Re: jump
Re: jump
- Subject: Re: jump
- From: Klaus Karcher <email@hidden>
- Date: Fri, 23 Nov 2007 02:44:51 +0100
eugene appert wrote:
There is something I don't understand about "applies your working
grayscale space".
It's hard for me to explain, but maybe you want to conduct an experiment:
Make a new Lab document, fill it with Lab 50/0/0 and set the first color
readout in the info palette to actual color and the second to Lab.
When you open the channels palette, you will notice that the L* channel
looks different (typically brighter) than the composite image.
Select the composite image. When you use the eyedropper tool, the info
palette will show Lab=50/0/0 in both halves.
Now click onto the L* channel in the channels palette and use the
eyedropper tool again: The first color readout will show 50% K now and
the second readout will show something *different* than Lab=50/0/0 (e.g.
L=54 when your grayscale working space is set to Gamma 2,2). When you
change the grayscale working space, the L* value will change, but the K
value will still be 50%: Photoshop did not change the file data: there
is still the same gray level (128, i.e. 50%) in the file, but Photoshop
interprets it differently.
I would have thought that the file data and not the L* would be
affected by the target working space. Such that though L*50 is
represented in a 1.8 space as 61% and as 54% in a 2.2 space, the L*
value remains the same.
That's true when you /convert/ e.g. L=50 to gray gamma 1.8 or 2.2.
Unless you mean the info pallet applies the algorithm normally
employed by Photoshop for mode conversions (.3 R + .59G + .11 B).
No: the info palette also uses the (embedded- or working space-)
profiles. But since a single channel has no embedded profile, Photoshop
uses the grayscale working space to interpret its contents.
But
then why would the anticipated conversion values that the info
palette displays change simply by extracting the a* and b* channels?
Photoshop no longer assumes a Lab but a grayscale image.
I am using the Gretag color chart as reference. If I convert the Lab
file to grayscale via Photoshop menu mode conversion, I will always
preserve the L* values, no matter what grayscale space is configured
by default in my color settings.
Yes, but this L* value is represented by different levels in different
colorspaces.
Equally the info pallet will
display the exact L* values of the Gretag chart regardless of the
grayscale space configured in my color settings.
... as long as the composite image is selected in the channels palette.
Yet if I deactivate
the a* and b* channels the file brightens visibly on the screen, even
though the info palette continues to display the original L* values.
The remaining L* channel will be shown as grayscale image on the screen.
As long as the other channels are only hidden, but still selected, the
info palette still shows the composite values.
Klaus Karcher wrote:
I made a custom dot gain curve that maps Y to L [...]
I expected to get L=50 for K=50 in the info panel with this curve ...
but I got L=56. ... curious. Can someone explain what's going on?
okay, I've got it now: If you want a grayscale working space that won't
change the lightness, you need a dot gain curve that maps K to 100-Y(L)
where L=100-K (i.e. 0%K <=> L=100; 100%K <=> L=0; 50%K <=> L=50 <=>
Y=18.4 => output=81.6)
+-----+-------+
| in | out |
+-----+-------+
| 0 | 0.0 |
| 2 | 5.1 |
| 4 | 10.0 |
| 6 | 14.7 |
| 8 | 19.3 |
| 10 | 23.7 |
| 20 | 43.3 |
| 30 | 59.3 |
| 40 | 71.9 |
| 50 | 81.6 |
| 60 | 88.7 |
| 70 | 93.8 |
| 80 | 97.0 |
| 90 | 98.9 |
| 100 | 100.0 |
+-----+-------+
Please note: this is *not* my recommendation for a production
workstation, but my setting to find out how Photoshop works.
Klaus
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Colorsync-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden
| References: | |
| | >re:jump (From: "eugene appert" <email@hidden>) |