Ad
Ad
Ad
Pages: [1] 2 »   Bottom of Page
Print
Author Topic: Looking for a good color space converter - found, thank you.  (Read 8450 times)
xpatUSA
Sr. Member
****
Offline Offline

Posts: 305



WWW
« on: April 05, 2013, 02:51:22 PM »
ReplyReply

Sometime back, I shot a sunflower RAW image which was not well converted by ACR to RGB color. I tracked that problem down to yellows that were out-of-gamut in sRGB but not in ProPhoto. As I understand it, there are various ways (familiar to people who print stuff) of bringing wide gamuts into smaller ones - absolute colorimetric, perceptual, shift the white point, don't, etc, etc.

What would this forum recommend that would give such choices and more when converting image files from one space to another, as opposed to what Adobe thinks is good for you? I have ACR 5.4 and PSE 6 and am not looking to upgradeeither  anytime soon. I don't intend to get LR, either. Nor GIMP, nor rawTherapee (not easy to open Sigma files in those).

Mainly I shoot Sigma X3F images and play around with some of them quite a lot for research purposes. I often use Dave Coffins' dcraw and lately have been looking at it's XYZ output in ImageJ, the background for that being lamp spectral emission shots using a spectroscope. Just downloaded a ImageJ plug-in, not tried it yet though.
« Last Edit: April 28, 2013, 12:11:28 PM by xpatUSA » Logged

best regards,

Ted
digitaldog
Sr. Member
****
Offline Offline

Posts: 9225



WWW
« Reply #1 on: April 05, 2013, 02:54:05 PM »
ReplyReply

Sometime back, I shot a sunflower RAW image which was not well converted by ACR to RGB color. I tracked that problem down to yellows that were out-of-gamut in sRGB but not in ProPhoto. As I understand it, there are various ways (familiar to people who print stuff) of bringing wide gamuts into smaller ones - absolute colorimetric, perceptual, shift the white point, don't, etc, etc.

That's what the profile will do for you, so just let it after soft proofing. For conversion using simple matrix profiles (sRGB, Adobe RGB  etc), there is only the Colorimetric table in such profiles. You'll have to use Relative Colorimetric despite what you select (but don't pick Absolute, that's the same table without a white shift).

http://digitaldog.net/files/LR4_softproof2.mov
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
xpatUSA
Sr. Member
****
Offline Offline

Posts: 305



WWW
« Reply #2 on: April 05, 2013, 03:31:55 PM »
ReplyReply

Thanks Andrew,

Sorry, I should have made it more clear that I don't have LR4 and that I don't print anything, hence I never soft-proof.

A specific case for me would be:

An image looks good (i.e. no clipping, no horrible discontinuities, etc.) on-screen (but of course, not accurate) in color space A, say ProPhoto. But, when it is converted by something into the much maligned sRGB space clipping occurs - whereas one might prefer a different conversion, but still for display on-screen. It would be expected that the output image would have the correct ICC for screen applied as part of the process.
Logged

best regards,

Ted
Schewe
Sr. Member
****
Offline Offline

Posts: 5532


WWW
« Reply #3 on: April 05, 2013, 04:37:23 PM »
ReplyReply

What would this forum recommend that would give such choices and more when converting image files from one space to another, as opposed to what Adobe thinks is good for you?

It's not a question of what Adobe thinks, but the ICC. ICC v2 color space profiles can only use Relative Colorimetric rendering intent...so, out of gamut clips. And this isn't a function of Adobe's software, simply following the ICC spec.

The ICC has release their v4 spec and offers an sRGB v4 profile that does have Perceptual rendering. You can read more about it here: sRGB v4 Preference. With this profile you can actually use a perceptual rendering intent going from PP RGB > sRGB v4.

As far as optimizing color space conversions, you can preview what a conversion will look like and use soft proofing to adjust the conversion before converting...but that takes an app that can soft proof like Photoshop or Lightroom. I don't think there is any way to use soft proofing in Elements.
Logged
Simon Garrett
Sr. Member
****
Offline Offline

Posts: 395


« Reply #4 on: April 05, 2013, 05:14:52 PM »
ReplyReply

An image looks good (i.e. no clipping, no horrible discontinuities, etc.) on-screen (but of course, not accurate) in color space A, say ProPhoto. But, when it is converted by something into the much maligned sRGB space clipping occurs - whereas one might prefer a different conversion, but still for display on-screen. It would be expected that the output image would have the correct ICC for screen applied as part of the process.
Just to eliminate some of the variables:
  • Is your monitor calibrated and profiled with a hardware device (ColorMunki, Spyder, Huey etc)?
  • Are you viewing the images with colour-manged programs or viewers?  Photoshop (and I assume Elements), Lightroom and the W7 and W8 photo viewers are colour managed.  Firefox and Safari are colour manged (the latter only if there's an embedded profile).  IE (even IE10), Chrome, and most other programs are not colour managed.
Unless both these are true, then images are likely to look different in different colour spaces, even if the conversion from one colour space to another is perfect. 
Logged
xpatUSA
Sr. Member
****
Offline Offline

Posts: 305



WWW
« Reply #5 on: April 05, 2013, 06:12:49 PM »
ReplyReply

It's not a question of what Adobe thinks, but the ICC. ICC v2 color space profiles can only use Relative Colorimetric rendering intent...so, out of gamut clips. And this isn't a function of Adobe's software, simply following the ICC spec.

The ICC has release their v4 spec and offers an sRGB v4 profile that does have Perceptual rendering. You can read more about it here: sRGB v4 Preference. With this profile you can actually use a perceptual rendering intent going from PP RGB > sRGB v4.

Thanks, I may have there before re: rendering intent and did actually download V4 but it got lost when my computer died last year, RIP. Even worse, I can barely remember which intent does what (getting old 70+) but I was interested mainly in the one that squeezes the larger gamut down, losing saturation in so doing.

Quote
As far as optimizing color space conversions, you can preview what a conversion will look like and use soft proofing to adjust the conversion before converting...but that takes an app that can soft proof like Photoshop or Lightroom. I don't think there is any way to use soft proofing in Elements.

Well, I don't print, so would be happy enough with an app that takes e.g. a 16-bit ProPhoto TIFF output from dcraw and handily turns it into a 8-bit sRGB JPEG. And it would mostly be research-based, not a regular workflow thing.

Simon,

I may have wasted your time by not making it clear that I am not trying solve a problem per se. Your reply seems to be in a trouble-shooting style, which is why I said that.

Just to answer your questions though, my monitor is totally uncalibrated apart from me turning down the color temp 'cause it's to bright for my old eyes. I am no purist, and the NEC Multi-sync shows it! My and other's pics look good enough on it for my purposes - which may seem at odds with the OP, I know.

I use many different viewers for different reasons: FastStone Viewer, FireFox, Picasa, ACR, PSE, ImageJ, IrfanView, Sigma Photo Pro, and (blush) the Windows XP thing. I'm used to all their differences which don't bother me that much, to be honest.
« Last Edit: April 05, 2013, 06:18:58 PM by xpatUSA » Logged

best regards,

Ted
digitaldog
Sr. Member
****
Offline Offline

Posts: 9225



WWW
« Reply #6 on: April 05, 2013, 06:16:47 PM »
ReplyReply

Sorry, I should have made it more clear that I don't have LR4 and that I don't print anything, hence I never soft-proof.

Soft proofing is for more than printing. IF you want to see the effects of moving into different color spaces AND you have the ability to control the rendering intent (and thus the gamut clipping/compression that MUST result), it's useful to soft proof. But until you calibrate and profile that display, none of this is even worth doing. FWIW, when moving from larger to smaller color space, gamut clipping is often going to happen and you have to live with it.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
xpatUSA
Sr. Member
****
Offline Offline

Posts: 305



WWW
« Reply #7 on: April 05, 2013, 07:43:01 PM »
ReplyReply

Soft proofing is for more than printing. IF you want to see the effects of moving into different color spaces AND you have the ability to control the rendering intent (and thus the gamut clipping/compression that MUST result), it's useful to soft proof. But until you calibrate and profile that display, none of this is even worth doing. FWIW, when moving from larger to smaller color space, gamut clipping is often going to happen and you have to live with it.

Hi Andrew, perhaps there is a need for more understanding - since "caps" are beginning to appear Wink

I'm talking about mostly research for my own amusement into things like lamp emission spectra, shooting van der Wolf targets, and stuff that does not require perfect color rendition on my monitor before my very eyes. Another case in point was ACR putting blotches on a flower whereas SPP did not . . . research was the answer - and might have been a lot easier if I had the OP subject to hand. In these cases, "proofing" for me means digging into an image with, e.g., ImageJ and examining file data values rather than what's on the screen.

I hasten to add the obvious: I am not a Pro, neither do I have a gallery of pics needing the ultimate in presentation.

Yet, photography can be a fascinating subject . . even with an un-calibrated monitor . . .

« Last Edit: April 05, 2013, 09:28:23 PM by xpatUSA » Logged

best regards,

Ted
xpatUSA
Sr. Member
****
Offline Offline

Posts: 305



WWW
« Reply #8 on: April 06, 2013, 08:45:48 AM »
ReplyReply

Perhaps the quest for "a good color space converter" is inappropriate?

Woke up this morning with the thought that a converter is a converter and no more than that. In other words, going from (easy example) XYZ to sRGB is a fixed conversion and, as Andrew said, if it's gonna clip it will. In which case, the question in the OP is not possible to answer in the way that the OP intended. Which is what I believe y'all were trying to tell this obstinate old man Wink

So, if I understand correctly, this raw file will probably not clip (triangle is sRGB):




But the following one could be quite a problem:



The second gamut being that of a sunflower shot in sunlight.

And, if understand further, correct gamut conversion per se does not include rendering intent which, duh, is what it says it is!

I said earlier that flowers can be troublesome. Here is a typical yellow flower spectral reflectivity:



Looking at the purple line on the graph, that flower is yellow with next to no blue in it.

Here's how ACR (left) rendered my sunflower versus Sigma PhotoPro (right) last year:



Nothing in ACR nor in PSE post could fix those blotches without messing up the image saturation or color balance, whereas SPP and dcraw produced a blotch-free image by magic. I'll stress that it was the blotches that irritated, not the rendered color of the petals (too golden).

Here's the final sRGB image of flower in question and I can't remember how I did it Sad



Thanks to you all for your time and advice,
« Last Edit: April 06, 2013, 09:00:35 AM by xpatUSA » Logged

best regards,

Ted
Schewe
Sr. Member
****
Offline Offline

Posts: 5532


WWW
« Reply #9 on: April 06, 2013, 11:30:46 AM »
ReplyReply

What you are seeing there is more a gamut clipping issue where the out of gamut color cause uneven gradations. Again, download and install the sRGB V4 profile and try that.

BTW, you should really read up a bit on the whole ICC workflow...the actual color transform algorithms are contained in the color matching method (CMM) but the choice of rendering intents is contained in the receiving profile. So if you are going from PP RGB > sRGB, it's the sRGB that must have the various rendering intents...and the sRGB v4 Preference profile I mentioned in the beginning has that and would allow a perceptual rendering from PP RGB > sRBG and get the advantage of gamut compression instead of simple clipping.
Logged
xpatUSA
Sr. Member
****
Offline Offline

Posts: 305



WWW
« Reply #10 on: April 06, 2013, 12:26:31 PM »
ReplyReply

What you are seeing there is more a gamut clipping issue where the out of gamut color cause uneven gradations. Again, download and install the sRGB V4 profile and try that.

Thanks again. Er - install it where? I forget where it went last time. (Windows XP Pro, SP3).

Quote
BTW, you should really read up a bit on the whole ICC workflow...the actual color transform algorithms are contained in the color matching method (CMM) but the choice of rendering intents is contained in the receiving profile. So if you are going from PP RGB > sRGB, it's the sRGB that must have the various rendering intents...and the sRGB v4 Preference profile I mentioned in the beginning has that and would allow a perceptual rendering from PP RGB > sRBG and get the advantage of gamut compression instead of simple clipping.

Excellent advice . . I've read it and read it 'til I'm blue in the face - then I get interested in other stuff and it just fades away. But you and the Dog have reminded me about rendering intent being being contained in the receiving file, e.g. embedded IC profile, if I understand it this time around.
Logged

best regards,

Ted
Simon Garrett
Sr. Member
****
Offline Offline

Posts: 395


« Reply #11 on: April 06, 2013, 12:45:05 PM »
ReplyReply

On Windows, colour profiles go in C:\Windows\System32\spool\drivers\color.

Apart from reference white papers on the ICC site (http://www.color.org/index.xalter) the best reference is "Color Management for Photographers" by Andrew Rodney (aka digitaldog, who posted above).  It's from 2005 and some of the examples relate to 2005 software, but virtually all of it is still relevant.
Logged
Schewe
Sr. Member
****
Offline Offline

Posts: 5532


WWW
« Reply #12 on: April 06, 2013, 02:09:26 PM »
ReplyReply

But you and the Dog have reminded me about rendering intent being being contained in the receiving file, e.g. embedded IC profile, if I understand it this time around.

Close...the way it goes is this: Source profile dictates the profile connection space (PCS) which could be L*a*b* or XYZ (usually XYZ in matrix profiles), which is given to the color matching module (CMM) which uses the rendering intents in the destination profile (and it's PCS) to do the color transform...so it's the destination profile whose rendering intents dictate how the color transform (the CMM) does it's conversion. That's why the sRGB V4 Preference profile is useful...it actually has other rendering intents than the normal ICC v2 Relative Colorimetric rendering.

The rendering intent controls how the source gamut is converted to the destination gamut. With V2 matrix profiles (display class working space profiles), there's no real choice (even if software allows you to change it), it's always, only Relative Colorimetric...the sRGB V4 Preference profile DOES allow the use of perceptual rendering, which is what you want when compressing the source gamut into the destination gamut.

Ultimately, if you want control over the PP RGB to sRGB conversion, you'll want something that can soft proof the sRGB V4 Preference profile destination. You might want to take a good look at Lightroom if you don't want to move up to a full Photoshop license. Elements is rather limited...
Logged
xpatUSA
Sr. Member
****
Offline Offline

Posts: 305



WWW
« Reply #13 on: April 06, 2013, 08:44:11 PM »
ReplyReply

Close...the way it goes is this: Source profile dictates the profile connection space (PCS) which could be L*a*b* or XYZ (usually XYZ in matrix profiles), which is given to the color matching module (CMM) which uses the rendering intents in the destination profile (and it's PCS) to do the color transform...so it's the destination profile whose rendering intents dictate how the color transform (the CMM) does it's conversion. That's why the sRGB V4 Preference profile is useful...it actually has other rendering intents than the normal ICC v2 Relative Colorimetric rendering.
Thanks again, Jeff,

I did a *.icc search on my machine and it found 155 (now 156) 15 (now) 16 including today's sRGB V4 Preference profile. Where does it go? Do I re-name it? Do I move "old" sRGB profiles? It seems that several applications have a place for profiles . . often the "same" ones, by filename anyway Sad

Quote
The rendering intent controls how the source gamut is converted to the destination gamut. With V2 matrix profiles (display class working space profiles), there's no real choice (even if software allows you to change it), it's always, only Relative Colorimetric...the sRGB V4 Preference profile DOES allow the use of perceptual rendering, which is what you want when compressing the source gamut into the destination gamut.

By "allow the use of", does that occur invisibly or do all my applications somehow now come up with choices. So like, in Elements 6, where it said "Convert to sRGB" under the image menu would it now say something different? (not talking about the choices offered for printing).

Quote
Ultimately, if you want control over the PP RGB to sRGB conversion, you'll want something that can soft proof the sRGB V4 Preference profile destination. You might want to take a good look at Lightroom if you don't want to move up to a full Photoshop license. Elements is rather limited...

Recalling that I don't print and that the destination is my monitor (or a web sRGB monitor), what would "soft proofing" be in that context? Sadly, PSE6 is still serving my main purposes very well . . so I'll still with it until it irritates me enough to move elsewhere which might be soon if I don't get a grip of this conversion business  Undecided


Simon,

Thanks for the link, will be heading there directly . .
Logged

best regards,

Ted
bjanes
Sr. Member
****
Offline Offline

Posts: 2866



« Reply #14 on: April 06, 2013, 10:02:43 PM »
ReplyReply

Ultimately, if you want control over the PP RGB to sRGB conversion, you'll want something that can soft proof the sRGB V4 Preference profile destination. You might want to take a good look at Lightroom if you don't want to move up to a full Photoshop license. Elements is rather limited...

Jeff,

I have my own shot of a sunflower with highly saturated yellow. Soft-proofing with the sRGB V4 Preference profile using the perceptual intent in Photoshop does show some reduction in clipping when converting to sRGB. However, in LR 4.4 on my 64 bit windows machine, the sRGB V4 Preference profile does not seem to be available--it does not show up in the list of installed profiles even though the profile is properly installed and does work with Photoshop.

Do you know what is going on and have you seen the same with LR on your Mac? Also, ColorThinkPro ver 3.0.3 does not work properly with the Ver 4 profile-the color plot and image view is way off.

Bill
Logged
Schewe
Sr. Member
****
Offline Offline

Posts: 5532


WWW
« Reply #15 on: April 06, 2013, 10:31:15 PM »
ReplyReply

However, in LR 4.4 on my 64 bit windows machine, the sRGB V4 Preference profile does not seem to be available--it does not show up in the list of installed profiles even though the profile is properly installed and does work with Photoshop.

Yeah, I just tested this and that profile does not show up in LR 4.4. Not sure why...I'll ask.
Logged
Schewe
Sr. Member
****
Offline Offline

Posts: 5532


WWW
« Reply #16 on: April 06, 2013, 10:34:13 PM »
ReplyReply

By "allow the use of", does that occur invisibly or do all my applications somehow now come up with choices. So like, in Elements 6, where it said "Convert to sRGB" under the image menu would it now say something different? (not talking about the choices offered for printing).
Simon,

No, it means that when selecting Perceptual in a profile conversion, it can actually use the perceptual rendering. In most color space profiles like PP RGB, Adobe RGB and the v2 sRGB, even though software allows you to select perceptual, there is no perceptual rendering intent in the profile, so regardless of what you select in rendering intent, it's always only doing a relative colorimetric rendering.
Logged
Simon Garrett
Sr. Member
****
Offline Offline

Posts: 395


« Reply #17 on: April 07, 2013, 02:55:55 AM »
ReplyReply

Yeah, I just tested this and that profile does not show up in LR 4.4. Not sure why...I'll ask.
I'm sure I read that LR doesn't support v4 profiles, or is that out of date or just a myth?
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 9225



WWW
« Reply #18 on: April 07, 2013, 12:08:12 PM »
ReplyReply

Yeah, I just tested this and that profile does not show up in LR 4.4. Not sure why...I'll ask.

It's because it's missing a tag that LR uses to filter profiles.  The "class" of the ICC profile needs to be Display or Output for it to show up for soft-proofing in Lr. I have a copy where this was updated if anyone's interested.
Logged

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

Posts: 9225



WWW
« Reply #19 on: April 07, 2013, 12:12:32 PM »
ReplyReply

I'm sure I read that LR doesn't support v4 profiles, or is that out of date or just a myth?

Depends on the profile, how it was built and where you can use it. Generally speaking, V4 profiles bring nothing useful to the party, they are half baked "V2 profiles in V4 sheep's clothing". I don't know what profile package today supports what is called the PRMG which is a key to V4 profiles being (more) useful.

Until PRMG is fully supported, we're back to the part of the conversion whereby the CMM gets Lab data and has no idea about the source (was it big like ProPhoto or small like sRGB?). Each package makes it's own assumptions which is, well an assumption. It is useful with the V4 sRGB profile to have a perceptual table, but by the time that profile get's it's data, the Perceptual Rendering Gamut is no more defined than if you were using a V2 profile.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
Pages: [1] 2 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad