the Nikon 5700.


correct me if I am wrong, but thats what I am reading. Anyone else have any other comments in this long discussion.

Jim
Jim Ghiringhelli is offline   Reply With Quote
Old Jan 30, 2003, 8:28 PM   #25
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default

Jim

Quote:
I switched to Tiff because I wanted to do large prints and have all the data I could get.
If you do this from the camera you're still at 8-bit (somebody correct me?), and raw(nef) is still better for you while the file size is also smaller than tiff and @ 12-bit to top it off.

If you do this from the PC to preserve jpeg then that's a different story... but then so is keeping your original jpegs safe! It is still more compact and always edit from the original and save the steps in an action (.atn file) in Photoshop! Tiff is huge unless you're doing commercial work...
NHL is offline   Reply With Quote
Old Jan 30, 2003, 8:44 PM   #26
Senior Member
 
Join Date: Dec 2002
Posts: 5,803
Default

I believe (but I'm not positive) that NHL is right. Some google searching shows that tiff is 8-bit (and raw is 12-bit.)

The other benefit of the NEF/RAW format is that it causes the camera to not do some processings (white balance is one, I believe.) The logic is that using a full blown desktop computer with a powerful processor (not the one in the camera) you can do much more complex and better image processing. The camera is a balance of battery life vs. quality of pictures vs. size of picture (ram & flash.) Among other balances. A PC removes the battery life part.

Personally, this means it relies on the skill of the editor (me) which means that right now the camera would probably do a better job.

Eric
eric s is offline   Reply With Quote
Old Jan 31, 2003, 6:19 AM   #27
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default

Eric
Quote:
Personally, this means it relies on the skill of the editor (me) which means that right now the camera would probably do a better job.
Not really... If you open the raw picture, the application that comes with the camera (NikonView in this case) will open the CCD's raw sensor readings along with the camera settings (ie shutter, aperture, as well as others camera parameters) that are also embedded in the nef file to reconstitute the same picture as if it was done in the camera! 8) (unless of course you played with the application's adjustments!)

Now I'm going to stick my head out again, since my only Nikon is a CP990... If you save this raw picture to either both jpeg or tiff format using NikonView, Irfanview etc... you'll still end up with only a 24-bit color picture at this point! Importing raw file to PhotoShop (you need a plug-in for this, .mrw in my case) however, will not only let you work in a 16-bit environment, but also will let you save natively with the higher color resolution option. Another reason not to use tiff unless one has the need for it!

Any expert from the Nikon school?
NHL is offline   Reply With Quote
Old Jan 31, 2003, 10:33 AM   #28
Senior Member
 
Join Date: Dec 2002
Posts: 5,803
Default

I have actually wonder about this. Now I have the perfect opertunity to ask!

Since most (all?) of the settings which the camera used when taking the picture are saved with the picture, does NikonView have the ability to do exactly the same processing as the camera would have done had I used "fine/jpg" (i.e. apply the white balance and other skipped processing)?

If it did, then I could directly see what the camera would have produced and what I could do with the same raw data... and see who does a better job!

NHL
One of the biggest reasons (to me) of using raw will be that it has more bits of color. That way photostop (or bibble, which I hear does a better job) can take that .NEF file and turn it into a 16-bit per channel standard format. From there photoshop will do wonders. The rounding errors that others have mentioned are annoying. What bugs me is that important features of photoshop are still 8-bit per channel only (I'm told.) When will they finish moving to 16-bit througt-out?!

I have to say, though, that someones idea that using RAW means they don't have to worry about white balance and a few other camera settings... and therefor they feel more free to take pictures (instead of worry over those settings.) I like that too. Less worry, more pictures. A great combination.

All this talk just makes me more excited to get my D100!
eric s is offline   Reply With Quote
Old Jan 31, 2003, 12:01 PM   #29
NHL
Senior Member
 
NHL's Avatar
 
Join Date: Jun 2002
Location: 39.18776, -77.311353333333
Posts: 11,547
Default

Eric

The WB, exposure compensation etc... are saved as tagged info along with the raw data from the CCD for subsequent use by the program in the computer. That's why you can somewhat overide the camera setting since the initial image capture is still intact from the CCD...
Quote:
If it did, then I could directly see what the camera would have produced and what I could do with the same raw data... and see who does a better job!
Does it really matter? Eventually it's you that the picture has to please, and most of the time you do the tweaking even in jpeg! The camera can't tell an artist apart... You will do the better job!

Quote:
One of the biggest reasons (to me) of using raw will be that it has more bits of color. That way photostop (or bibble, which I hear does a better job) can take that .NEF file and turn it into a 16-bit per channel standard format. From there photoshop will do wonders. The rounding errors that others have mentioned are annoying. What bugs me is that important features of photoshop are still 8-bit per channel only (I'm told.) When will they finish moving to 16-bit througt-out?!
Same here... it's fine for precise multi-sessions editing and for keeping rounding errors to the minimum, but in the end:
Quote:
Most of the time colors are expressed as an 8-bit RGB value like (86, 121, 55 - a grassy yellow green) That's because most devices (and the human eye) cannot easily distinguish between two successive values (e.g. that color and (87, 121, 55) would look almost identical)
Now it also becomes a practical issue: would (86.4321, 121, 55) makes a difference when you can't distinguish it from (87, 121, 55)? But requires significantly larger files, bigger memories (or slower execution time)?... Cost also becomes a factor and that's probably why the CCDs still capture only 12-bit to keep it commercially practical

Remember the above AdobeRGB (the D100 can also map to) colorspace discussion? You are also limited in what you are seeing anyway on your sRGB PC screen! What good is the dynamic range if your monitor is limited?... More bit just increase the resolution in between, the end points stay the same in sRGB(ie gamut), IMO you just make your PC run harder that's all!

I see it now... with enough bit one can make the steps continuous and back to analog again... Let's just go with Film!!! :lol: :lol: :lol:
NHL is offline   Reply With Quote
Old Jan 31, 2003, 6:31 PM   #30
Senior Member
 
Join Date: Oct 2002
Posts: 173
Default comment

To add another question to this colorspace issue is if your photos are only for your viewing on a monitor or to be printed on an inkjet
printer or other print device.

Inkjets are CMYK devices and have a more limiting gamut than your monitor's RGB gamut. If you use Photoshop you can see that by setting gamut alert on and seeing dark brown on colors of your image that wont print CMYK.

This also gets into areas of printer profiles to match printer/ink/media to the CYMK image in Photoshop.

which makes me think that a 24 bit image (8X3) is just fine.
Even if you do use raw more which has certain advantages
you will have to convert back to a Tiff or Jpeg to print.

Jim
Jim Ghiringhelli 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 10:43 AM.