Ad
Ad
Ad
Pages: « 1 2 [3] 4 5 »   Bottom of Page
Print
Author Topic: C4 vs. LR color rendition  (Read 19122 times)
Peter_DL
Sr. Member
****
Offline Offline

Posts: 421


« Reply #40 on: April 14, 2008, 02:56:13 PM »
ReplyReply

Quote
     
So LR renders wrong color because of my cultural background or that I (suposely) don't like orange or magenta? So why does C4 give the right red and yellow tones? Because the program likes me?

That's a good one.
[a href=\"index.php?act=findpost&pid=189476\"][{POST_SNAPBACK}][/a]
There’s less hope based on your comment, but I will try again with a serious response:

assume there would be an issue at the level of colorimetric interpretation & color reconstruction which would finally render a blue sky more saturated than “accurate“, or fresh green foliage somewhat darker than following the inevitable tone curve, plus a hue shift towards red – surprise: most people would be quite happy, would find the colors to be pleasing. Maybe except those photographers who have to reproduce company logos.

But, with red hues and skin tones, things are differently: the requirement for initial hue accuracy tends to be about the same whether you attempt to capture people or logos.

Good luck.
& Cheers, Peter

--
Logged
samuel_js
Guest
« Reply #41 on: April 14, 2008, 03:03:01 PM »
ReplyReply

Quote
There’s less hope based on your comment, but I will try again with a serious response:

assume there would be an issue at the level of colorimetric interpretation & color reconstruction which would finally render a blue sky more saturated than “accurate“, or fresh green foliage somewhat darker than following the inevitable tone curve, plus a hue shift towards red – surprise: most people would be quite happy, would find the colors to be pleasing. Maybe except those photographers who have to reproduce company logos.

But, with red hues and skin tones, things are differently: the requirement for initial hue accuracy tends to be about the same whether you attempt to capture people or logos.

Good luck.
& Cheers, Peter

--
[a href=\"index.php?act=findpost&pid=189505\"][{POST_SNAPBACK}][/a]

Peter, there's a simple answer. I want the red to be red as I saw it in reality. And as I look at these red jeans today, they are still red, not magenta. As simnple as that.
Logged
Peter_DL
Sr. Member
****
Offline Offline

Posts: 421


« Reply #42 on: April 14, 2008, 03:13:49 PM »
ReplyReply

Quote
Peter, there's a simple answer. I want the red to be red as I saw it in reality. And as I look at these red jeans today, they are still red, not magenta. As simnple as that.
[a href=\"index.php?act=findpost&pid=189509\"][{POST_SNAPBACK}][/a]
Not so. The output dynamic range of monitor or print is typically much smaller than reality.
You would be shocked upon seeing such a dark, dull but strictly colorimetric reproduction (i.e. linearly downscaled).

Peter

--
Logged
samuel_js
Guest
« Reply #43 on: April 14, 2008, 03:38:12 PM »
ReplyReply

Quote
Not so. The output dynamic range of monitor or print is typically much smaller than reality.
You would be shocked upon seeing such a dark, dull but strictly colorimetric reproduction (i.e. linearly downscaled).

Peter

--
[a href=\"index.php?act=findpost&pid=189512\"][{POST_SNAPBACK}][/a]

Not so small to not be accepted  as a "correct rendering".
Anyway, this has nothing to do with the intention of my initial post. It's nothing that color psychology can fix really...
« Last Edit: April 14, 2008, 03:39:06 PM by samuel_js » Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8628



WWW
« Reply #44 on: April 14, 2008, 03:48:02 PM »
ReplyReply

Quote
Peter, there's a simple answer. I want the red to be red as I saw it in reality. And as I look at these red jeans today, they are still red, not magenta. As simnple as that.
[a href=\"index.php?act=findpost&pid=189509\"][{POST_SNAPBACK}][/a]

No, you want red in an output referred image to appear as you recall the red at the scene. Yes, you want red (not magenta or any other color). But when you start talking about what you saw in reality, then the only way we can define this without subjectivity is to measure the red you saw (scene referred) which as to be rendered output referred.
Logged

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

Posts: 2756



« Reply #45 on: April 14, 2008, 04:00:00 PM »
ReplyReply

Quote
OK so you're not going to answer the question, instead go back to color geek speak and talk about matrixes. I think its pretty clear where you're going with all this.
[a href=\"index.php?act=findpost&pid=189469\"][{POST_SNAPBACK}][/a]

I just purchased the LL Camera Raw Tutorial and have just finished watching the interview with Thomas Knoll and the Adobe development team. At the 26 minute mark, Michael discusses the issue of profiles with TK, who stated that support for external profiles will be available at about the time when the tutorial is released. As I suspected, the current ACR profile is limited to matrix operations. TK explained further that ICC profiles are for rendered images and have problems with scene referred data but their multidimensional lookup tables do allow for more flexible color correction. The exact nature of the external profile was not defined.

Despite canine protestations to the contrary, it does seem as if some users have convinced Mr. Knoll that external profiles might have some utility. It will be quite interesting to see what develops.

Bill
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8628



WWW
« Reply #46 on: April 14, 2008, 04:03:02 PM »
ReplyReply

Quote
K explained further that ICC profiles are for rendered images and have problems with scene referred data but their multidimensional lookup tables do allow for more flexible color correction
Despite canine protestations to the contrary, it does seem as if some users have convinced Mr. Knoll that external profiles might have some utility. It will be quite interesting to see what develops.

Bill
[a href=\"index.php?act=findpost&pid=189526\"][{POST_SNAPBACK}][/a]

My NDA doesn't permit me to say much more than you're off base with this assumption!
« Last Edit: April 14, 2008, 04:04:38 PM by digitaldog » Logged

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

Posts: 270


WWW
« Reply #47 on: April 14, 2008, 04:11:20 PM »
ReplyReply

In 1984 I signed a NDA as assistant when we were one of the first to shoot the new MacIntosh.  I can't remember exactly how long that NDA was for.  Wonder if I can talk about it yet?

Quote
My NDA doesn't permit me to say much more than you're off base with this assumption!
[a href=\"index.php?act=findpost&pid=189527\"][{POST_SNAPBACK}][/a]
Logged
Peter_DL
Sr. Member
****
Offline Offline

Posts: 421


« Reply #48 on: April 14, 2008, 04:17:30 PM »
ReplyReply

Quote
Not so small to not be accepted  as a "correct rendering".
[{POST_SNAPBACK}][/a]
Well, there's probably good reason why some guys are spending their lifetime to develop [a href=\"http://www.cis.rit.edu/people/faculty/fairchild/]Color appearance models[/url], to maintain color appearance at different candlepower and illuminants. Color appearance changes significantly at the low luminance level of monitor or print. That is why we have to shift all the sliders, to come to a pleasing rendition (again).

That said, as mentioned earlier, pure reds as well as skin tones are more a subject of accuracy as far as the hue angle is concerned rather than any superimposed tweaks. The magenta skin tones you reported are symptomatic for the issue described in above link.

Perhaps you might wish to start with this article,
even though it includes a bit too much Adobe evangelism for my taste.

Quote
Anyway, this has nothing to do with the intention of my initial post. It's nothing that color psychology can fix really...
Oh, I'm sorry.

Peter

--
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8628



WWW
« Reply #49 on: April 14, 2008, 04:33:10 PM »
ReplyReply

Quote
In 1984 I signed a NDA as assistant when we were one of the first to shoot the new MacIntosh.  I can't remember exactly how long that NDA was for.  Wonder if I can talk about it yet?
[a href=\"index.php?act=findpost&pid=189529\"][{POST_SNAPBACK}][/a]

Well considering what Thomas said on the video, I'll simply say this: Listen to what Hamburg says about presets, what Zalman says about the number of people who should make "profiles" then what Michael says about his experience building ICC camera profiles (something Adobe wouldn't want to subject users having to experience), then draw your own conclusions until we can talk more legally!

The term "profiles" is loosely used in the discussion (for good reason!). So are presets. It will be interesting to discuss the distinctions.
Logged

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

Posts: 3878



WWW
« Reply #50 on: April 14, 2008, 04:39:36 PM »
ReplyReply

Bill -

 I guess we should say "Well Done" to Michael for asking the right questions.

 We have some floating point ICC profiles now, they've been voted through although the published spec doesn't reflect this yet.

 Brian of Raw Developer fame has just informed me that much of what we have been talking about is actually implemented in RD.

I have created a new post in the Processing section of the forum to discuss this as it is technical.

http://luminous-landscape.com/forum/index....showtopic=24706

Edmund

Quote
I just purchased the LL Camera Raw Tutorial and have just finished watching the interview with Thomas Knoll and the Adobe development team. At the 26 minute mark, Michael discusses the issue of profiles with TK, who stated that support for external profiles will be available at about the time when the tutorial is released. As I suspected, the current ACR profile is limited to matrix operations. TK explained further that ICC profiles are for rendered images and have problems with scene referred data but their multidimensional lookup tables do allow for more flexible color correction. The exact nature of the external profile was not defined.

Despite canine protestations to the contrary, it does seem as if some users have convinced Mr. Knoll that external profiles might have some utility. It will be quite interesting to see what develops.

Bill
[a href=\"index.php?act=findpost&pid=189526\"][{POST_SNAPBACK}][/a]
« Last Edit: April 14, 2008, 04:47:56 PM by eronald » Logged

Edmund Ronald, Ph.D. 
digitaldog
Sr. Member
****
Offline Offline

Posts: 8628



WWW
« Reply #51 on: April 14, 2008, 04:45:10 PM »
ReplyReply

Quote
We have some floating point ICC profiles now, they've been voted through although the published spec doesn't reflect this yet.
[a href=\"index.php?act=findpost&pid=189534\"][{POST_SNAPBACK}][/a]

The sooner you forget ICC profiles in respect to this product, the better.

Yes, well done to Michael. He's also a smart enough guy not to try to connect dots where there is no connection or he wasn't explicitly informed about. You should follow suite!
Logged

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

Posts: 3878



WWW
« Reply #52 on: April 14, 2008, 04:51:12 PM »
ReplyReply

Quote
The sooner you forget ICC profiles in respect to this product, the better.

Yes, well done to Michael. He's also a smart enough guy not to try to connect dots where there is no connection or he wasn't explicitly informed about. You should follow suite!
[{POST_SNAPBACK}][/a]

Andrew, we each have our NDA's - and I also have my own self-written code which can do some very cute things

I suggest you read the info by Brian carefully and see what opportunities it offers to consultants.
[a href=\"http://luminous-landscape.com/forum/index.php?showtopic=24706]http://luminous-landscape.com/forum/index....showtopic=24706[/url]


By the way, most of what I'm doing here is targeted at getting things right for the camera industry geeks to supply results to the end-users. It's not targeted directly at having normal end-users without special hardware  build profiles.

Edmund
« Last Edit: April 14, 2008, 04:55:53 PM by eronald » Logged

Edmund Ronald, Ph.D. 
digitaldog
Sr. Member
****
Offline Offline

Posts: 8628



WWW
« Reply #53 on: April 14, 2008, 05:50:36 PM »
ReplyReply

And this has what to do with the ACR pipeline?
Logged

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

Posts: 2756



« Reply #54 on: April 14, 2008, 06:11:37 PM »
ReplyReply

Quote
My NDA doesn't permit me to say much more than you're off base with this assumption!
[a href=\"index.php?act=findpost&pid=189527\"][{POST_SNAPBACK}][/a]

Way off base? Hardly. I merely repeated in my post what Mr. Knoll stated in the interview. He said, "there are some things the ICC profiles can do that we can't do, yet. ... version of Camera Raw ... will have support for external profiles that you can load." He did not say ICC profiles, but the implication is that they are more than presets, which can not do what ICC profiles can.

Bill
Logged
digitaldog
Sr. Member
****
Offline Offline

Posts: 8628



WWW
« Reply #55 on: April 14, 2008, 06:18:06 PM »
ReplyReply

Quote
Way off base? Hardly. I merely repeated in my post what Mr. Knoll stated in the interview. He said, "there are some things the ICC profiles can do that we can't do, yet. ... version of Camera Raw ... will have support for external profiles that you can load." He did not say ICC profiles, but the implication is that they are more than presets, which can not do what ICC profiles can.

You heard what you wanted to hear, some of it using terms that you think are familiar when they may not be. Don't fall in love with the word profile any more than the word preset. But I'm not going to say much more (I've said enough on the subject).
Logged

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

Posts: 3878



WWW
« Reply #56 on: April 15, 2008, 06:10:19 AM »
ReplyReply

Quote
You heard what you wanted to hear, some of it using terms that you think are familiar when they may not be. Don't fall in love with the word profile any more than the word preset. But I'm not going to say much more (I've said enough on the subject).
[{POST_SNAPBACK}][/a]

I just got another email from Brian indicating that Leaf's Raw format can do a lot of interesting things, including embedding "look" profiles. The MF world is clearly using more complex solutions than Lightroom, which explains what the poll is showing regarding color quality

[a href=\"http://luminous-landscape.com/forum/index.php?showtopic=24695]http://luminous-landscape.com/forum/index....showtopic=24695[/url]

Edmund
« Last Edit: April 15, 2008, 06:14:44 AM by eronald » Logged

Edmund Ronald, Ph.D. 
digitaldog
Sr. Member
****
Offline Offline

Posts: 8628



WWW
« Reply #57 on: April 15, 2008, 08:44:59 AM »
ReplyReply

Quote
I just got another email from Brian indicating that Leaf's Raw format can do a lot of interesting things, including embedding "look" profiles.

It would be useful for you define what a profile is, then what a Look Profile is, then what a rendering preset is and how they all differ. Math isn't necessary, end results to the end user are.
Logged

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

Posts: 984



WWW
« Reply #58 on: April 15, 2008, 09:25:41 AM »
ReplyReply

Quote
It would be useful for you define what a profile is, then what a Look Profile is, then what a rendering preset is and how they all differ. Math isn't necessary, end results to the end user are.
[a href=\"index.php?act=findpost&pid=189664\"][{POST_SNAPBACK}][/a]


Regardless of what the look is called, profile, presets, etc. it would be of tremendous value if what you set in C1, Dpp, LC11, etc. would be picked up exactl the same in lightroom and ACR.

This is not to say lightroom is bad, there is no bad in terms of converters and lightroom is the most full featured, but I find with all of my cameras, Nikon, Canon, Leica and Phase it does take a lot of movements to get to the original look in lightxroom I get from the manufacturer's converters, at least as a starting point.

I break digital capture down to three parts.   The on set view which is to assure yourself and the client that your going in the right direction.  The second step is processing for web galleries and/or contacts which is a defined but general view of the look and the selection and of course the third is the  final finish out that should be dead on perfect.

What would be nice actually what would  be great, is if we could preset a series of "film looks" for all the cameras and backs, embed them into the cameras and all of the converters would read this looks as we shot and when we went to process.

Right now it's a mixed bag and most of us shoot, then make the film look up at a later stage and taking a p30 file for instance, going from c1 3.7, c1 v4, Raw Developer, Lightroom etc. in initial view looks like you shot with three or four different cameras.  Sometimes this is ok, but usually just confusing and requires a lot of work to get back to the basics or to the "film look" you envisioned when you were shooting.

JR
Logged

digitaldog
Sr. Member
****
Offline Offline

Posts: 8628



WWW
« Reply #59 on: April 15, 2008, 09:31:16 AM »
ReplyReply

Quote
Regardless of what the look is called, profile, presets, etc. it would be of tremendous value if what you set in C1, Dpp, LC11, etc. would be picked up exactl the same in lightroom and ACR.

Not sure I'm following you. You want a "look" that could be used equally and identically in all the above converters? Or you want to have a "look" you create somehow, applied in LR/ACR?

I seriously doubt its possible to build a cross application "Look" if that's what you're suggesting. And if you did, it probably would have to be applied after the rendering since all the other stuff is totally proprietary from converter to converter.
Logged

Andrew Rodney
Author “Color Management for Photographers”
http://digitaldog.net/
Pages: « 1 2 [3] 4 5 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad