| Author | Thread |
|
|
11/14/2003 01:35:45 PM · #26 |
Originally posted by kiwiness:
Thanx for the post Pedro. Next Monday, Tuesday and Wednesday I am going to a seminar on Photoshop CS combined with a photography project, so I'll be learning about all the new features. So far from what I have seen of the new version, I am impressed. |
I'm jealous Gary!
Have lots of fun...
JD
|
|
|
|
11/14/2003 01:38:04 PM · #27 |
This is the first I've seen this option. It was in other versions?
My PS7 didn't have it.
that's weird. |
|
|
|
11/14/2003 01:42:00 PM · #28 |
Originally posted by JC Homola:
This is the first I've seen this option. It was in other versions?
My PS7 didn't have it.
that's weird. |
Its under Image... Adjust....Variations - at least in my Photoshop 7 and my Photoshop 5 LE
Just yet another in a long line of stuff that's hard to find and non-intuiative in Adobe products :) Photoshop is a monster
Message edited by author 2003-11-14 13:42:38.
|
|
|
|
11/14/2003 01:42:49 PM · #29 |
 This image is adjusted w/ USM, auto levels/contrast/colour.
 Is using one of the new "photo filter" options. warm
 Is "underwater" photo filter.
All of these are adjustable for density or colour. There are 19 different options.
|
|
|
|
11/14/2003 01:59:17 PM · #30 |
Does that new version of PS work with out-of-the-camera RAW images from Canon? i.e. CRW files?
|
|
|
|
11/14/2003 02:12:36 PM · #31 |
Speaking of photoshop stuff, I just found //www.photoworkshop.com seems to have a load of interesting content
|
|
|
|
11/14/2003 02:30:37 PM · #32 |
Originally posted by cykhansen:
Does that new version of PS work with out-of-the-camera RAW images from Canon? i.e. CRW files? |
yes
|
|
|
|
11/14/2003 03:49:16 PM · #33 |
One of my favourite features is Layer Comp (found at the top menu bar on the right, tabbed section). This brilliant item allows you to save at-a-glance layer combinations which is ideal for web site design and busy documents.
Basically, you can set up a layer set up for a Home Page for example, save a Layer Comp then enable all the layers for the Services page and save another Layer Comp. You can then toggle between them by clicking the < > buttons, which allows immediate comparison of mutiple layers and layer sets.
The advantages for image manipulation and photo adjustment are limitless. |
|
|
|
11/14/2003 04:25:30 PM · #34 |
AMEN!!!
Originally posted by ronners:
#1 reason for upgrading to PhotoShop CS... 16-bit editing.
Ron. |
|
|
|
|
11/14/2003 04:43:19 PM · #35 |
Originally posted by ronners:
#1 reason for upgrading to PhotoShop CS... 16-bit editing.
Ron. |
Originally posted by Patella:
AMEN!!! |
... say the disk and data-storage media makers.
Some questions about 16-bit color:
1). If my camera only captures 8-bit color there's no point in converting to edit, is there?
2). How can I tell the difference if I don't have a 16-bit monitor?
3). What happens to the colors when it's saved back to 8-bit JPEG for printing?
4). I have a scanner which will work in 12-bit color. What do I do with that? |
|
|
|
11/14/2003 05:10:35 PM · #36 |
Originally posted by Gordon:
Originally posted by JC Homola:
This is the first I've seen this option. It was in other versions?
My PS7 didn't have it.
that's weird. |
Its under Image... Adjust....Variations - at least in my Photoshop 7 and my Photoshop 5 LE
Just yet another in a long line of stuff that's hard to find and non-intuiative in Adobe products :) Photoshop is a monster |
Nope not in mine!
How bizarre. |
|
|
|
11/14/2003 05:30:43 PM · #37 |
Originally posted by GeneralE:
... say the disk and data-storage media makers.
Some questions about 16-bit color:
1). If my camera only captures 8-bit color there's no point in converting to edit, is there?
2). How can I tell the difference if I don't have a 16-bit monitor?
3). What happens to the colors when it's saved back to 8-bit JPEG for printing?
4). I have a scanner which will work in 12-bit color. What do I do with that? |
Yes, file sizes will double -- I am MORE than willing to deal with that -- looking into DVD-drives now...
In answer to your questions:
1. No, there is no practical reason to convert to 16-bit from an 8-bit original -- at least not that I am aware of. (I'm sure someone will prove me wrong just for the sake of it.)
2. Well, the easiest way you can "tell" the difference needs a 16-bit file to play with. Make a duplicate of the original file, convert to 8-bit, and keep both 8-bit and 16-bit original up and on the screen. Now, using either curves or levels, make a fairly radical adjustment to both that makes either highlights WAY too dark, or shadows WAY too bright. Now, using your favorite method, compare the histograms. On the 16-bit, you will still have a nice smooth curve. On the 8-bit you will see "gapping" or "combing" -- areas where the image has gaps between values. The more gapping you have, the less data you have to work with/print/etc.
3. 16-bit mode is basically for editing. 8-bit is basically for printing (exterme generalization here). So, you actually want to convert it to 8-bit mode before printing. (But I'd hesitate to convert it to jpg -- unless you mean an online printer?)
4. A scanner that works in 12 bit? I'm not sure about that. I know I can scan my negs in 16-bit mode and will do so as soon as I get my hands on CS and have a reason to do so -- and finally be a LOT more happy with camera RAW and its 16-bit mode now that I can edit the RAW images.
Message edited by author 2003-11-14 17:31:46. |
|
|
|
11/14/2003 05:35:28 PM · #38 |
Thanks -- helpful as always. I'll double-check that scanner setting ...
Yes I'm talking about JPEG for online/local print facilities which only take that format; otherwise I'd use TIFF. I always keep all the file formats anyway.
How is this different from the 16-bit mode in PS 5.0? |
|
|
|
11/14/2003 05:43:24 PM · #39 |
Originally posted by Patella:
In answer to your questions:
1. No, there is no practical reason to convert to 16-bit from an 8-bit original -- at least not that I am aware of. (I'm sure someone will prove me wrong just for the sake of it.)
2. Well, the easiest way you can "tell" the difference needs a 16-bit file to play with. Make a duplicate of the original file, convert to 8-bit, and keep both 8-bit and 16-bit original up and on the screen. Now, using either curves or levels, make a fairly radical adjustment to both that makes either highlights WAY too dark, or shadows WAY too bright. Now, using your favorite method, compare the histograms. On the 16-bit, you will still have a nice smooth curve. On the 8-bit you will see "gapping" or "combing" -- areas where the image has gaps between values. The more gapping you have, the less data you have to work with/print/etc. |
Answer 2 actually answers (and contradicts) 1.
|
|
|
|
11/14/2003 05:45:54 PM · #40 |
Yes, you edit in the 16-bit space and then convert to 8-bit jpegs then for printing.
Editing in 16-bit in previous versions of Photoshop is, if not impossible, extremely difficult. Convert an image into 16-bit mode and see what happens. Suddenly all your layers and blend modes are impossible to access. 75% of filters are also "gone."
Half the reason some of the convoluted sharpening/contrast/etc methods exist is as work-arounds to the 16-bit accessibility issue. They aren't necessarily inherently better than another, easier technique -- they just work in a 16-bit space.
With CS, there isn't an accessinbility issue -- it all works. Hooray! |
|
|
|
11/14/2003 05:48:50 PM · #41 |
Originally posted by Gordon:
Answer 2 actually answers (and contradicts) 1. |
As far as I've learned, Gordon, that's not true. Converting to 16-bit from an 8-bit original doesn't suddenly conjure up all that extra color depth.
If you do the same test, but instead of converting a 16-bit file to 8-bit, you start with an 8-bit original and convert the duplicate from 8 to 16, then both histograms will have the same gapping problem.
Message edited by author 2003-11-14 17:55:47. |
|
|
|
11/14/2003 06:19:37 PM · #42 |
Originally posted by Patella:
Originally posted by Gordon:
Answer 2 actually answers (and contradicts) 1. |
As far as I've learned, Gordon, that's not true. Converting to 16-bit from an 8-bit original doesn't suddenly conjure up all that extra color depth.
If you do the same test, but instead of converting a 16-bit file to 8-bit, you start with an 8-bit original and convert the duplicate from 8 to 16, then both histograms will have the same gapping problem. |
Yes - I realise that - you don't create the extra information. What you do enable is more resolution when you do manipulations. I'm having some problem getting my head around this too - but hypothetically :
Suppose you had a set of integer numbers. Your resolution was integers.
1,2,3,4,5
Now suppose you multiplied all of those values by 0.5 then by 2 again.
you'd end up with
0, 2, 2, 4, 4
Now suppose you start off with the same set of integers
1,2,3,4,5 Then change your resolution to fractional halves and do the same 0.5 followed by 2 transform
Your result is
1,2,3,4,5
I'm suggesting that switching to 16 bit gives you the same flexibility, that when you start manipulating the data 16 bit you don't magically gain extra precision in the existing data, but you gain space and resolution to manipulate that data more accurately.
Its probably pretty subtle and hard to notice in real world cases, but intuitively it seems correct.
|
|
|
|
11/14/2003 06:30:29 PM · #43 |
I think I see where you're going, but since we're not dealing with doublings...
With 8-bit, we have 2 to the 8th power = 256 values
With 16-bit, we have 2 to the 16th power = 65,536 values
In light of the comparison of those relative values, I think any benefit is going to be VERY subtle. Perhaps subtle to the point of one divided by infinity => 0. OK, that's extreme -- the value is more like 0.0003?
But it's late on a Friday afternoon and I know I'm not thinking straight -- so I'll give it some more thought when I'm better rested. |
|
|
|
11/14/2003 06:32:10 PM · #44 |
Originally posted by Patella:
..8-bit jpegs.. |
Maybe I'm being stupid here. I presume by the above you mean 8 bits per channel? |
|
|
|
11/14/2003 06:36:08 PM · #45 |
Not being stupid, I'm just using short-hand and it may not be clear.
Yes, 8 bits per channel and 16 bits per channel |
|
|
|
11/14/2003 06:44:47 PM · #46 |
Originally posted by Patella:
I think I see where you're going, but since we're not dealing with doublings...
With 8-bit, we have 2 to the 8th power = 256 values
With 16-bit, we have 2 to the 16th power = 65,536 values
In light of the comparison of those relative values, I think any benefit is going to be VERY subtle. Perhaps subtle to the point of one divided by infinity => 0. OK, that's extreme -- the value is more like 0.0003?
But it's late on a Friday afternoon and I know I'm not thinking straight -- so I'll give it some more thought when I'm better rested. |
Actually, no.
Its 8 bit per channel - so a typical '8bit' value is actually 24bit per pixel - or 32bit if you consider Alpha (transparency) channels
RGB 888 is the simplest form of encoding in this scheme.
16 bit images are 16 bit per channel - so you need to cube all of your values - or consider them on a per channel basis.
This is one of the biggest issues with current B&W images - effectively those are 8 bit per pixel (as a grey value - you set R=G=B - so you do only get 256 levels in an 8-bit B&W JPEG) This is also why I prefer working with duotones and printing duotones - you can move back to the entire colour palette and have the 65,000 grey levels or approximations there in.
8-bit images have a potential pallete of 16,777,216 colours.
16-bit images can represent a potential pallete of 281,474,976,710,656
This is why mostly its an argument off in the noise about working with these levels of colour resolution - when you are eye can make out a few thousand on a good day. However, the extra space to manipulate images is were the benefit lies - 16 bit images is equivalent to allowing fractions to be considered when doing the adjustments - not a big deal for the first calcuation, but keeps things accurate over the longer term. 8 bit images accumulate errors very quickly compared to 16-bit images over multiple manipulations (at least I think! )
Now, black is still black, and white is still white - so these additional levels appear as increased precision between the levels.
So previously per channel you went from 1 to 2, now you can go to 256 distinct levels between 1 and 2, giving much, much more resolution - roughly equivalent to 3 decimal places of additional precision (very roughly)
Message edited by author 2003-11-14 18:50:21.
|
|
|
|
11/14/2003 06:52:10 PM · #47 |
| Does all that mean you'd get a more accurate upsampling by converting to 16-bit before using Image Size? |
|
|
|
11/14/2003 07:01:25 PM · #48 |
*throwing up hands in surrender*
I guess that all the PS gurus in NY going on about the difference between 256 and 65,000 are as mixed up as I am then...
Message edited by author 2003-11-14 19:02:22. |
|
|
|
11/14/2003 11:44:40 PM · #49 |
Originally posted by Patella:
*throwing up hands in surrender*
I guess that all the PS gurus in NY going on about the difference between 256 and 65,000 are as mixed up as I am then... |
Its correct_per_ channel - but you have 4 channels per layer 3 RGB and 1 alpha
So 8 bit images are 24bit colour and 8 bit alpha. 16 bit images are 48bit colour and quite often still 8 bit alpha.
This, by the way, is one of the reasons GIF is out of favour for image storage - in that case it really was only 8 bit colours - but you had an indexed 'paintbox' of 256 colours from the 16.8 million colour options availble from 24bit colour (8bit red, 8bit green, 8 bit blue)
In gif each colour had an index into that table - so you could only represent 256 indivdiual colours - and would get quite nasty banding on skin tones - JPEG, with the complete RGB888 bits per pixel allows for much more flexibility and tonal range. (Though quite often pixels aren't actually coded as RGB888 - other compression codings, such as RGB565 are common too - leaving more detail in the green channel but suiting a 16 bit processor or memory system more easily (5+6+5)
RGB888+A, being 32-bit fits quite well within a typical 32 bit processor/ memory hierarchy and the new 16 bit data sizes is going to get a significant step up on 64 bit processors.
This is also fairly clearly explained in the photoshop help files
Photoshop's RGB Color mode uses the RGB model, assigning an intensity value to each pixel ranging from 0 (black) to 255 (white) for each of the RGB (red, green, blue) components in a color image. For example, a bright red color might have an R value of 246, a G value of 20, and a B value of 50. When the values of all three components are equal, the result is a shade of neutral gray. When the value of all components are 255, the result is pure white; when the values are 0, pure black.
RGB images use three colors, or channels, to reproduce colors on-screen. The three channels translate to 24 (8 bits x 3 channels) bits of color information per pixel. With 24-bit images, up to 16.7 million colors can be reproduced. With 48-bit images (16 bits per channel), even more colors can be reproduced. In addition to being the default mode for new Photoshop images, the RGB model is used by computer monitors to display colors. This means that when working in color modes other than RGB, such as CMYK, Photoshop interpolates the CMYK image to RGB for display on-screen.
Although RGB is a standard color model, the exact range of colors represented can vary, depending on the application or display device. Photoshop's RGB Color mode varies according to the working space setting that you have specified in the Color Settings dialog box. (See About working spaces.)
Message edited by author 2003-11-15 00:10:36.
|
|
|
|
11/16/2003 12:12:56 AM · #50 |
I've just had some great success using this feature but for another reason.
Just went and took a heap of motocross photos. Heaps were hopeless due to all the dust in the air. Too many to do complicated adjustments on.
Ran the Shadow/Highlight adjust and it really cleared up the dust, and would presumable work for fog etc as well. Photo is pretty average, but I should stress it is dust in the air that is the effect here that spoils it.
This isn't the best example, but an average one, so it is more realistic.
May not be the very best way to do it, but it was a great fast way for me to process an awful lot of snapshots.
Before:
After:
No adjustment other than this was made (and a resize). Auto levels on top of this then worked as another quick/dirty adjustment, but only for some pics.
|
|
