8-bit vs 12-bit scanning

8-bit vs 12-bit scanning

Post by David Blanchard, » Tue, 12 Mar 2002 20:29:22



Greetings,

I'm puzzled about the choice between using 8-bit and 12-bit
scanning on my Nikon LS-2000.  I've done both and saved them
as two files.  Then I've tried to open them with Adobe
Photoshop Elements (not the full-blown Photoshop).  There is
no problem with the 8-bit scan. Elements complains about the
12-bit image indicating it is not a supported bit depth...
and then reduces it to 8 bits.

Questions:

1.  Is Photoshop Elements different from Photoshop, i.e.,
does Photoshop support 12-bit scans while the less expensive
Elements does not?  (I'd test this myself if I had the full
version of Photoshop...)

2.  And is there much value in 12-bit scanning vs. 8-bit
scanning, anyway?

David
--
+----------------------------------------------------------------------+

+----------------------------------------------------------------------+

 
 
 

8-bit vs 12-bit scanning

Post by Mac McDougal » Tue, 12 Mar 2002 23:46:35



Quote:> 1.  Is Photoshop Elements different from Photoshop, i.e.,
> does Photoshop support 12-bit scans while the less expensive
> Elements does not?  (I'd test this myself if I had the full
> version of Photoshop...)

Yup, that's it.

Quote:> 2.  And is there much value in 12-bit scanning vs. 8-bit
> scanning, anyway?
> David

The value is in the scanner/software SAMPLING at high bit and then
choosing the best to keep.

In my opinion, actually manipulating a 48 bit file has minimal, if any,
practical purpose (for my needs). Btw, full PhotoShop has limited tools
that work on 48 bit.

I output to film also, and can see no diff between 24 and 48 bit output,
and if *film* can't see it, well....

I understand there are certainly some scientific uses for the extra data,
but monitor/inkjet/film recorder can't see it, so I don't use it.

--
        Mac McDougald
Doogle Digital - www.doogle.com

 
 
 

8-bit vs 12-bit scanning

Post by Bob Shomle » Wed, 13 Mar 2002 00:13:04


Quote:> 2.  And is there much value in 12-bit scanning vs. 8-bit
> scanning, anyway?

Value of passing 12-bit [per channel] data to Photoshop depends in part
on how much tonal and color editing you will there do on the image data
vs how much is performed to your satisfaction by the scanner subsystem
before it reduces its internal 12-bit data to 8-bit for image file
output.

Bruce Fraser has a good explanatory article on the pros and cons of
editing in 16-bit mode at creativepro: "The High-Bit Advantage -- Why
bother working with 36- or 48-bit color when you're stuck with 24-bit
output?"

  www.creativepro.com/story/news/7627.html?cprose=I20

Bob Shomler
www.shomler.com

 
 
 

8-bit vs 12-bit scanning

Post by Wayne Fulto » Wed, 13 Mar 2002 01:54:08



Quote:>1.  Is Photoshop Elements different from Photoshop, i.e.,
>does Photoshop support 12-bit scans while the less expensive
>Elements does not?  (I'd test this myself if I had the full
>version of Photoshop...)

Right, the Photoshop full and LE versions have a 16 bit mode.  
But Elements does not.

Quote:>2.  And is there much value in 12-bit scanning vs. 8-bit
>scanning, anyway?

It is real important that the scanner do it internally, to aid the
results of doing the gamma and tonal corrections.  All scanners do,
since there are no more 24 or 30 bit scanners currently manufactured.

But if the scanner does the tonal/contrast adjustments halfway well, and
it should, then there is much less benefit of outputting the full data.  
24 bits is all monitors and printers can use, and I doubt the human eye
could see it if they did more.  

Relatively mild adjustments to 24 bit data is no big deal, it has been
done for years.  Extreme adjustments on 8 bit data can leave detectable
effects, but usually do not, and there is nothing to compare it to
anyway.  If the scanner did it halfway well while it had the 12 bit data
available, this should not be an issue.  

But some people do prefer to output the full data and to do most of the
tonal edit work externally in 16 bit mode.  Whether it really matters or
not, it removes all doubt. <g>  After completion of this tonal editing,
then the image must be converted to 24 bits for the rest of the world.

--
Wayne

http://www.scantips.com    "A few scanning tips"

 
 
 

8-bit vs 12-bit scanning

Post by Bart van der Wol » Wed, 13 Mar 2002 01:21:48




SNIP
> > 2.  And is there much value in 12-bit scanning vs. 8-bit
> > scanning, anyway?

> > David

> The value is in the scanner/software SAMPLING at high bit and then
> choosing the best to keep.

Yes, like Mac said, sampling in 12 or 16-bit per color mode is very
important.
One of the most important reasons is that CCD scanners have a linear gamma
response. In converting that linear capture gamma to something that displays
well on a monitor, a gamma correction needs to be applied. Any subsequent
image editing increases the possibility of so called banding in the lighter
areas of the image (e.g. the sky, or smooth skin). The risk can be reduced
by doing most color/brightness corrections in 16-bit mode, e.g. with the
scanner software.

The 16-bit per color mode in the full Photoshop version helps to preserve
accuracy even when severe manupulations are applied to the image. Only when
outputting the final workpiece, it will usually convert back to 8-bit per
color (e.g. printing).

Bart

 
 
 

8-bit vs 12-bit scanning

Post by David Blanchard, » Wed, 13 Mar 2002 04:23:43





>>1.  Is Photoshop Elements different from Photoshop, i.e.,
>>does Photoshop support 12-bit scans while the less expensive
>>Elements does not?  (I'd test this myself if I had the full
>>version of Photoshop...)

>Right, the Photoshop full and LE versions have a 16 bit mode.  
>But Elements does not.

Well, I have an old version of LE floating around somewhere.  Maybe I'll
relaod it and see if I can use it for an intermediate stage.

I also have GraphicConverter which appears to support the greater bit depth.

Thanks, Wayne, Mac, Bob, and Bart for your informative answers.

David
--
+----------------------------------------------------------------------+

+----------------------------------------------------------------------+

 
 
 

8-bit vs 12-bit scanning

Post by Bill Perkin » Wed, 13 Mar 2002 23:17:28


Gentlemen,
Could one of you clarify some points to a novice.  When you speak of 8, 12,
or 16-bit, are you referring to a single color channel (did I say that
correctly)?  And when you talk of 24, 32, 48-bits, are you referring to the
combined RGB channels?  If not, would you please explain the difference.

Also, what arithmetic is involved in calculating total available colors?

Thanks in advance,
Bill Perkins

 
 
 

8-bit vs 12-bit scanning

Post by Wayne Fulto » Thu, 14 Mar 2002 01:15:47




Quote:

>Gentlemen,
>Could one of you clarify some points to a novice.  When you speak of 8, 12,
>or 16-bit, are you referring to a single color channel (did I say that
>correctly)?  And when you talk of 24, 32, 48-bits, are you referring to the
>combined RGB channels?  If not, would you please explain the difference.

>Also, what arithmetic is involved in calculating total available colors?

Yes, one channel or three RGB channels (only one for grayscale).

Three 12 bit channels of RGB give 36 bits total.  
This is normally truncated to 8 bits / 24 bits for output.

Some scanners can optionally output three 16 bit words called 48 bits,
containing the full 36 bits of data.  Computers only want words in 8 bit
byte multiples.   Only a few photo editors can accept this (called 16 bit
mode), but Photoshop full and LE version can.

The largest number representable in 8 binary bits is 255.
The range of 0 to 255 is 256 possible colors.

The possible color combinations in three 8 bit color channels is
 256x256x256 = 16.7 million color combinations.

--
Wayne

http://www.scantips.com    "A few scanning tips"

 
 
 

8-bit vs 12-bit scanning

Post by s » Sat, 16 Mar 2002 16:47:18



>Gentlemen,
>Could one of you clarify some points to a novice.  When you speak of 8, 12,
>or 16-bit, are you referring to a single color channel (did I say that
>correctly)?

Yes * 2!

Quote:>And when you talk of 24, 32, 48-bits, are you referring to the
>combined RGB channels?

Yes, allthough in that respect the concept of "pixel" is most widely used. A
pixel holds an 8 or 16 bits storage space per channel (color, either of R, G
or B), that may hold up to 8 or 16 bits of data.

Quote:>If not, would you please explain the difference.

>Also, what arithmetic is involved in calculating total available colors?

The largest number using 8 bits is 255. Accordingly, it holds 256 ( 0-255 )
different values.

256*256*256 (or 256^3) = 16 777 216 colors, or about 16.7 million different
color combinations. This goes well beyond the perceptive ability of the human
eye, and also any output device or output medium.

Regards,
Carsten J. Arnholm, Oslo, Norway.

http://carnholm.home.online.no

 
 
 

8-bit vs 12-bit scanning

Post by Andy Spra » Fri, 05 Apr 2002 09:50:51



the queue on Mon, 11 Mar 2002 17:21:48 +0100, and nailed this to the
shed door:




^ SNIP
^ > > 2.  And is there much value in 12-bit scanning vs. 8-bit
^ > > scanning, anyway?

^ > The value is in the scanner/software SAMPLING at high bit and then
^ > choosing the best to keep.

^ Yes, like Mac said, sampling in 12 or 16-bit per color mode is very
^ important.
^ One of the most important reasons is that CCD scanners have a linear gamma
^ response.

Except that they don't. I've investigated several different makes and
models of scanners in the last couple of years, and invariably they
have an intrinsic gamma in the range 1.05 - 1.1. WHY, is a question I
am still unable to answer. At the moment, I am looking into
differential and integral non-linearity of the ADC as a possible
cause, but I find it difficult to believe that this could account for
such a substantial departure from linearity.

Andy

--
sparge at globalnet point co point uk

 Hire a man a car and you
 transport him into work for a day.
 Teach a man to drive and you
 transform him into an arsehole for life.

 
 
 

8-bit vs 12-bit scanning

Post by Bart van der Wol » Sat, 06 Apr 2002 07:40:07




> the queue on Mon, 11 Mar 2002 17:21:48 +0100, and nailed this to the
> shed door:
SNIP
> ^ Yes, like Mac said, sampling in 12 or 16-bit per color mode is very
> ^ important.
> ^ One of the most important reasons is that CCD scanners have a linear
gamma
> ^ response.

> Except that they don't. I've investigated several different makes and
> models of scanners in the last couple of years, and invariably they
> have an intrinsic gamma in the range 1.05 - 1.1. WHY, is a question I
> am still unable to answer. At the moment, I am looking into
> differential and integral non-linearity of the ADC as a possible
> cause, but I find it difficult to believe that this could account for
> such a substantial departure from linearity.

Interesting. After the "conversion" of photons into electrons, the electrons
are just collected and read-out. So that would point in the direction of a
non-linearity being introduced in the ADC. It can hardly be intentional,
because what's the benefit? On the other hand, not all CCDs are created
equal. Some have optional anti-blooming functionality, which could introduce
deviations (depending on how it's implemented).

Would be interesting to know which CCDs/scanners do this gamma "correction",
because it could perhaps increase the accuracy of programs like VueScan if
factored into the calculations.

Bart

 
 
 

8-bit vs 12-bit scanning

Post by Andy Spra » Sat, 06 Apr 2002 16:57:11



the queue on Fri, 5 Apr 2002 00:40:07 +0200, and nailed this to the
shed door:




^ > the queue on Mon, 11 Mar 2002 17:21:48 +0100, and nailed this to the
^ > shed door:

^ > ^ One of the most important reasons is that CCD scanners have a linear
^ > ^ gamma response.

^ > Except that they don't. I've investigated several different makes and
^ > models of scanners in the last couple of years, and invariably they
^ > have an intrinsic gamma in the range 1.05 - 1.1. WHY, is a question I
^ > am still unable to answer. At the moment, I am looking into
^ > differential and integral non-linearity of the ADC as a possible
^ > cause, but I find it difficult to believe that this could account for
^ > such a substantial departure from linearity.

^ Interesting. After the "conversion" of photons into electrons, the electrons
^ are just collected and read-out. So that would point in the direction of a
^ non-linearity being introduced in the ADC. It can hardly be intentional,
^ because what's the benefit? On the other hand, not all CCDs are created
^ equal. Some have optional anti-blooming functionality, which could introduce
^ deviations (depending on how it's implemented).

I looked at anti-blooming as a possibile explanation. If it's there at
all, it takes the "gamma" the other way (I mean, a sub-linear response
to increasing intensity, rather than the super-linear which I have
noted).

^ Would be interesting to know which CCDs/scanners do this gamma "correction",
^ because it could perhaps increase the accuracy of programs like VueScan if
^ factored into the calculations.

And I would be interested to know if it is absent in higher-price
scanners. The ones I have measured it in so far:

HP 6100C/ScanJet 4C, 6200C, 6300C
Microtek ScanMaker X6USL, X12 USL

Andy

--
sparge at globalnet point co point uk

 All men are arseholes
 but some are more arseholes than others

 
 
 

8-bit vs 12-bit scanning

Post by Bernard Leve » Sun, 07 Apr 2002 07:25:38



> 1.  Is Photoshop Elements different from Photoshop, i.e.,
> does Photoshop support 12-bit scans while the less expensive
> Elements does not?  (I'd test this myself if I had the full
> version of Photoshop...)

David,
If you want to do your main editing in 48 bits mode, and find
Photoshop (full)a bit expensive, look at picture Window pro
(www.dc-l.com).
For photographers, it can do more and better than Photoshop (but you
don't have all the graphic arts bells and wisles).
Bernard
 
 
 

8-bit vs 12-bit scanning

Post by Bernard Leve » Tue, 09 Apr 2002 00:56:03




> > 1.  Is Photoshop Elements different from Photoshop, i.e.,
> > does Photoshop support 12-bit scans while the less expensive
> > Elements does not?  (I'd test this myself if I had the full
> > version of Photoshop...)

> David,
> If you want to do your main editing in 48 bits mode, and find
> Photoshop (full)a bit expensive, look at picture Window pro
> (www.dc-l.com).
> For photographers, it can do more and better than Photoshop (but you
> don't have all the graphic arts bells and wisles).
> Bernard

Sorry, it is www.dl-c.com

Bernard

 
 
 

1. small MCU with 12-bit ADC and 8-bit PWM output ?

Does anyone know of such Microcontroller ? I'm interested especially in the
12-bit ADC.

Martin.
--
        martin "at" quickstep "punkt" dirnet "punkt" com
--
 Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------

2. VLISP Dev Bible book

3. 8-bit vs. 24-bit color on Windows

4. Gnu Switches for Listing files

5. 12-bit Data in 16-bit VOI LUT

6. MIDI/WinNT4.0? Attach instrument definition file to MPU-401 driver

7. SCAN BIT DEPTH vs CROP BIT DEPTH Vuescan HPs20

8. AWARD bios for INTEL PD440FX motherboard?

9. PIC How to: 24 Bit / 16 bit = 16 bit (division routine on a 8 bit PIC )

10. 8 bit or 12 bit

11. 35mm Film scanner - 10 bit or 12 bit - whats gained ?

12. 10 bits Versus 12 bits for image recognition

13. convertir un mot de 16 bits en un mot de 12 bits