Ad
Ad
Ad
Pages: « 1 [2]   Bottom of Page
Print
Author Topic: ETTR and LR 4 (and DxO)  (Read 6513 times)
elied
Sr. Member
****
Offline Offline

Posts: 269


« Reply #20 on: April 23, 2012, 06:09:51 PM »
ReplyReply

Quote
So a reliable (RAW) clipping indicator would be really good to have in LR
..
In the meantime .... I have found that putting the Temperature and Tint sliders all the way over to the left (2000, -150) will produce a histogram that is very much like a Uni-WB histogram. It may not be a true null WB, but it is close. (I often shoot with a Uni-WB and at the "As Shot" setting LR's sliders go to 2050 and -150). It also is quite close to histograms in Raw Digger and Rawnalyze and I believe it gives a fairly good representation of the nature of the Raw capture, especially with regard to clipping.
Logged

Roll over Ed Weston,
Tell Ansel Adams the news
Hans Kruse
Sr. Member
****
Offline Offline

Posts: 760



WWW
« Reply #21 on: April 24, 2012, 06:04:08 AM »
ReplyReply

I'm somewhat embarrased not to have seen Uni-WB mentioned before and not knowing about it....does anybody have a file for the 1Ds mkIII that could be used for setting the Uni-WB in camera? I'm not sure I would like to shoot with that color tempeture (I shoot RAW of course), but it would be interesting to check it. Also if this helps checking in LR this would be good to have. I do think LR PV2012 does a beautiful job in handling the situations where just channel is clipped and I have checked some pictures where the sky (clouds) contained clippings and I checked one version not clipped versus one clipped and I could not see the difference in the areas clipped. But the fact that LR PV2012 gives me no warnings about clipping is a concern as mentioned a couple of times. Now I know that just changing to PV2010 will give a hint in some cases and a fairly quick turn around for critical pictures via Rawdigger is fairly quick by right clock an image and choose show in Finder and the open the RAW file with Rawdigger. But if LR could do it directly that would be nice Smiley
Logged

elied
Sr. Member
****
Offline Offline

Posts: 269


« Reply #22 on: April 24, 2012, 06:36:36 AM »
ReplyReply

Back several years ago when I first started messing with Uni-WB, people were creating magenta jpgs that could be loaded into the camera and used to create a Uni-WB custom WB setting. Then somebody (E-K at the POTN forum) discovered a much easier way. He reasoned that if the (Canon) camera were given a target image that was already perfectly balanced (R=G=B), the calculated custom WB multipliers would be 1/1/1. There are two ways to easily create such a target; either shoot a black frame (0/0/0) by leaving the lens cap on or completely overexpose a frame (255/255/255). I prefer the black frame. When you use it to make the CWB, the camera will warn you that the result will not be pretty. Ignore that.
Logged

Roll over Ed Weston,
Tell Ansel Adams the news
Hans Kruse
Sr. Member
****
Offline Offline

Posts: 760



WWW
« Reply #23 on: April 24, 2012, 07:18:04 AM »
ReplyReply

Back several years ago when I first started messing with Uni-WB, people were creating magenta jpgs that could be loaded into the camera and used to create a Uni-WB custom WB setting. Then somebody (E-K at the POTN forum) discovered a much easier way. He reasoned that if the (Canon) camera were given a target image that was already perfectly balanced (R=G=B), the calculated custom WB multipliers would be 1/1/1. There are two ways to easily create such a target; either shoot a black frame (0/0/0) by leaving the lens cap on or completely overexpose a frame (255/255/255). I prefer the black frame. When you use it to make the CWB, the camera will warn you that the result will not be pretty. Ignore that.

Thanks, just did this and the WB setting in LR (as shot) becomes 2700, -150. For some testshots bracketed by 1/3 there was a perfect relation between blinkies on the camera LCD and the clipping in Rawdigger. LR did with PV2012 show clipping at the as shot WB although small and PV2010 larger. Of course that WB could not be used, so going to e.g. daylight there were no clipping warnings in LR. Will do some more testing.
Logged

elied
Sr. Member
****
Offline Offline

Posts: 269


« Reply #24 on: April 24, 2012, 08:20:56 AM »
ReplyReply

Thanks, just did this and the WB setting in LR (as shot) becomes 2700, -150. For some testshots bracketed by 1/3 there was a perfect relation between blinkies on the camera LCD and the clipping in Rawdigger. LR did with PV2012 show clipping at the as shot WB although small and PV2010 larger. Of course that WB could not be used, so going to e.g. daylight there were no clipping warnings in LR. Will do some more testing.
When I use the eye-dropper WB tool in LR on a black frame it goes to 2700/-150. I think the important factor is the Tint -150 because in normal broad spectrum light the green channel will always be most exposed. If the green is not clipped the chances are very good that the other two are also not clipped.
Logged

Roll over Ed Weston,
Tell Ansel Adams the news
Hans Kruse
Sr. Member
****
Offline Offline

Posts: 760



WWW
« Reply #25 on: April 25, 2012, 02:51:01 PM »
ReplyReply

I have checked further and found the clipping warning on the LCD pretty accurate compared to Rawdigger when setting the camera WB to Uni-WB. And even better this means that the live histogram in live view when the Uni-WB is set is a real RAW histogram (or at least very close to). The only bad thing is that libe view on the 1Ds mkIII do not have electronic first curtain Wink So ETTR is easy in live view in this way.
« Last Edit: April 25, 2012, 03:23:57 PM by Hans Kruse » Logged

Hans Kruse
Sr. Member
****
Offline Offline

Posts: 760



WWW
« Reply #26 on: May 12, 2012, 06:35:38 AM »
ReplyReply

I guess it all depends on what you consider to be "highlight recovery"...the new image adaptive PV 2012 has an auto-ranging capability which can appear as "recovery" but I don't think it really is in a classic sense. What I think is going on is that the way the highlights are rendered in PV 2012 has a different roll off which extracts more highlight detail. Note that in a linear capture, the brightest area of the capture has a lot of data even if there is no clipping. Recurving the tone mapping to extract more usable image data is certainly what PV 2012 is doing. That's what I think people are seeing...and I don't call that "highlight recovery" which to me means extracting image detail when one or more channels are clipped. I'm pretty sure there is a small degree of classic highlight recovery going on even when all Basic panel controls are at zero which I think is what Eric is talking about and I think it's the new tone mapping that is making the biggest differences. But a minus Exposure or Highlights is clearly when the classic "highlight recovery" really kicks in and that has been substantially improved in PV 2012.

I'm still coming to grips with PV 2012 and I've been working with it in various forms since late last summer...what I think is going on is that this new tone mapping logic is really, really leading edge and traditional descriptions and terms fail to really describe what going on under the hood. I've read the post on Lightroom Journal titled Magic or Local Laplacian Filters? and that stuff is too deep for me to understand...but what I will say it that PV 2012 has made a major improvement in the ability to tone map raw captures. I can't say it's "perfect" (there are a couple of types of images where tone bleeding can occur) but compared to PV 2010 it's clear that one can get a lot more image detail out of highlights and shadows. I hesitate to call this a change in the dynamic range of the image but is sure does make it easier to deal with higher contrast scenes...I'm going to be in the process of writing about this stuff in the very near future...the way I usually do that is ask some questions of Eric and Thomas, swirl it around and spit it out and see if they agree with me. I'll let ya all know how that goes :~)

Not to push you Wink and I don't I could even if I tried, but it would be really great with a more comprehensive writeup on what the new PV2012 algorithms do in the highlight are and in the shadows as well. Since I came home from my Tuscany trip, I have looked at many of the older captures where I have kept overexposed ones for exposure blending and found that in many cases the overexposed image (clipped by typically one channel in the RAW histogram) which was not usable in PV2010 now looks indistinguishable in 1:1 view from the exposure one stop lower (and under the clipping limit according to the RAW histogram) in the highlights. This means, of course, that in high DR scenes I can in many cases do tone mapping in LR4 on a single exposure rather than use Photomatix Pro, Photoshop HDR Pro (or manual blending in Photoshop) or other tools. This is wonderful news compared to LR3 and in my opinion alone a strong reason to upgrade to LR4 (if anyone still doubts that LR3 to LR4 is a worth while upgrade!!).

So do I need a writeup to continue using LR4 for what I have now found to be an amazing new capability? No, certainly not, but I'm curious to see an explanation of what it really does and where the limits are so to speak! For example, can it recover more than one channel clipped? Does it work from interpolation across an area where clipping occurs and estimate color and luminosity from neighboring areas to fill in the gaps in the clipped areas? Etc. etc. and it is probably much more advanced that what I hint at here. It would be good to know how this works rather than reverse engineer the understanding from working with the pictures alone.

Using a Canon 1Ds mkIII this is very good news and does away with many of the situations where I previously would do a form of exposure blending, since I can now edit the picture in LR4 alone. So for any Adobe engineers reading this, a huge thanks for this new capability in LR4.
Logged

jrsforums
Sr. Member
****
Offline Offline

Posts: 746


« Reply #27 on: May 12, 2012, 09:04:21 AM »
ReplyReply

If I may, let me share some of the things, I think, I have learned and some of the exchanges I have had with Eric...

First, I do not believe there is any easy or cookbook way to rationalize what your camera histogram shows and how far you can push ETTR to not blow highlights in LR.  Each camera system and usually each camera is going to be a little different. 

You need to learn how you camera is going to work....i.e. a calibrate it.  Lee Varis, in his book 'Mastering Exposure..." (or a free PDF extract from it) shows how one can calibrate it...i.e. how much room do you have after you hit the blinkies and right side of the chart. This, obviously, needs to be done in oncert with the raw converter you are using.

One thing I noticed, and is cussed with Eric, is that 2010 and 2012 are quite different in what hey tell you about over exposed highlights.  In 2010, if you have red indicators of over exposure in your image, that does not tell you if they are truly blown or just that the other settings have pushed them there.  Only later, if you try to recover hem will you find that some are really blown and have lost all, or most, detail.

I have found 2012 to be quite wonderful in this area.  At the initial default settings, the only areas which show red are are those that are totally or partially "blown", i.e. misinterpret detail.  This makes it much easier to determine if these areas are significant to the image...and if so...to not even bother to try to work on the image as these areas will never look good and will usually at best provide a silvery or grayish area with no texture.

Anyway, this was first on my list of "likes" about LR 4.  The more I use it, the longer the list becomes.

John
Logged

John
Hans Kruse
Sr. Member
****
Offline Offline

Posts: 760



WWW
« Reply #28 on: May 12, 2012, 09:32:00 AM »
ReplyReply

In the past I found a pretty ok correlation between the blinkies on the camera and in LR. That could be improved a bit using the Uni-WB setting which I tried.

Using LR4 PV2012 there is no longer such a relationship as mentioned. If you take a picture that has a blown channel seen in the RAW histogram using e.g. Rawdigger then often LR4 PV2012 will not show any clipping in the default input and when adjusted for the baseline offset. So what is interesting is how I can rely on the clipping indicators in LR4 PV2012 or not. In essence what are the algorithms in lay mans terms that make it possible to not show clipping indicators in areas where there is a blown channel (and yes, I know that the white balance will change this hugely). What I do in practical terms is to compare pictures from an exposure that is not clipped with one that is clipped but LR4 PV2012 says is ok and see if there is any visible difference in the areas of clipping (which I can check in addition in PV2010 using the Uni-WB (2700, -150) which corresponds well with the Raw histogram).
Logged

jrsforums
Sr. Member
****
Offline Offline

Posts: 746


« Reply #29 on: May 12, 2012, 10:14:25 AM »
ReplyReply

In the past I found a pretty ok correlation between the blinkies on the camera and in LR. That could be improved a bit using the Uni-WB setting which I tried.

Using LR4 PV2012 there is no longer such a relationship as mentioned. If you take a picture that has a blown channel seen in the RAW histogram using e.g. Rawdigger then often LR4 PV2012 will not show any clipping in the default input and when adjusted for the baseline offset. So what is interesting is how I can rely on the clipping indicators in LR4 PV2012 or not. In essence what are the algorithms in lay mans terms that make it possible to not show clipping indicators in areas where there is a blown channel (and yes, I know that the white balance will change this hugely). What I do in practical terms is to compare pictures from an exposure that is not clipped with one that is clipped but LR4 PV2012 says is ok and see if there is any visible difference in the areas of clipping (which I can check in addition in PV2010 using the Uni-WB (2700, -150) which corresponds well with the Raw histogram).

Since the camera histogram is showing the jpeg, there should not be a correlation with truly blown highlights in LR.  This is the exact problem I mentioned in LR 3.  At the default settings, the red over exposure indications were not necessarily truly blown....and in most cases, most were not.

In 2012, all the red areas (you may need to blow up to 100 or 200%) to ensure all the area is red) will have RGB of 100, possible 99 which Eric said could be a rounding problem).  Any RGB. Pixel lower than that will not show red....unless you force it with slider changes.

I am kind of a pragmatic person.  While I like to understand what is happening under the covers, practically, if the area under consideration has correct color and textural detail, I really don't care how it was arrived at or whether it was partially blown and recovered.  I do want to know quickly what cannot be recovered so I do not waste my time in pp.

So I am happy with LR 4, even if I need to sorta guess how much I can push past the warnings of the jpeg blinkies and histogram.  A little exploring/comparing allows me to pretty much nail the ETTR for 2012.  If it is critical, I will bracket.
« Last Edit: May 12, 2012, 10:23:38 AM by jrsforums » Logged

John
Hans Kruse
Sr. Member
****
Offline Offline

Posts: 760



WWW
« Reply #30 on: May 12, 2012, 10:48:41 AM »
ReplyReply

Since the camera histogram is showing the jpeg, there should not be a correlation with truly blown highlights in LR.  This is the exact problem I mentioned in LR 3.  At the default settings, the red over exposure indications were not necessarily truly blown....and in most cases, most were not.

In 2012, all the red areas (you may need to blow up to 100 or 200%) to ensure all the area is red) will have RGB of 100, possible 99 which Eric said could be a rounding problem).  Any RGB. Pixel lower than that will not show red....unless you force it with slider changes.

I am kind of a pragmatic person.  While I like to understand what is happening under the covers, practically, if the area under consideration has correct color and textural detail, I really don't care how it was arrived at or whether it was partially blown and recovered.  I do want to know quickly what cannot be recovered so I do not waste my time in pp.

So I am happy with LR 4, even if I need to sorta guess how much I can push past the warnings of the jpeg blinkies and histogram.  A little exploring/comparing allows me to pretty much nail the ETTR for 2012.  If it is critical, I will bracket.

When setting the Uni-WB and the baseline offset there is a good correlation between tje JPG blinkies on the camera and what LR shows with PV2010. Try it. Getting the exposure right is not really difficult, I bracket for landscape shooting to get the optimal exposure, especially in high DR situations. But I do it most of the time anyway because this gives me the largest flexibility in post processing where I often edit a lot to get the look I'm after.

I'm also quite a pragmatic person, but I like to have the guess work out of the place if possible. In fact what I miss is RAW clipping indicator (different color) to overlay the normal LR colored clipping indicator area. In this way I would be able to see where the problems might be. I can use a virtual copy set with the above Wb, offset and PV2010, but then I have to switch back and forth between them. This is doable but a bit cumbersome.

Despite being pragmatic I like to know how things work, a left over from the IT profession in the past, I guess Smiley
Logged

jrsforums
Sr. Member
****
Offline Offline

Posts: 746


« Reply #31 on: May 12, 2012, 11:05:25 AM »
ReplyReply

I did try UniWB and did not like it (to each their own)....and some guess work still remained.  I do try to get the jpeg histogram closer (less guess) by lowering the saturation and contrast setting as much as possible...I should also lower. Sharpening, but actually raise it to better judge image.

At default settings, 2012 red indicators are the "RAW clipping indicators" at least as far as LR conversion is concerned.  In your case, however, I expect the UniWB throws that off.

John
Logged

John
Hans Kruse
Sr. Member
****
Offline Offline

Posts: 760



WWW
« Reply #32 on: May 12, 2012, 11:24:32 AM »
ReplyReply

At default settings, 2012 red indicators are the "RAW clipping indicators" at least as far as LR conversion is concerned.  In your case, however, I expect the UniWB throws that off.


No, that's not true and you can easily see this by looking at a RAW image in Rawdigger. I do not use UniWB when checking in PV2012 and in general do not use UniWB. Just checked it.
« Last Edit: May 12, 2012, 11:39:57 AM by Hans Kruse » Logged

jrsforums
Sr. Member
****
Offline Offline

Posts: 746


« Reply #33 on: May 12, 2012, 02:34:36 PM »
ReplyReply

No, that's not true and you can easily see this by looking at a RAW image in Rawdigger. I do not use UniWB when checking in PV2012 and in general do not use UniWB. Just checked it.

Not sure I understand or am missing your point.

I specifically said "as far as LR conversion was concerned".  I don't kow enough about intricacies of raw conversion, but I would think than comparing raw conversion between LR and RAwdigger would be apples and oranges.

John
Logged

John
Hans Kruse
Sr. Member
****
Offline Offline

Posts: 760



WWW
« Reply #34 on: May 12, 2012, 02:44:56 PM »
ReplyReply

Not sure I understand or am missing your point.

I specifically said "as far as LR conversion was concerned".  I don't kow enough about intricacies of raw conversion, but I would think than comparing raw conversion between LR and RAwdigger would be apples and oranges.

John

No, it is not apples and oranges. Rawdigger indicates the true histogram and where clipping occurs. That's why I'm saying that rather than using an approximation using PV 2010 and Uni-WB og base level offset, it would be much better to be able to see where the RAW clipping is since LR will not show clipping in areas where there are clipping in the RAW data. I hope it is clear now Wink Just having LR show where LR will clip because of the editing is not enough for me at least.
« Last Edit: May 13, 2012, 06:25:32 AM by Hans Kruse » Logged

BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3771


« Reply #35 on: May 12, 2012, 04:45:34 PM »
ReplyReply

[...] it would be much better to be able to see where the RAW clipping is since LR will not shot clipping in areas where there are clipping in the RAW data.

Yes, I agree with Hans. The Exposure control has stopped being a post-exposure gain control, and in PV2012 has been turned into a mid-tone control. It is not a reliable tool to check exposure anymore, because the connection of highlight rendering and underlying Raw data in not obvious. The simple solution is, like implemented in "RawTherapee", an optional Raw histogram (and with adjustable vertical scale).

Cheers,
Bart
Logged
aduke
Sr. Member
****
Offline Offline

Posts: 365


« Reply #36 on: May 12, 2012, 05:28:28 PM »
ReplyReply

It seems to me that a RAW histogram would be useful in LRx, where x > 4.

Alan
Logged
Pages: « 1 [2]   Top of Page
Print
Jump to:  

Ad
Ad
Ad