Thick Headed

[Onerider]

I have read threads about 8 and 16 bit manipulation but I still don't quit get something.
When you use LR or maybe some other 16 bit editor, then move it to PSE or GIMP, some actions require a conversion to 8 bit.
When you move it back to LR, don't you loose something because it's now an 8 bit image? In my thick head it seems you have lost some of the quality that you were trying to achieve.

 

[480sparky]

Yep.. once it's 8-bit, it's always gonna be 8-bit. You can't get that data back. That's why converting to an 8-bit format should be your last step.

 

[KmH]

If you move an 8-bit depth image into Lightroom, you still have an 8-bit depth image. If you move an 16-bit depth image into Lightroom, you still have an 16-bit depth image.
Some tools are 8-bit only tools because if they were 16-bit capable the processing time it would take to do the edit is to long.

Lightroom/Camera Raw (ACR) cannot edit pixels, which is why ACR is known as a parametric image editor. Parametric edits are XML line commands in a .XMP text file that change how the algorithms render the image data file.

It is also helpful to know that Lightroom uses a very broad color space akin to ProPhoto RGB that you cannot change. No doubt you can change the input or output color space, but there is no way you can change the color space Lightroom uses to display images.

PsE, Gimp, and Ps are raster (bitmapped pixels) graphics editors that can and do edit pixels. There is also vector (mathematically defined lines shapes) graphics editors. Adobe Illustrator, and Inkscape are vector graphics editors.

 

[Onerider]

Ok, thanks for clearing that up for me. It seems a shame that you loose half the image info though. Is there a way that this is compensated for?

 

[KmH]

You lose way lots more than 1/2, and no there is not a way to compensate for what is lost.

8-bits can only represent 256 discrete values. Particularly in smooth gradients like the sky, 256 colors doesn't allow for smooth transitions and steps can often be seen (banding/posterization) in the gradient.

Digital cameras actually record photos as 12 or 14 bit depth files. 12-bits can represent 4096 discrete values (7.8x), and 14-bits can represent 16,384 discrete values (64x).

Photo Editing Tutorials

The colors in a digital photo are rendered using the RGB color model.
RGB stands for Red, Green, and Blue.
So there are 3 color channels that define how many colors can possibly be rendered.

8-bit depth files are also known a 24 bit files, 8 x the 3 RGB color channels so 8-bits can render a total of 256 * 256 * 256 = 16,777,216 colors.
12-bit depth files (36-bit files) can render 4096 * 4096 * 4096 = 68,719,476,736 colors
14-bit depth files (42-bit files) can render 16 384 * 16 384 * 16 384 = 4,398,046,511,104 colors

 

[Onerider]

KmH Thanks for the help. I guess I didn't appreciate just how everything interacted.
Thnx

 

[KmH]

It's time to throw you a couple of curves then.

Photoshop doesn't actually use 16-bits. Photoshop only uses 15-bits.

When a photo is converted to JPEG, one of the ways they achieve the JPEG file compression is by converting the image pixels into 8x8 pixel, 8x16 pixel or 16x16 pixel groups known as Minimum Coded Units (MCU's).

Zoom into a JPEG and you can see the individual pixels and the MCU groups they have been locked into.

 

[Onerider]

Well, I won't worry about one little ole bit..haha
Can you actually see a difference in an 8 and 16 bit image? I have old eyes and can't tell that one is better than the other.

 

[snowbear]

Well, I won't worry about one little ole bit..haha
Can you actually see a difference in an 8 and 16 bit image? I have old eyes and can't tell that one is better than the other.

Interesting you bring this up. IIRC, the average human eye can detect about 13,000,000 colors. A while back I did a quick test to see if I could tell the difference between 8- and 16-bit photos. I couldn't with the image I chose, but someone may be able to, and perhaps a different image would show variations.

Processing was done with the original raw, then exported to 8-bit and 16-bit PNG files.

 

[Onerider]

Very nice. If we can't see the difference, then why bother with 16 bit manipulation?
I think I'm still not tying something together here.

 

[snowbear]

I think it may make a difference when the range of color (or even light/dark value) is very broad. The larger files (more bits) definitely give you more to work with - that's just basic math. Maybe even the file type will make a difference. I think I might play with this a little more this week.

 

[KmH]

The bit-depth is about editing headroom, and ensuring your edits don't produce edit artifacts that can be seen in the final image.

Every edit, including parametric edits, has an associated image quality reduction cost.

Many 8-bit images have no editing headroom at all, and all 8-bit images have very limited editing headroom.

Do as much editing in 16-bit mode as possible before you reduce the bit-depth to 8-bits to use 8-bit only editing tools.

 

[snowbear]

The bit-depth is about editing headroom, and ensuring your edits don't produce edit artifacts that can be seen in the final image.

Every edit, including parametric edits, has an associated image quality reduction cost.

So, i may see some differences if I edit two different depth copies of the same image - say an 8-bit and 16-bit TIF or PNG?

 

[Onerider]

So you are saying, in 16 bit, there are more colors to choose from and the conversion to 8 bit chooses those colors. Whereas in an 8 bit original, there are fewer colors to choose from? If that is right, then I now understand.

 

[snowbear]

So you are saying, in 16 bit, there are more colors to choose from and the conversion to 8 bit chooses those colors. Whereas in an 8 bit original, there are fewer colors to choose from? If that is right, then I now understand.

In a way.

For an analogy, let's say you are given a very large box of apples, but the box won't fit into your car (8-bit file), so you have to put some of the apples in a smaller box in order to get them home. You can't take all of the apples, so you grab what will fit and move along. On the other hand, if you had a larger car or truck (16-bit file) the large box will fit just fine and you don't lose any apples.