Digital BW, The Print

Hi, I need some clarification on mode changing from 8 bit to 16 bit. 
It's my understanding that unless the file originated as a 16 bit file 
there is no advantage to change an 8 bit (or whatever) to 16. No net 
gain. Is that right? In a specialized photoshop class there is 
instructions to convert an 8 bit file to 16 but as I recall there is 
not an advantage. Can anyone clarify? Thanks in advance.
Jack

Changing a file from 8 bit t to 16 bit gives you nothing other than a
larger file size.
-- 
amadou diallo
Author, Mastering Digital Black and White
www.masteringdigitalbwbook.com

In a message dated 4/9/07 5:08:31 PM, amadiallo@... writes:


> Changing a file from 8 bit t to 16 bit gives you nothing other than a
> larger file size.
> 

It also gives you more places to spread it around... if you shift the gamma 
on a file that has been shifted from 8 to 16 bit, the side that needs extra 
data won't get it. But the side that is dumping data now has someplace to dump 
it, so its not entirely missing later if you move the other way. But yes, there 
is very little virtue to a low big file being increased in bit depth, even if 
functions such as blurring moves data into the previously empty sections...

C. David Tobie
Product Technology Manager
ColorVision Business Division
DataColor Inc.
CDTobie@...
www.colorvision.com


**************************************
 See what's free at 
http://www.aol.com.


[Non-text portions of this message have been removed]

Hello Jack,

>Hi, I need some clarification on mode changing from 8 bit to 16 bit. 
>It's my understanding that unless the file originated as a 16 bit file 
>there is no advantage to change an 8 bit (or whatever) to 16. No net 
>gain. Is that right? In a specialized photoshop class there is 
>instructions to convert an 8 bit file to 16 but as I recall there is 
>not an advantage. Can anyone clarify? Thanks in advance.

I have found some benefit in doing this when using 8-bit jpgs from my
pocket digicam.  Because I'm converting to grayscale and applying
various curves and layers on the way to a BW print, the resulting
histogram is much less combed and ragged if I first convert to 16-bit
before doing anything else.  I've found it does make a difference in
the quality of the final print.


Regards,
Clayton


Info on black and white digital printing at    
http://www.cjcom.net/digiprnarts.htm

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "jnhugo" 
<jacknadelle@...> wrote:
>
> Hi, I need some clarification on mode changing from 8 bit to 16 bit. 
> It's my understanding that unless the file originated as a 16 bit 
file 
> there is no advantage to change an 8 bit (or whatever) to 16. No net 
> gain. Is that right? In a specialized photoshop class there is 
> instructions to convert an 8 bit file to 16 but as I recall there is 
> not an advantage. Can anyone clarify? Thanks in advance.
> Jack
>
 Thanks for the replies, this list is great and I appreciate you all 
taking time to respond. 
So the bit depth on capture determines how detailed the information is 
sampled, if an image is sampled at 8 bit changing it to 16 will not 
add information that was not originally sampled at that depth. 
Although Clayton is saying that he sees less destruction to the image 
if he converts 8 bit to 16 bit before applying a fair amount of 
adjustments. So while converting doesn't add info it does less damage 
to the file when applying adjustments? Amadou, C David what about 
image degrading when lots of adjustments are applied? 
Thanks again
Jack

Clayton,
Ignoring what the histogram looks like, are you saying that an 8bit
file, converted to 16bit, with various adjustments applied yields a
superior print when compared to the same file with identical
adjustments that has stayed in 8bit? You've done a side-by-side print
comparison in which the only variable is the 8bit to 16bit conversion?

On 4/9/07, Clayton Jones <cj@...> wrote:
>  I have found some benefit in doing this when using 8-bit jpgs from my
>  pocket digicam. Because I'm converting to grayscale and applying
>  various curves and layers on the way to a BW print, the resulting
>  histogram is much less combed and ragged if I first convert to 16-bit
>  before doing anything else. I've found it does make a difference in
>  the quality of the final print.
-- 
amadou diallo
Author, Mastering Digital Black and White
www.masteringdigitalbwbook.com

At 8:09 PM +0000 4/9/07, jnhugo wrote:
Hi, I need some clarification on mode changing from 8 bit to 16 bit.
It's my understanding that unless the file originated as a 16 bit file
there is no advantage to change an 8 bit (or whatever) to 16. No net
gain. Is that right? In a specialized photoshop class there is
instructions to convert an 8 bit file to 16 but as I recall there is
not an advantage. Can anyone clarify? Thanks in advance.
Jack


As others have mentioned, if you are making some corrections to your image,
it may behoove you to go to 16-bit. If you are using Shadow/Highlight in
Photoshop, going to 16 for that tool (and eventually back to 8) has been
found to create an advantage by no less an authority than Dan Margulis,
author of *Professional Photoshop.* I also like it for subtle Curves moves,
but no, I have not done a side-by-side print comparison.

Please note that most of Margulis's expertise deals with color images bound
for a CMYK color space and commercial (offset) printing. But his work is
not confined to that situation, nor relevant only to that situation.

As CD Tobie mentions, capture in 16-bit in the first place gives you more
advantages than capture in 8-bit and upsizing to 16.
-- 

jonathan wills
_______
wills design
portland, oregon, usa

Hello Amadou,

>Ignoring what the histogram looks like, are you saying that an 8bit
>file, converted to 16bit, with various adjustments applied yields a
>superior print when compared to the same file with identical
>adjustments that has stayed in 8bit? You've done a side-by-side 
>print comparison in which the only variable is the 8bit to 16bit 
>conversion?

Yes, the first time I tried it was after my trip to Death Valley NP
two years ago.  I had a number of nice images on my pocket digicam
that were worth printing and weren't on my bigger camera.  One in
particular had a large smooth dark gray area of early morning sand
dune (see photo #4, Morning On The Dunes, here  
http://www.cjcom.net/cal05-a.htm) with a lot of very subtle tonal
transitions.  The difference showed in this area. That was my only
real attempt at a comparison test.  I'd read about it somewhere, and
after that I was convinced and made it a normal part of my workflow
for that camera.  BTW, as Paul mentioned, part of my procedure is to
convert the jpg to psd or tif as well.


Also want to add my congrats on your new book.  It looks really well done.

Regards,
Clayton


Info on black and white digital printing at    
http://www.cjcom.net/digiprnarts.htm

My 2 cents,

I "sometimes" increase the bit-depth.

But, I'm a superstitious fool for the most part.

When I think I want more data, I up-sample one way or another.

Finding out how to mix ingredients in the right order is the important part
of imaging.


gar

-----Original Message-----
From: DigitalBlackandWhiteThePrint@yahoogroups.com
[mailto:DigitalBlackandWhiteThePrint@yahoogroups.com]On Behalf Of jnhugo
Sent: Monday, April 09, 2007 3:40 PM
To: DigitalBlackandWhiteThePrint@yahoogroups.com
Subject: [Digital BW] Re: 8bit to 16 bit


--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "jnhugo"
<jacknadelle@...> wrote:
>
> Hi, I need some clarification on mode changing from 8 bit to 16 bit.
> It's my understanding that unless the file originated as a 16 bit
file
> there is no advantage to change an 8 bit (or whatever) to 16. No net
> gain. Is that right? In a specialized photoshop class there is
> instructions to convert an 8 bit file to 16 but as I recall there is
> not an advantage. Can anyone clarify? Thanks in advance.
> Jack
>
 Thanks for the replies, this list is great and I appreciate you all
taking time to respond.
So the bit depth on capture determines how detailed the information is
sampled, if an image is sampled at 8 bit changing it to 16 will not
add information that was not originally sampled at that depth.
Although Clayton is saying that he sees less destruction to the image
if he converts 8 bit to 16 bit before applying a fair amount of
adjustments. So while converting doesn't add info it does less damage
to the file when applying adjustments? Amadou, C David what about
image degrading when lots of adjustments are applied?
Thanks again
Jack



Please visit the Group Homepage to check the Files, and other resources as
they are often being updated.

http://groups.yahoo.com/group/DigitalBlackandWhiteThePrint

If you wish to receive no emails or just a daily digest, or you wish to
unsubscribe, please edit your Membership preferences by visiting this same
page.

Please follow these basic guidelines:
- As threads develop, trim off excess portions of earlier messages to keep
them short.
- Good manners are required at all time. No personal attacks or flames.
Hostile, aggressive or argumentative users may be removed from the
membership without notice.
- Keep your posts and threads related to the group topic of digital B&W
printing. Users who persistently make off-topic posts may be removed from
the membership.
- By posting on this forum you agree to abide by the group rules and
guidelines, and to abide by the actions and decisions of the group Owner and
Moderators. See \ufffdGroup Topic, Rules and Guidelines\ufffd in the Files section:
http://groups.yahoo.com/group/DigitalBlackandWhiteThePrint/files/

BY PARTICIPATING IN AND/OR POSTING MESSAGES TO THE DIGITAL BW, THE PRINT
YAHOO! GROUP YOU EXPRESSLY UNDERSTAND AND AGREE THAT THE \ufffdOWNER\ufffd AND
\ufffdMODERATORS\ufffd OF DIGITAL BW, THE PRINT YAHOO GROUP SHALL NOT BE LIABLE TO YOU
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL OR EXEMPLARY
DAMAGES, INCLUDING BUT NOT LIMITED TO, DAMAGES FOR LOSS OF PROFITS,
GOODWILL, USE, DATA OR OTHER INTANGIBLE LOSSES (EVEN IF THE  \ufffdOWNER\ufffd AND
\ufffdMODERATORS\ufffd OF DIGITAL BW, THE PRINT YAHOO GROUP HAVE BEEN ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES), RESULTING FROM: (i) THE USE OR THE INABILITY
TO USE THE DIGITAL BW, THE PRINT YAHOO GROUP; (ii) UNAUTHORIZED ACCESS TO OR
ALTERATION OF YOUR TRANSMISSIONS OR DATA; (iii) STATEMENTS OR CONDUCT OF ANY
THIRD PARTY ON THE DIGITAL BW, THE PRINT YAHOO GROUP; OR (iv) ANY OTHER
MATTER RELATING TO THE DIGITAL BW, THE PRINT YAHOO GROUP.

Yahoo! Groups Links

A little bit of computer science here.

An 8 bit BW file can have at most 256 levels of gray tones, a 16 bit
file can have 65536 independent tones. Converting from 8 to 16 bit
simply maps those original 256 tones into the 65536 space. If you had
a linear gradient going from black to white at the start it would
appear exactly the same after the conversion. You would still have 256
independent tones. However you would also have room between each tone,
256 blank spots with no gray vales on either side. By this I mean that
for the gray value that was originally 128 in the 8 bit file it would
now be 32768. The difference being that in the 8 bit file there were
picture elements represented by values 127 and 129. While in the 16
bit file those values would be represented by 32512 and 330224, there
would be no pixels in the image with values between 32768 and these
other two numbers.

Next lets say that you adjust the global contrast of your image. In a
simplified algorithm, an increase in contrast makes the values below
the midpoint (128 8-bit) darker and the values above the midpoint
lighter. In an 8 bit file if we subtract 1 from each value less than
128 then we now have 126 tones where we had 127 tones before the
subtraction. This is because the pixels that were formerly at value 1
are now at zero along with all the pixels that were previously at
zero. The same logic applies above the midpoint, so where we
originally had 256 tones before the contrast adjustment we now have
254. In a 16 bit file applying the same type of correction still
leaves you with the original 256 independent tones, because there was
256 bits of headroom between each original pixel value. You would have
to adjust the contrast by more than 255 points before you started
reducing the number of tones in your image.

The scenario I've described above leads to the posterization that we
see when making large adjustments in 8-bit files. It is not true that
all image editing leads to the loss of data, but large movements
certainly can. There are other issues with 8 vs 16 bit related to
rounding errors also.

With all that said, I rarely change my 8-bit files to 16-bit before
editing. I would only do so if there was something really wrong with
my exposure and I needed to make large adjustments to the file.

Hope I didn't confuse the issue too much.

Derek

Will any future digital camera ever shoot in 32 bit?
Mark

On Apr 10, 2007, at 11:21 AM, dealy663 wrote:

> A little bit of computer science here.
>
> An 8 bit BW file can have at most 256 levels of gray tones, a 16 bit
> file can have 65536 independent tones. Converting from 8 to 16 bit
> simply maps those original 256 tones into the 65536 space. If you had
> a linear gradient going from black to white at the start it would
> appear exactly the same after the conversion. You would still have 256
> independent tones. However you would also have room between each tone,
> 256 blank spots with no gray vales on either side. By this I mean that
> for the gray value that was originally 128 in the 8 bit file it would
> now be 32768. The difference being that in the 8 bit file there were
> picture elements represented by values 127 and 129. While in the 16
> bit file those values would be represented by 32512 and 330224, there
> would be no pixels in the image with values between 32768 and these
> other two numbers.
>
> Next lets say that you adjust the global contrast of your image. In a
> simplified algorithm, an increase in contrast makes the values below
> the midpoint (128 8-bit) darker and the values above the midpoint
> lighter. In an 8 bit file if we subtract 1 from each value less than
> 128 then we now have 126 tones where we had 127 tones before the
> subtraction. This is because the pixels that were formerly at value 1
> are now at zero along with all the pixels that were previously at
> zero. The same logic applies above the midpoint, so where we
> originally had 256 tones before the contrast adjustment we now have
> 254. In a 16 bit file applying the same type of correction still
> leaves you with the original 256 independent tones, because there was
> 256 bits of headroom between each original pixel value. You would have
> to adjust the contrast by more than 255 points before you started
> reducing the number of tones in your image.
>
> The scenario I've described above leads to the posterization that we
> see when making large adjustments in 8-bit files. It is not true that
> all image editing leads to the loss of data, but large movements
> certainly can. There are other issues with 8 vs 16 bit related to
> rounding errors also.
>
> With all that said, I rarely change my 8-bit files to 16-bit before
> editing. I would only do so if there was something really wrong with
> my exposure and I needed to make large adjustments to the file.
>
> Hope I didn't confuse the issue too much.
>
> Derek
>
>
>
> Please visit the Group Homepage to check the Files, and other  
> resources as they are often being updated.
>
> http://groups.yahoo.com/group/DigitalBlackandWhiteThePrint
>
> If you wish to receive no emails or just a daily digest, or you  
> wish to unsubscribe, please edit your Membership preferences by  
> visiting this same page.
>
> Please follow these basic guidelines:
> - As threads develop, trim off excess portions of earlier messages  
> to keep them short.
> - Good manners are required at all time. No personal attacks or  
> flames. Hostile, aggressive or argumentative users may be removed  
> from the membership without notice.
> - Keep your posts and threads related to the group topic of digital  
> B&W printing. Users who persistently make off-topic posts may be  
> removed from the membership.
> - By posting on this forum you agree to abide by the group rules  
> and guidelines, and to abide by the actions and decisions of the  
> group Owner and Moderators. See “Group Topic, Rules and Guidelines”  
> in the Files section:
> http://groups.yahoo.com/group/DigitalBlackandWhiteThePrint/files/
>
> BY PARTICIPATING IN AND/OR POSTING MESSAGES TO THE DIGITAL BW, THE  
> PRINT YAHOO! GROUP YOU EXPRESSLY UNDERSTAND AND AGREE THAT THE  
> “OWNER” AND “MODERATORS” OF DIGITAL BW, THE PRINT YAHOO GROUP SHALL  
> NOT BE LIABLE TO YOU FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,  
> CONSEQUENTIAL OR EXEMPLARY DAMAGES, INCLUDING BUT NOT LIMITED TO,  
> DAMAGES FOR LOSS OF PROFITS, GOODWILL, USE, DATA OR OTHER  
> INTANGIBLE LOSSES (EVEN IF THE  “OWNER” AND “MODERATORS” OF DIGITAL  
> BW, THE PRINT YAHOO GROUP HAVE BEEN ADVISED OF THE POSSIBILITY OF  
> SUCH DAMAGES), RESULTING FROM: (i) THE USE OR THE INABILITY TO USE  
> THE DIGITAL BW, THE PRINT YAHOO GROUP; (ii) UNAUTHORIZED ACCESS TO  
> OR ALTERATION OF YOUR TRANSMISSIONS OR DATA; (iii) STATEMENTS OR  
> CONDUCT OF ANY THIRD PARTY ON THE DIGITAL BW, THE PRINT YAHOO  
> GROUP; OR (iv) ANY OTHER MATTER RELATING TO THE DIGITAL BW, THE  
> PRINT YAHOO GROUP.
>
> Yahoo! Groups Links
>
>
>
>
>

Derek,

>A little bit of computer science here.

>... lets say that you adjust the global contrast of your image. ...
> In an 8 bit file if we subtract 1 from each value less 
> than 128 then we now have 126 tones where we had 127 tones 
> before the subtraction. This is because the pixels that were 
> formerly at value 1 are now at zero ...

If you'd used curves for this adjustment, would there have been any loss?

Paul
www.PaulRoark.com

I doubt there will ever be a need for a 32 bit camera. As things stand
right now I don't believe that there is much in the way of an
appreciable difference whether the final image is in 8 or 16 bit
values. 4294967296 (2 to the 32nd power) independent tones in an image
is complete overkill and would ridiculously bloat the size of the
files even further. As things stand right now I believe that most high
end digital cameras capture a 12 bit image 4096 independent tones.

Paul, whether the redistribution of the pixels is done by curves, the
contrast slider or other tool makes no difference. If your adjustment
causes a change which exceeds the amount of headroom between each
discrete tonal value then you are going to reduce the number of
independent tones in your image. In an image that begins with 256
tones in an 8-bit space then your headroom is 0, while in a 16-bit
space your headroom is 255. An image that begins with 128 (evenly
distributed) tones in an 8-bit space has 1 bit of headroom between
values, while in 16-bit land it would have 512 bits of headroom.

Please keep in mind that I've massively simplified how these things
work in the real world. A contrast adjustment is not likely to be a
simple subtraction/addition of a single value to all pixels in the
image not equal to the midpoint.

Maybe I should take the time to write a full fledged article on how
this stuff works.
--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Paul Roark"
<paul.roark@...> wrote:
>
> Derek,
> 
> >A little bit of computer science here.
> 
> >... lets say that you adjust the global contrast of your image. ...
> > In an 8 bit file if we subtract 1 from each value less 
> > than 128 then we now have 126 tones where we had 127 tones 
> > before the subtraction. This is because the pixels that were 
> > formerly at value 1 are now at zero ...
> 
> If you'd used curves for this adjustment, would there have been any
loss?

> 
> Paul
> www.PaulRoark.com
>

Hello Derek,

>A little bit of computer science here...
>Converting from 8 to 16 bit simply maps those original 256 tones 
>into the 65536 space...
>There are other issues with 8 vs 16 bit related to rounding errors 
>also.

Thanks very much, it makes good sense.


>The scenario I've described above leads to the posterization that we
>see when making large adjustments in 8-bit files...

As I mentioned before, I did see in a large smooth area containing
subtle transitions some negative affects that went away in the 16-bit
version.  It convinced me that there is some merit to the technique,
at least for smooth tone areas.


>With all that said, I rarely change my 8-bit files to 16-bit before
>editing. I would only do so if there was something really wrong with
>my exposure and I needed to make large adjustments to the file.

Since I don't know beforehand what sorts of manipulations I'll be
doing, I change to 16-bit any file I'm going to convert to grayscale
for a BW print.  The added file size for me is a small price to pay
for the insurance of "overhead" if it's needed (and because it's
grayscale it actually ends up being 1/3 smaller than the 8-bit color
version).  

I use the pocket cam frequently for various things and occasionally
find an image that's worthy of a print.  It's already at a
disadvantage not being RAW, so I want to give it all the help I can.

Thanks for the good explanation.


Regards,
Clayton


Info on black and white digital printing at    
http://www.cjcom.net/digiprnarts.htm

Derek,

>If your adjustment causes a change which exceeds the amount of 
>headroom between each discrete tonal value then you are going to 
>reduce the number of independent tones in your image. 

Isn't that what we're seeing in a badly combed histogram?

Regards,
Clayton


Info on black and white digital printing at    
http://www.cjcom.net/digiprnarts.htm

So, if my understanding is correct, 

8-bit rgb = 256 levels x 3 channels = 768 levels (if converted to b&w)

or the equivalent of 9.5 greyscale bits. 

Converting the 8-bit file to a 16-bit file produces 64k possible
levels, which will be initially discretely distributed into 256
levels, i.e. each pixel color will have a value equal to 256*x, 
where x is the original 8-bit pixel value. So just after the
conversion the information is the same in both the 8- and 16-bit
files. The numbers are just bigger but you still have use 256 levels
per color.

However, after the conversion, you can adjust the pixel values more
finely because you have the 64k levels.  So processing the image in
16-bits should create less artifacts that doing so in 8-bits. The
difference might be visible depending on how good your sight is 
and on the amount of processing.

Right?

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "dealy663"
<dealy663@...> wrote:

>
> A little bit of computer science here.
> 
> An 8 bit BW file can have at most 256 levels of gray tones, a 16 bit
> file can have 65536 independent tones. Converting from 8 to 16 bit
> simply maps those original 256 tones into the 65536 space. If you had
> a linear gradient going from black to white at the start it would
> appear exactly the same after the conversion. You would still have 256
> independent tones. However you would also have room between each tone,
> 256 blank spots with no gray vales on either side. By this I mean that
> for the gray value that was originally 128 in the 8 bit file it would
> now be 32768. The difference being that in the 8 bit file there were
> picture elements represented by values 127 and 129. While in the 16
> bit file those values would be represented by 32512 and 330224, there
> would be no pixels in the image with values between 32768 and these
> other two numbers.
> 
> Next lets say that you adjust the global contrast of your image. In a
> simplified algorithm, an increase in contrast makes the values below
> the midpoint (128 8-bit) darker and the values above the midpoint
> lighter. In an 8 bit file if we subtract 1 from each value less than
> 128 then we now have 126 tones where we had 127 tones before the
> subtraction. This is because the pixels that were formerly at value 1
> are now at zero along with all the pixels that were previously at
> zero. The same logic applies above the midpoint, so where we
> originally had 256 tones before the contrast adjustment we now have
> 254. In a 16 bit file applying the same type of correction still
> leaves you with the original 256 independent tones, because there was
> 256 bits of headroom between each original pixel value. You would have
> to adjust the contrast by more than 255 points before you started
> reducing the number of tones in your image.
> 
> The scenario I've described above leads to the posterization that we
> see when making large adjustments in 8-bit files. It is not true that
> all image editing leads to the loss of data, but large movements
> certainly can. There are other issues with 8 vs 16 bit related to
> rounding errors also.
> 
> With all that said, I rarely change my 8-bit files to 16-bit before
> editing. I would only do so if there was something really wrong with
> my exposure and I needed to make large adjustments to the file.
> 
> Hope I didn't confuse the issue too much.
> 
> Derek
>

Clayton,
I ran the gist of the early part of this thread by some Adobe
engineers and beta testers. Aside from the standard "no benefit"
response (of which I was guilty as well), those who had actually tried
it concurred with your results, that in some cases bumping from 8 to
16bits before large edits yielded a benefit. The explanations were
more anecdotal, than an official Adobe position on the matter, but
thanks for putting this information out there. If I get any definitive
technical explanations from Adobe I'll pass them on.
-- 
amadou diallo
Author, Mastering Digital Black and White
www.masteringdigitalbwbook.com

Hello,

I've been following this debate with interest and would like to point out 
the following:

Canon just came out with a 1D Mark III which is 14 bits thus provides 16384 
possible tones.

Specialized equipment such as Astronomy cameras already produce 16 bit 
output thus 65536 possible tones. Some models even provide special features 
to prevent filling up the wells in order to manage extreme dynamic range 
issues.

As typical photographers (amateur or pro) we may never need to go beyond 12, 
14 or 16 bits but keeping an eye on specialized equipment and potential 
directions is always interesting.

Most importantly coding application to ensure you can support new wider 
formats without having to recode everything is important regardless of how 
the implementation is done.

I earn a living as a programmer and I have been involved recently in project 
to access our ability to undergo a porting activity from 32 bit code to 64 
bits. Now compound this for operating systems across windows and unix who do 
not handle the bit depth in the same way.

A lot of the code was written based on the idea that we will never ever need 
more than 32 bits to express something in and as always a few years later we 
are stuck with the results of making assumptions that didn't hold up to the 
test of time.

Who know what's next but I am sure enjoying the ride and its possible future 
directions.

Cheers,
Pierre

>From: "dealy663" <dealy663@...>
>Reply-To: DigitalBlackandWhiteThePrint@yahoogroups.com
>To: DigitalBlackandWhiteThePrint@yahoogroups.com
>Subject: [Digital BW] Re: 8bit to 16 bit
>Date: Tue, 10 Apr 2007 18:24:53 -0000
>
>I doubt there will ever be a need for a 32 bit camera. As things stand
>right now I don't believe that there is much in the way of an
>appreciable difference whether the final image is in 8 or 16 bit
>values. 4294967296 (2 to the 32nd power) independent tones in an image
>is complete overkill and would ridiculously bloat the size of the
>files even further. As things stand right now I believe that most high
>end digital cameras capture a 12 bit image 4096 independent tones.
>
>Paul, whether the redistribution of the pixels is done by curves, the
>contrast slider or other tool makes no difference. If your adjustment
>causes a change which exceeds the amount of headroom between each
>discrete tonal value then you are going to reduce the number of
>independent tones in your image. In an image that begins with 256
>tones in an 8-bit space then your headroom is 0, while in a 16-bit
>space your headroom is 255. An image that begins with 128 (evenly
>distributed) tones in an 8-bit space has 1 bit of headroom between
>values, while in 16-bit land it would have 512 bits of headroom.
>
>Please keep in mind that I've massively simplified how these things
>work in the real world. A contrast adjustment is not likely to be a
>simple subtraction/addition of a single value to all pixels in the
>image not equal to the midpoint.
>
>Maybe I should take the time to write a full fledged article on how
>this stuff works.

_________________________________________________________________
RealLiveMoms: Share your experience with Real Live Moms just like you 
http://www.reallivemoms.ca/

Part of the problem here is in thinking linearly. In fact, the way the 256  
steps of an 8-bit image are distributed are logarithmic and a great deal  
hinges on the dynamic range of the image. 128 steps of the total 256 steps  
are all used for the brightest 1 f-stop of the image. The next f-stop has  
only 64 steps, the next stop has 32. At the 5th f-stop you are down to  
only 8 steps. Fortunately our eyes are less sensitive to tonal changes in  
the darker areas that in the brighter areas. Never-the-less, with a 16-bit  
file, the 5th f-stop range has 2048 possible steps and you can therefore  
manipulate the data with less chance of introducing errors than with only  
8 steps available. Also, even though you end up printing in an 8 bit  
environment, that does not automatically limit you to 256 steps. As Roy  
Harrington has pointed out in his many discussions on the QTR mailing  
list, you can still get a smoother tonal range in the final print through  
the use of dithering if you have more than 8 bits of data available.
Steve

On Tue, 10 Apr 2007 12:12:22 -0700, amadou diallo <amadiallo@...>  
wrote:

> Clayton,
> I ran the gist of the early part of this thread by some Adobe
> engineers and beta testers. Aside from the standard "no benefit"
> response (of which I was guilty as well), those who had actually tried
> it concurred with your results, that in some cases bumping from 8 to
> 16bits before large edits yielded a benefit. The explanations were
> more anecdotal, than an official Adobe position on the matter, but
> thanks for putting this information out there. If I get any definitive
> technical explanations from Adobe I'll pass them on.



-- 
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/

I can without a doubt, stand behind converting from 8 to 16 for EXTENSIVE
editing on smooth areas. I have seen this over all my years of working with
PS. (not that that has been  a life time but.7 years)   I found it to be a
questionable trade of before PS could handle 16 bit in many of the more
commonly used filters. There are of course times when even changing will not
help, and it may be more obvious to users of different out put devices and
output controls.

 

Eric 

 

Eric Neilsen Photography

4101 Commerce Street, Suite 9

Dallas, TX 75226

214-827-8301

http://ericneilsenphotography.com

 

Skype : ejprinter

  _____  

From: DigitalBlackandWhiteThePrint@yahoogroups.com
[mailto:DigitalBlackandWhiteThePrint@yahoogroups.com] On Behalf Of amadou
diallo
Sent: Tuesday, April 10, 2007 2:12 PM
To: DigitalBlackandWhiteThePrint@yahoogroups.com
Subject: Re: [Digital BW] Re: 8bit to 16 bit

 

Clayton,
I ran the gist of the early part of this thread by some Adobe
engineers and beta testers. Aside from the standard "no benefit"
response (of which I was guilty as well), those who had actually tried
it concurred with your results, that in some cases bumping from 8 to
16bits before large edits yielded a benefit. The explanations were
more anecdotal, than an official Adobe position on the matter, but
thanks for putting this information out there. If I get any definitive
technical explanations from Adobe I'll pass them on.
-- 
amadou diallo
Author, Mastering Digital Black and White
www.masteringdigitalbwbook.com

 



[Non-text portions of this message have been removed]

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Manuel Toledo
Quinones" <mtoledo@...> wrote:
>
> So, if my understanding is correct, 
> 
> 8-bit rgb = 256 levels x 3 channels = 768 levels (if converted to b&w)
> 


Actually, the above is incorrect. A color file that is in 8-bit rgb
mode converted to BW only has 256 tones (levels) even if still in rgb
mode. This is because all three of the rgb values become the same for
each pixel (r = b = g in that scenario).

Clayton, yes I do believe that posterization and histogram combing are
diretly related to one another. Also Clayton, I sent you a direct
email with questions about a replacement for Epson enhanced matte
(Kayenta?) yesterday. Maybe it went to your spam folder?

Derek

8 bits is two bucks. 16 bits is four bucks. End of discussion. :-)

Rick

Derek,

I believe this discussion started with the question of whether there is
anything to be gained from converting 8 bit to 16 bit.  So, my comments are
in that context, and I believe there are significant advantages to such a
conversion if one is going to do any manipulation of the image.

>... whether the redistribution of the pixels is done by curves, 
> the contrast slider or other tool makes no difference. 
> If your adjustment causes a change which exceeds the amount 
> of headroom between each discrete tonal value then you 
> are going to reduce the number of independent tones in your image.

No question there. But it misses the 2 issue I wanted to point out.

First, if one starts with 8 bit and converts to 16 bit, there is, as you
later point out, enormous headroom.  This allows one to do significant
adjustments with no or virtually no loss of information -- if the right
tools are used.

Second, the "Brightness/Contrast" adjustment is different than the curves.
With the former I believe you slide the entire image scale up or down and
are much more likely to push things out the end, with a resulting loss of
information.  With the curves, you'll compress one side and stretch out the
other, with a much lower chance of losing information.  Again, this is in
the context of an 8 bit file that has been converted to 16 bit such that it
has headroom to absorb the compression without loss of information.

Paul
www.PaulRoark.com

-----Original Message-----

From: DigitalBlackandWhiteThePrint@yahoogroups.com
[mailto:DigitalBlackandWhiteThePrint@yahoogroups.com]On Behalf Of Manuel
Toledo Quinones
Sent: Tuesday, April 10, 2007 12:12 PM
To: DigitalBlackandWhiteThePrint@yahoogroups.com
Subject: [Digital BW] Re: 8bit to 16 bit


So, if my understanding is correct,

8-bit rgb = 256 levels x 3 channels = 768 levels (if converted to b&w)

or the equivalent of 9.5 greyscale bits.

Converting the 8-bit file to a 16-bit file produces 64k possible
levels, which will be initially discretely distributed into 256
levels, i.e. each pixel color will have a value equal to 256*x,
where x is the original 8-bit pixel value. So just after the
conversion the information is the same in both the 8- and 16-bit
files. The numbers are just bigger but you still have use 256 levels
per color.

However, after the conversion, you can adjust the pixel values more
finely because you have the 64k levels.  So processing the image in
16-bits should create less artifacts that doing so in 8-bits. The
difference might be visible depending on how good your sight is
and on the amount of processing.

Right?

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "dealy663"
<dealy663@...> wrote:
>
> A little bit of computer science here.
>
> An 8 bit BW file can have at most 256 levels of gray tones, a 16 bit
> file can have 65536 independent tones. Converting from 8 to 16 bit
> simply maps those original 256 tones into the 65536 space. If you had
> a linear gradient going from black to white at the start it would
> appear exactly the same after the conversion. You would still have 256
> independent tones. However you would also have room between each tone,
> 256 blank spots with no gray vales on either side. By this I mean that
> for the gray value that was originally 128 in the 8 bit file it would
> now be 32768. The difference being that in the 8 bit file there were
> picture elements represented by values 127 and 129. While in the 16
> bit file those values would be represented by 32512 and 330224, there
> would be no pixels in the image with values between 32768 and these
> other two numbers.
>
> Next lets say that you adjust the global contrast of your image. In a
> simplified algorithm, an increase in contrast makes the values below
> the midpoint (128 8-bit) darker and the values above the midpoint
> lighter. In an 8 bit file if we subtract 1 from each value less than
> 128 then we now have 126 tones where we had 127 tones before the
> subtraction. This is because the pixels that were formerly at value 1
> are now at zero along with all the pixels that were previously at
> zero. The same logic applies above the midpoint, so where we
> originally had 256 tones before the contrast adjustment we now have
> 254. In a 16 bit file applying the same type of correction still
> leaves you with the original 256 independent tones, because there was
> 256 bits of headroom between each original pixel value. You would have
> to adjust the contrast by more than 255 points before you started
> reducing the number of tones in your image.
>
> The scenario I've described above leads to the posterization that we
> see when making large adjustments in 8-bit files. It is not true that
> all image editing leads to the loss of data, but large movements
> certainly can. There are other issues with 8 vs 16 bit related to
> rounding errors also.
>
> With all that said, I rarely change my 8-bit files to 16-bit before
> editing. I would only do so if there was something really wrong with
> my exposure and I needed to make large adjustments to the file.
>
> Hope I didn't confuse the issue too much.
>
> Derek
>




Please visit the Group Homepage to check the Files, and other resources as
they are often being updated.

http://groups.yahoo.com/group/DigitalBlackandWhiteThePrint

If you wish to receive no emails or just a daily digest, or you wish to
unsubscribe, please edit your Membership preferences by visiting this same
page.

Please follow these basic guidelines:
- As threads develop, trim off excess portions of earlier messages to keep
them short.
- Good manners are required at all time. No personal attacks or flames.
Hostile, aggressive or argumentative users may be removed from the
membership without notice.
- Keep your posts and threads related to the group topic of digital B&W
printing. Users who persistently make off-topic posts may be removed from
the membership.
- By posting on this forum you agree to abide by the group rules and
guidelines, and to abide by the actions and decisions of the group Owner and
Moderators. See \ufffdGroup Topic, Rules and Guidelines\ufffd in the Files section:
http://groups.yahoo.com/group/DigitalBlackandWhiteThePrint/files/

BY PARTICIPATING IN AND/OR POSTING MESSAGES TO THE DIGITAL BW, THE PRINT
YAHOO! GROUP YOU EXPRESSLY UNDERSTAND AND AGREE THAT THE \ufffdOWNER\ufffd AND
\ufffdMODERATORS\ufffd OF DIGITAL BW, THE PRINT YAHOO GROUP SHALL NOT BE LIABLE TO YOU
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL OR EXEMPLARY
DAMAGES, INCLUDING BUT NOT LIMITED TO, DAMAGES FOR LOSS OF PROFITS,
GOODWILL, USE, DATA OR OTHER INTANGIBLE LOSSES (EVEN IF THE  \ufffdOWNER\ufffd AND
\ufffdMODERATORS\ufffd OF DIGITAL BW, THE PRINT YAHOO GROUP HAVE BEEN ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES), RESULTING FROM: (i) THE USE OR THE INABILITY
TO USE THE DIGITAL BW, THE PRINT YAHOO GROUP; (ii) UNAUTHORIZED ACCESS TO OR
ALTERATION OF YOUR TRANSMISSIONS OR DATA; (iii) STATEMENTS OR CONDUCT OF ANY
THIRD PARTY ON THE DIGITAL BW, THE PRINT YAHOO GROUP; OR (iv) ANY OTHER
MATTER RELATING TO THE DIGITAL BW, THE PRINT YAHOO GROUP.

Yahoo! Groups Links

Ya? Rick,

But which is better?  :  )

In any case, the monitor doesn't display all the data. Only the programmers
so-called "best representation" - which I need like another hole in the
head.

gar

-----Original Message-----
From: DigitalBlackandWhiteThePrint@yahoogroups.com
[mailto:DigitalBlackandWhiteThePrint@yahoogroups.com]On Behalf Of Rick
Colson
Sent: Tuesday, April 10, 2007 1:11 PM
To: DigitalBlackandWhiteThePrint@yahoogroups.com
Subject: [Digital BW] Re: 8bit to 16 bit


8 bits is two bucks. 16 bits is four bucks. End of discussion. :-)

Rick




Please visit the Group Homepage to check the Files, and other resources as
they are often being updated.

http://groups.yahoo.com/group/DigitalBlackandWhiteThePrint

If you wish to receive no emails or just a daily digest, or you wish to
unsubscribe, please edit your Membership preferences by visiting this same
page.

Please follow these basic guidelines:
- As threads develop, trim off excess portions of earlier messages to keep
them short.
- Good manners are required at all time. No personal attacks or flames.
Hostile, aggressive or argumentative users may be removed from the
membership without notice.
- Keep your posts and threads related to the group topic of digital B&W
printing. Users who persistently make off-topic posts may be removed from
the membership.
- By posting on this forum you agree to abide by the group rules and
guidelines, and to abide by the actions and decisions of the group Owner and
Moderators. See \ufffdGroup Topic, Rules and Guidelines\ufffd in the Files section:
http://groups.yahoo.com/group/DigitalBlackandWhiteThePrint/files/

BY PARTICIPATING IN AND/OR POSTING MESSAGES TO THE DIGITAL BW, THE PRINT
YAHOO! GROUP YOU EXPRESSLY UNDERSTAND AND AGREE THAT THE \ufffdOWNER\ufffd AND
\ufffdMODERATORS\ufffd OF DIGITAL BW, THE PRINT YAHOO GROUP SHALL NOT BE LIABLE TO YOU
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL OR EXEMPLARY
DAMAGES, INCLUDING BUT NOT LIMITED TO, DAMAGES FOR LOSS OF PROFITS,
GOODWILL, USE, DATA OR OTHER INTANGIBLE LOSSES (EVEN IF THE  \ufffdOWNER\ufffd AND
\ufffdMODERATORS\ufffd OF DIGITAL BW, THE PRINT YAHOO GROUP HAVE BEEN ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES), RESULTING FROM: (i) THE USE OR THE INABILITY
TO USE THE DIGITAL BW, THE PRINT YAHOO GROUP; (ii) UNAUTHORIZED ACCESS TO OR
ALTERATION OF YOUR TRANSMISSIONS OR DATA; (iii) STATEMENTS OR CONDUCT OF ANY
THIRD PARTY ON THE DIGITAL BW, THE PRINT YAHOO GROUP; OR (iv) ANY OTHER
MATTER RELATING TO THE DIGITAL BW, THE PRINT YAHOO GROUP.

Yahoo! Groups Links

Paul,

You can see how some of these things works just by opening an image
and looking at the histogram.

Bring up the brightness/contrast slider and watch when you make a
brightness adjustment. Here all that is happening is a fixed value
(either positive or negative) is being added to every pixel value in
the image hence the histogram slides as a whole up and down the range.

The contrast slider and the curves slider can be thought of as
behaving in a similar way. As you increase the contrast the data will
fall off the sides of the histogram as more pixels are assigned values
0 and 255 and the number of tones represented in your image decreases.
You can see this as evidenced by the walls that start going up on the
left and right sides of the histogram while the number of pixels
making up the remaining tones in your image decreases the middle of
your histogram begins getting lower and lower. Make a large enough
adjustment and you will be left with just 2 tones, black and white
with no data in the middle of the histogram.

In reality the contrast adjustment is likely more elegant than the
simple scenario I described in my first missive on this subject. You
can get a picture of how it works by doing something similar with the
curves control. Place 3 anchor points equidistant on the curves
control with one in the center, resulting in the line being divided up
into four equal length segments. Now take the two non-centered anchors
and move them equal but opposite distances (resulting in an s shaped
line) and you'll see the histogram behave in exactly the same way as
with the contrast slider. The curves control can dump pixels just as
quickly as the contrast control. The big difference is that with
curves you have much more say in the re-distribution of those pixels
as opposed to the programmer that coded up the contrast implementation.

I didn't mean to say that going to 16-bit doesn't allow you more room
with editing at all. My comment at the end of my first email on the
subject was merely to state that I don't do it very often, simply
because 1. I rarely shoot 8-bit and 2. when I do I rarely am off by so
much that I need to make large adjustments to the 8-bit file.

Derek

P.S. I've been a pretty quiet lurker on this list for a couple of
years now. Its good finally contribute something in a substantial way.

> Second, the "Brightness/Contrast" adjustment is different than the
curves.
> With the former I believe you slide the entire image scale up or
down and
> are much more likely to push things out the end, with a resulting
loss of
> information.  With the curves, you'll compress one side and stretch
out the
> other, with a much lower chance of losing information.  Again, this
is in
> the context of an 8 bit file that has been converted to 16 bit such
that it

> has headroom to absorb the compression without loss of information.
> 
> Paul
> www.PaulRoark.com
>

Derek,

> Clayton, yes I do believe that posterization and histogram combing are
> diretly related to one another. Also Clayton, I sent you a direct
> email with questions about a replacement for Epson enhanced matte
> (Kayenta?) yesterday. Maybe it went to your spam folder?

I got it and sent a reply at 4:28 pm (EDT).  Let me know if you can't
find it and I'll resend.


Regards,
Clayton


Info on black and white digital printing at    
http://www.cjcom.net/digiprnarts.htm

> I got it and sent a reply at 4:28 pm (EDT).  Let me know if you can't
> find it and I'll resend.
> 
> 
> Regards,
> Clayton
> 

Hi Clayton,

Sorry, but for some reason it didn't go through to my account at
hotmail. My main real email address is dme AT grandprixsw DOT com.
Whatever you send should make it to that account.

Thanks, Derek