Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: LR histogram changes in PS  (Read 6446 times)
gvaughn
Newbie
*
Offline Offline

Posts: 27


WWW
« on: May 22, 2009, 02:10:37 PM »
ReplyReply

A number of times it seems that when I process a photo in Lightroom (v1.4.1) and make adjustments for both appearance and to keep all tone values within a normal histogram, when I export the image and then open it in Photoshop (CS3), the RGB histogram is significantly different and shows quite a bit of clipping.

Which histogram is right one to go by?  I'm primarily processing for stock photo sales, not inkjet printing, and have always been told to avoid all clipping for that purpose.
Logged

- Greg

www.GregVaughn.com
author, Photographing Oregon
James R
Sr. Member
****
Offline Offline

Posts: 260


« Reply #1 on: May 22, 2009, 04:14:36 PM »
ReplyReply

Quote from: gvaughn
A number of times it seems that when I process a photo in Lightroom (v1.4.1) and make adjustments for both appearance and to keep all tone values within a normal histogram, when I export the image and then open it in Photoshop (CS3), the RGB histogram is significantly different and shows quite a bit of clipping.

Which histogram is right one to go by?  I'm primarily processing for stock photo sales, not inkjet printing, and have always been told to avoid all clipping for that purpose.

I just checked the histogram of a few images in LR2 (ver. 2.3) and PS (CS4).  They were pretty much the same.  There may be some issues with the much older version of LR; but, I don't have that version.  I'm surprised you have not upgraded LR, given the many improvements.  I would think your images are OK, if the blinkies are controlled in LR.
Logged
madmanchan
Sr. Member
****
Offline Offline

Posts: 2110


« Reply #2 on: May 22, 2009, 04:27:05 PM »
ReplyReply

You are likely changing color spaces as you go into CS3. If your CS3 working space is set to sRGB, or you have chosen to export your images from LR in sRGB, then you may (likely will) clip some colors and that's why the histogram looks like some colors are clipped.
Logged

Guillermo Luijk
Sr. Member
****
Offline Offline

Posts: 1336



WWW
« Reply #3 on: May 22, 2009, 05:17:17 PM »
ReplyReply

Quote from: madmanchan
You are likely changing color spaces as you go into CS3. If your CS3 working space is set to sRGB, or you have chosen to export your images from LR in sRGB, then you may (likely will) clip some colors and that's why the histogram looks like some colors are clipped.
This is something that was always strange to me. I am not a LR user (never tried it actually), but if I am not wrong histograms in LR always correspond to the ProPhoto RGB space, but with a sRGB-like gamma curve (i.e. 2.2 gamma linear in the deep shadows). This means that a near to saturate image in LR can easily get clipped once exported to PS in sRGB or even Adobe RGB, and the only way to solve this is to go back to LR and do trial-error adjustments to avoid the clipping.

Eric could you please try to give some information about why Adobe decided to introduce this behaviour in LR? is the reason that LR is intended for photographers that are not so deeply interested in the technical details of the image levels (i.e. the histogram)?

I used to use ACR, and found very useful that ACR's histograms do depend on the output working space selected. I never saw the point of LR working in a different way.

Regards.
« Last Edit: May 22, 2009, 05:23:57 PM by GLuijk » Logged

madmanchan
Sr. Member
****
Offline Offline

Posts: 2110


« Reply #4 on: May 22, 2009, 09:53:54 PM »
ReplyReply

Guillermo, I agree that one of LR's limits is that currently there is not a good way to tell whether certain colors will clip upon conversion to another working space or device space. Output-specific histograms are better, like in ACR as you've observed, but still have limitations in that they don't tell you where in the image the clipping is occurring. I do not know the history behind LR's difference in design in this regard. However, I expect to see improvements in both CR/LR in this area going forward.
Logged

gvaughn
Newbie
*
Offline Offline

Posts: 27


WWW
« Reply #5 on: May 22, 2009, 11:46:21 PM »
ReplyReply

Quote from: madmanchan
You are likely changing color spaces as you go into CS3. If your CS3 working space is set to sRGB, or you have chosen to export your images from LR in sRGB, then you may (likely will) clip some colors and that's why the histogram looks like some colors are clipped.

Ahh! That is indeed the problem. I was telling Lightroom to export files as AdobeRGB, which I use as default in Photoshop for images destined for stock distribution. I didn't realize that LR works in ProPhotoRGB. If I tell LR to export with that profile, the histogram in PS looks the same as in LR. Is there a way to set LR to work in aRGB, since that is what I mostly need to use?

James R: I do intend to upgrade to LR2 and PS CS4 soon, just have recently had to put the available funds into upgrading, repairing and/or replacing laptop, hard drives, cameras, printer and other software first.
Logged

- Greg

www.GregVaughn.com
author, Photographing Oregon
digitaldog
Sr. Member
****
Offline Offline

Posts: 9315



WWW
« Reply #6 on: May 23, 2009, 08:44:29 AM »
ReplyReply

http://www.ppmag.com/reviews/200701_rodneycm.pdf

The Histogram in LR is Melissa RGB (ProPhoto primaries with an sRGB 2.2 TRC "gamma"). That's why they don't match.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
CDL
Newbie
*
Offline Offline

Posts: 15


« Reply #7 on: May 24, 2009, 09:11:27 AM »
ReplyReply

Quote from: madmanchan
Guillermo, I agree that one of LR's limits is that currently there is not a good way to tell whether certain colors will clip upon conversion to another working space or device space. Output-specific histograms are better, like in ACR as you've observed, but still have limitations in that they don't tell you where in the image the clipping is occurring. I do not know the history behind LR's difference in design in this regard. However, I expect to see improvements in both CR/LR in this area going forward.

Madmanchan, what we are talking about here (i.e. also what you call "output-specific histograms"), would't that actually be what is usually called "softproofing"? Or do I understand this wrong? Is there anything more (or less) to the concept of softproofing as to what we discuss here or is it actually exactly this? Thanks very much for your answer as I'm far from being an expert in this, but thought that softproofing could possibly be the solution (if it would exist in LR  ).
Logged
madmanchan
Sr. Member
****
Offline Offline

Posts: 2110


« Reply #8 on: May 24, 2009, 10:02:00 AM »
ReplyReply

Yes, soft proofing is indeed a possible solution ... an imperfect one, but certainly improved relative to the current implementation.
Logged

CDL
Newbie
*
Offline Offline

Posts: 15


« Reply #9 on: May 24, 2009, 12:02:33 PM »
ReplyReply

Just out of curiosity: Why is it imperfect? What would be missing?
Logged
madmanchan
Sr. Member
****
Offline Offline

Posts: 2110


« Reply #10 on: May 24, 2009, 02:20:27 PM »
ReplyReply

It is not perfect for a number of technical reasons, one being that the printer can also print colors that the monitor cannot display. Hence such colors cannot be previewed accurately on the display.
Logged

scott morrish
Full Member
***
Offline Offline

Posts: 112


« Reply #11 on: May 25, 2009, 06:23:45 AM »
ReplyReply

Quote from: madmanchan
It is not perfect for a number of technical reasons, one being that the printer can also print colors that the monitor cannot display. Hence such colors cannot be previewed accurately on the display.


Hi Eric,

In the context of the specific issue you raise above, couldn't we have soft-proofing (preferably with flexibility to use ANY output profile users want to work with) and simply add an out-of gamut indicator (specific to the display rather than the input / output icc profiles) as a variation on the out-of-gamut warning already in Photoshop?  

Admittedly it is not perfect, but it would be workable and much better than the present situation... Especially as most colours can be displayed, and not all images are affected by out-of-gamut issues anyway.

Regards,
Scott
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2882



« Reply #12 on: May 26, 2009, 07:31:55 PM »
ReplyReply

Quote from: madmanchan
Output-specific histograms are better, like in ACR as you've observed, but still have limitations in that they don't tell you where in the image the clipping is occurring.

Eric,

To demonstrate output clipping and where it occurs in the image, why not use the alt key (windows) in conjunction with the exposure control in ACR?

Bill

Logged
gvaughn
Newbie
*
Offline Offline

Posts: 27


WWW
« Reply #13 on: September 18, 2009, 09:52:14 PM »
ReplyReply

A follow-up to this and another thread where I was commenting on clipping the blue channel when exporting an image from Lightroom to Photoshop.

I now am using Lightroom 2.5 and PS CS4 and still find that an image that gives me a good histogram in LR shows significant clipping on the shadow end of the blue channel. And this is with LR set to Export to the Adobe RGB profile.

I discovered today that by simply applying a Vibrance adjustment of even +1 or -1 in PS, the blue channel clipping suddenly disappears. Not sure what's happening here, and there's no apparent visible change in the image on my monitor.

 - Greg

   www.GregVaughn.com
« Last Edit: September 18, 2009, 09:52:27 PM by gvaughn » Logged

- Greg

www.GregVaughn.com
author, Photographing Oregon
Guillermo Luijk
Sr. Member
****
Offline Offline

Posts: 1336



WWW
« Reply #14 on: September 19, 2009, 08:20:58 AM »
ReplyReply

Quote from: gvaughn
A follow-up to this and another thread where I was commenting on clipping the blue channel when exporting an image from Lightroom to Photoshop.

I now am using Lightroom 2.5 and PS CS4 and still find that an image that gives me a good histogram in LR shows significant clipping on the shadow end of the blue channel. And this is with LR set to Export to the Adobe RGB profile.

I discovered today that by simply applying a Vibrance adjustment of even +1 or -1 in PS, the blue channel clipping suddenly disappears. Not sure what's happening here, and there's no apparent visible change in the image on my monitor.

The B channel clipped in the exported image means Adobe RGB couldn't keep all the colour gamut when converting from LR's colour space. That means you already lost that information in the B channel.

If now in PS, by desaturating the B clipping disappears, is just an illusion in terms of quality. The information was already lost. The variations in that channel that were present in LR, are not present in PS anymore. Desaturating increases the B channel in those pixels to some positive value, that's why it looks non-clipped. But the hurt was already done so you are not fixing the problem here.

If you can't see differences, is because they are actually very slight and impossible to notice in a standard monitor. Maybe this last sentence is the one to take into account from a practical point of view.

Regards
« Last Edit: September 19, 2009, 08:21:10 AM by GLuijk » Logged

gvaughn
Newbie
*
Offline Offline

Posts: 27


WWW
« Reply #15 on: September 21, 2009, 10:24:18 AM »
ReplyReply

Quote from: GLuijk
If you can't see differences, is because they are actually very slight and impossible to notice in a standard monitor. Maybe this last sentence is the one to take into account from a practical point of view.

Yup. It really doesn't seem to make any significant difference on screen or in print, so I've been ignoring it even though I know that I'm losing some data. But from the early days of digital, I've had it drummed into me by stock agencies, magazines and books that a histogram should show no clipping, and I found it interesting that just by adding the slightest Vibrance adjustment the histogram suddenly looked pretty. : )
Logged

- Greg

www.GregVaughn.com
author, Photographing Oregon
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad