Pages: [1] 2   Go Down

Author Topic: Interesting LR Threshold Behaviour  (Read 7662 times)

luxborealis

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2798
    • luxBorealis.com - photography by Terry McDonald
Interesting LR Threshold Behaviour
« on: June 26, 2012, 10:52:07 am »

This is a long post with 6 photos, so bear with me...

Last night was a clear night so I thought I would put the D800e to the test by photographing the waxing crescent moon. I should have been out a 1/2 hour earlier to keep some of the dusk sky colour, but such is family life.

Settings: Nikkor 300mm ƒ4 AF-S IF-ED; ƒ5.6 @ 1/60; ISO200; no filter; on a sturdy tripod & head & MLU; NEF 14-bit raw capture. All images are screen captures of Lightroom 4.1 with each image zoomed to 100%

For comparison sake, here is the Unmanipulated raw file and what I have called "Normal" processing to enhance this specific image to my liking.





Frankly, I am amazed at the detail and sharpness of even the unmanipulated raw file at 100%. Also, there appears to be some significant Highlight "headroom" compared to Nikon's blinking highlights as the right of edge of the moon was blinking on the LCD preview. I'm aware of the inaccuracies of the Preview, but was surprised by how much is actually there "behind" the blinking highlights. When imported into LR, there were no clipped highlights - a function of LR's reading of the image file. While this is all fine and good, it is not the point of this post...

I came across some interesting thresholds with the Shadow and Highlight adjustments. I'm one of those persnickety types who do not use the sliders, but use the value boxes so that I can increase and decrease the values with the cursor keys with greater precision and consistency (+/–1 with the up/down Cursor key; +/–10 with Shift+Cursor Up/Down). I tend not to "creep up to" settings, but rather overshoot and come back (not unlike the days of test strips and variable contrast filtration in the enlarger!). I know, I'm anal!

So, here's what happened when I decreased Shadows by just –1 (from 0 to –1): this 1 point decrease in Shadows blew out the Highlights and marginally increased the Blacks. Hmmm...very interesting! Even more interesting is that there is no significant change decreasing by one point at a time until I reach -27 then boom: Highlights are in balance again (as per Normal above). Then at -28, boom: the Highlights are back to being blown out and continue that way until -35 when they instantly go back to "Normal". From -36 to -38 the Highlights gradually brighten, then at -39: boom they are back to being blown out and stay they way through to -100. Remember - this is the Shadow adjustment I am changing, yet its affect is greatest on the Highlights - go figure!

Here is the "Normal–1 Shadows" screen capture:


So, after investigating changes to Shadow adjustment, I turned my attention to Highlight adjustment. Here's what happened when I increased Highlights by just +1 from "Normal" (from -48 to -47): this 1 point increase blew out the Highlights and raised the Shadows/Blacks significantly.



And this is what happened when I decreased the Highlights by just –1 from "Normal" (from -48 to -49): the Highlights actually brightened slightly:



Finally, when I decreased Highlights by –3 from "Normal" (from -48 to -51): the Highlights blew out entirely with a moderate gain in Blacks. A further decrease in Highlights erred to increase the Blacks/Shadows until around -60, then the Highlights started decreasing again back to "Normal" at around -68, but the Blacks/Shadows were still elevated producing a noisy blue sky. From -68 to -100, the Highlight adjustment continued to behave normally in that it gradually dressed the Highlights into a muddy grey.

Here's the Normal–3 Highlights" capture:


Bizarre, at least to my mind. Perhaps someone in the know can explain what's going on here with respect to these seemingly bizarre behaviours at certain and apparently random thresholds.
Logged
Terry McDonald - luxBorealis.com

Slobodan Blagojevic

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 18090
  • When everyone thinks the same, nobody thinks
    • My website
Re: Interesting LR Threshold Behaviour
« Reply #1 on: June 26, 2012, 11:38:21 am »

Had it been a full moon, the bizarre behavior could be explained easily ;)

madmanchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2115
    • Web
Re: Interesting LR Threshold Behaviour
« Reply #2 on: June 26, 2012, 08:19:05 pm »

The issue is that you are adjusting the sliders out of order, e.g., you were adjusting Highlights/Shadows even though Whites/Blacks had already been adjusted (e.g., Whites +60).  The sliders are generally arranged so that the later sliders depend on the earlier ones, e.g., the range and methods used to implement Whites/Blacks depends on what you've set Highlights/Shadows to.  Thus I suggest (generally) editing top-down.
Logged
Eric Chan

jljonathan

  • Full Member
  • ***
  • Offline Offline
  • Posts: 144
Re: Interesting LR Threshold Behaviour
« Reply #3 on: June 26, 2012, 10:26:38 pm »

Interesting comment; I always thought that no matter what order you adjust, LR would do its own thing undercover. I'm not sure how I would adjust Highlight/Shadow before White/Black; I usually do these first and then I see whether I need to move the shadows/highlights. Anyone have any thoughts on this slider sequence information?
Logged

wolfnowl

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5824
    • M&M's Musings
Re: Interesting LR Threshold Behaviour
« Reply #4 on: June 27, 2012, 01:36:47 am »

Yes... same question here!  I was given to understand that the 'top down' approach was considered 'usual' for lack of a better word, but not that failure to do so would give such dramatic errors.

Mike.
Logged
If your mind is attuned t

Schewe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6229
    • http:www.schewephoto.com
Re: Interesting LR Threshold Behaviour
« Reply #5 on: June 27, 2012, 02:17:37 am »

I'm not sure how I would adjust Highlight/Shadow before White/Black; I usually do these first and then I see whether I need to move the shadows/highlights.

Eric is the final arbitrator on this question but...there is a real good reason why the controls are set in the order they are in PV 2012. That is the optimal order based on the image adaptive, auto ranging controls. Yes, you can set Whites and Blacks first if you want but...you will end up bouncing back and forth between ALL of the Basic panel adjustments over and over if you do.

There really is a logic to the top down (not bottom up) nature of PV 2012. You may not like it if you've adopted a hybrid non-traditional approach to tone mapping. But hey, that's life, things change, learn to adapt. Exposure is no longer the optimal place to set clipping in whites but the best place to set the overall brightness of the image. Next up is Highlights and Shadows to fine-tune the 1/4 and 3/4 tone areas of the image. The last settings to use are Whites and Blacks to finesse (only if needed and less so in PV 2012) the exact clipping points.

Look, you can try to fight against the changes (many did on the LR 4 beta forums) but it's easier to completely divorce yourself from your pre PV 2012 approach and adopt a top down approach and take advantage of the PV 2012 behaviors. Mike hated PV 2012 when he first saw it because, well it was so totally different in behavior. The more advanced a user is, the more they tend to cling to past behavior expectations and that's not an optimal strategy for using PV 2012. You are much better off forgetting everything you thought you know and start over. Do that and you'll be able to adapt pretty quickly (and get better results). Reject that approach and you'll be fighting against the new controls and lamenting the loss of the old approach–which will only frustrate you further. Move on and move up...

To the OP...the "default" rendering looked pretty darn good...yes, flailing about with other setting can indeed make it look less-good (which is true with most image adjustments). The question I'll ask if what, EXACTLY did you hope to achieve over the original Default rendering? Sure looks like the flailing about didn't help much (although you got nice and confused, right?).
« Last Edit: June 27, 2012, 02:22:22 am by Schewe »
Logged

John R Smith

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1357
  • Still crazy, after all these years
Re: Interesting LR Threshold Behaviour
« Reply #6 on: June 27, 2012, 03:22:18 am »


The above reply from Jeff pretty much summarises all the reasons that (after several weeks with LR 4.1 RC) I shall not be upgrading to LR 4.

LR 3 may be not so clever in absolute terms, but at least it doesn't try to do things behind your back. Image adaptive? No thanks. I'll do the adapting, if you don't mind. And of course you should be able to set the end points of the histogram and your clipping threshold with the White and Black sliders. Any sane person would expect it.

John
Logged
Hasselblad 500 C/M, SWC and CFV-39 DB
an

Schewe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6229
    • http:www.schewephoto.com
Re: Interesting LR Threshold Behaviour
« Reply #7 on: June 27, 2012, 03:31:54 am »

Any sane person would expect it.

That's a very backwards approach...you really need to accept the next-gen image adaptive behaviors because, well it's the next step for really advanced image adjustments. In the photo industry it's expected that a major change and improvement is made to radically advance what you can "get" out of your raw images....and PV 2012 is no exception.

You may "think" you don't want image adaptive, auto ranging adjustments but I think if you try to ignore the advance, you'll be stuck in a non-progressive situation you'll end up regretting...seriously, I think you need to rethink your ludite approach. Do you care about preserving your approach to image processing or do you care about extracting the best image you can create? If the latter, you have no choice but upgrade...if the former, you are welcome to stick your head in the sand and pretend stuff isn't advancing. yeah, I know, upgrading means learning new shit...tough, that's called progress and you either progress or stagnant–your choice.
Logged

John R Smith

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1357
  • Still crazy, after all these years
Re: Interesting LR Threshold Behaviour
« Reply #8 on: June 27, 2012, 04:15:06 am »

Jeff

I was perfectly content with the tools available to me with film and the wet darkroom, forty years ago. I look back at my pictures from those times and I still enjoy them. All I ask from my gear now is to be able to do as well as I did then, but using digital "negatives" and printing with ink rather than chemistry. For almost ten years I used Corel Photopaint 9 on Win 98 and XP and it did everything I wanted it to do with scanned film TIFFs. The only reason I moved to LR was because I got a digital back and needed to work with the new RAW files.

When I work on my pictures in LR I am not in a state of mind where I am looking to do something "new" or "better" or "different" with my pictures. All of that stuff happens when I am out in the countryside with my camera. All I want to be able to do is develop and print a B/W picture as near as possible to the same way that I could in the wet darkroom. With LR 2 and LR 3 I felt that I was pretty close to that. Not so with LR 4.

And yes, I live in a two-hundred year old cottage, love classic cars and motorcycles, listen to music through valve amplifiers and I do not possess a mobile phone. So I think I am indeed a luddite. But that does not make me or my opinions irrelevant  ;)

John
« Last Edit: June 27, 2012, 04:17:03 am by John R Smith »
Logged
Hasselblad 500 C/M, SWC and CFV-39 DB
an

Brad Smith

  • Jr. Member
  • **
  • Offline Offline
  • Posts: 97
Re: Interesting LR Threshold Behaviour
« Reply #9 on: June 27, 2012, 07:02:25 am »

OK--so we are to do things in order and Exposure, for midtone adjustment, is the first slider after WB . . .

To my eye, the OP's "Normal" image is quite and improvement over the "Unmanipulated". I recognize these are only screen captures, but they seem nevertheless quite informative. When I compare these first two images, I would not have guessed that my first step, to arrive at the "Normal" improvement, should have been to INCREASE the Exposure setting (much less to do so by +2.5 stops). The histogram, to me, offers little help, because there are so few pixels in the midtones. So, I ask, if we are indeed supposed to adjust the Exposure slider first, what would have been my clue that this would have been a good move--trial and error? And if the answer is by trial and error, could I have arrived at +2.5 Exposure without adjusting the other sliders as well, but then you are into the back and forth game anyway??

Thanks in advance to anyone who can clarify.
Logged

luxborealis

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2798
    • luxBorealis.com - photography by Terry McDonald
Re: Interesting LR Threshold Behaviour
« Reply #10 on: June 27, 2012, 07:11:00 am »

To the OP...the "default" rendering looked pretty darn good...yes, flailing about with other setting can indeed make it look less-good (which is true with most image adjustments). The question I'll ask if what, EXACTLY did you hope to achieve over the original Default rendering? Sure looks like the flailing about didn't help much (although you got nice and confused, right?).

Thanks for responding Jeff and Eric. Your views on this forum are not only helpful, but really complete the picture of what LR is doing under the hood.

I mostly (not completely) understand what is happening with LR and its image adaptive behaviours and have been a strong proponent of LR4 as I teach it at a nearby college.

What I take issue with is the tone of your reply. I have not criticized LR or you, merely pointed out an interesting behaviour. You seem to assume that when ever someone has a comment about LR4, they are a rank amateur with no background knowledge. Making orderly changes in values to see what would happen to learn more about how an app reacts is not "flailing about"!

What, EXACTLY did I hope to achieve over the original Default rendering? - Like many, I am still learning the nuances of PV2012. To do so, I need to put it through its paces to see how far it can be pushed. Once I learn about the boundaries of any tool, be it hardware or software, I can then work within and with those limits and sometime push those limits still further.

I've learned something from this experimenting and I hope others have as well. It's good to experiment and share results!

Logged
Terry McDonald - luxBorealis.com

stamper

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 5882
Re: Interesting LR Threshold Behaviour
« Reply #11 on: June 27, 2012, 09:05:49 am »

Quote  luxborealis

Finally, when I decreased Highlights by –3 from "Normal" (from -48 to -51): the Highlights blew out

unquote

I am like the rest trying to get my head around these controls. My reading of the controls is if you go from -48 to -51 you aren't decreasing anything but increasing. You are adding illumination to the lighter tones. Decreasing, going from -51 to -48, would be decreasing and "darkening" the lighter tones and possibly adding detail. I am stating this in the context of clouds. There has been comment on other forums about users using the controls in reverse. Some of the controls in version 2012 should be used in reverse of 2010?

kwalsh

  • Full Member
  • ***
  • Offline Offline
  • Posts: 101
Re: Interesting LR Threshold Behaviour
« Reply #12 on: June 27, 2012, 10:32:25 am »

EDIT: I'm leaving the post below unedited, because I think it reflects an honest concern about how the adaptive controls work.  However, on further reflection while making my second post of questions I now wonder if the "answer" is that users would only end up with control settings that could provoke such behavior if they start their editing by adjusting sliders out of order.

Is that what Eric and Jeff have been driving at?  Yes, the UI can become oversensitive, but only if you do big adjustments out of order first?  That gets you into a point of astablility as far as the adaptive controls go?

I think I may have been misinterpreting the comments.  I was reading "if you make minor adjustments out of order things will blow up" and that really seems like wrong behavior to me.  But maybe I'm missing the bigger picture?  Are Jeff and Eric really saying "if you make your big movements out of order then all bets are off for future editing and small adjustments can blow up, but if instead you make your big movements in order then everything will play well and small movements will result in small changes to the image as expected"?

--------------------------------------------

With respect to both Eric and Jeff, you can talk about the order of controls all you want but any slider based system in which a control movement of +1 or -1 takes an optimal image to a completely broken image is a system with a UI flaw.  Plain and simple, no way to weasel word around it.  That is broken - period, end of story.  Maybe it is coded properly, maybe the algorithm is working as intended, but the final behavior is most definitely broken - either by design or implementation.

Both of you, usually so extremely helpful, and for which we are all deeply grateful, seem to either be not reading or not thinking very clearly in this case.  Go back, look at the examples, and tell me if honestly you think any user interface in the world should be so sensitive to someone tweaking a slider by such a vanishingly small amount regardless of the "preferred" order.  He moved the slider by 0.5% and the image exploded - think about that and its implications.  Honestly ask yourself if "out of order" slider movement is really an adequate explanation or if any user should really be expected to anticipate or tolerate such behavior.

PV2012 is awesome, and I am very, very impressed and pleased that such a functional adaptive algorithm was implemented "real time" in a tool like LR.  If I had even the tiniest association with such an accomplishment I would be very proud (and justifiably so).  Don't let that justifiable pride blind you to what seems to be a pretty clear example of the UI behaving in a pathological way.

Ken

P.S. And still, thank you both very much for responding to the OPs post.  I'm sure you are both correct that adjusting in a different order would give a different result and it is so helpful to have two folks deeply involved in the software development to post and point out a likely solution.
« Last Edit: June 27, 2012, 11:10:12 am by kwalsh »
Logged

kwalsh

  • Full Member
  • ***
  • Offline Offline
  • Posts: 101
Re: Interesting LR Threshold Behaviour
« Reply #13 on: June 27, 2012, 10:57:11 am »

And some actual questions (instead of just unsolicited and perhaps unwelcome criticism), so I can hopefully learn as much as possible from this example.

I'm guessing what is making this particular example so sensitive to small changes is the fact Highlights is very negative while whites is very positive?  Presumably this also has to do with the scene data itself, these particular control values might not be super sensitive on a different image?

Also, the presumption here is that if the controls had been adjusted in the intended order it is unlikely the user would have ever ended up at this setting?  They would have ended up at a similar rendering but with different control settings that were less sensitive to such small changes for this particular image?

Does this imply we should avoid positive/negative splits like this, or if we were to arrive at such control points having adjusted things in the intended order would such splits be OK (i.e. by OK I mean not result in pathological sensitivity to fine adjustments).

I know adaptive controls while sometimes twitchy can also be very powerful, this could be a particularly good example for illustrating what to do and not to do.

Thanks,

Ken
Logged

Schewe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6229
    • http:www.schewephoto.com
Re: Interesting LR Threshold Behaviour
« Reply #14 on: June 27, 2012, 12:45:52 pm »

What I take issue with is the tone of your reply. I have not criticized LR or you, merely pointed out an interesting behaviour. You seem to assume that when ever someone has a comment about LR4, they are a rank amateur with no background knowledge.

Hum...I just reread what I wrote and I fail to see any "tone" in my post. Exactly how did I offend you?
Logged

Schewe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6229
    • http:www.schewephoto.com
Re: Interesting LR Threshold Behaviour
« Reply #15 on: June 27, 2012, 01:03:04 pm »

I'm guessing what is making this particular example so sensitive to small changes is the fact Highlights is very negative while whites is very positive?  Presumably this also has to do with the scene data itself, these particular control values might not be super sensitive on a different image?

Correct...based on the image at default you have a ton of black with zero detail and a relatively small area of textured highlights and a small sliver of midtones. How the OP arrived at his "normal" settings, I don't know but a +2.5 exposure combined with a +60 whites sure seems like an extreme push. Due to the image adaptive nature of the controls that makes the highlights slider very, very tweaky.

Quote
Also, the presumption here is that if the controls had been adjusted in the intended order it is unlikely the user would have ever ended up at this setting?  They would have ended up at a similar rendering but with different control settings that were less sensitive to such small changes for this particular image?

That would be my guess...

Quote
Does this imply we should avoid positive/negative splits like this, or if we were to arrive at such control points having adjusted things in the intended order would such splits be OK (i.e. by OK I mean not result in pathological sensitivity to fine adjustments).

I'm not so sure the "order" of the adjustments really plays a role here...I think it's more the contradictory nature of a huge +whites adjustment combined with a strong -highlights adjustment that is sort of "breaking" the image tones when other adjustments are made. This is also a case where less extreme basic panel controls combined with fine tuned point curve edits might have been less twitchy. Heck, throw in a touch of adjustment brush for good measure.
« Last Edit: June 27, 2012, 01:04:57 pm by Schewe »
Logged

kwalsh

  • Full Member
  • ***
  • Offline Offline
  • Posts: 101
Re: Interesting LR Threshold Behaviour
« Reply #16 on: June 27, 2012, 03:42:49 pm »

Thanks Jeff!

I actually missed the exposure setting.  You're right, that would make things even more wonky. 

My interpretation (possibly flawed) is that the extreme exposure setting is pushing the midtones way up, but some adaptive fanciness is protecting the whites/highlights at the same time.  Next, the highlights setting is whipped rather negative and if I understand correctly this can affect a rather narrow range of tones and is perhaps stacked in a difficult to predict way on top of the highlight protection already in place with the big exposure push.  Then the whites are cranked the other way again, and if I understand correctly while this is focused on the extreme whites it does interact with other tones a bit (more so than the highlights possibly).  That fairly extreme zig-zag of plus, minus, plus on the mid-tones, highlights and whites is really creating a nearly pathological tone correction.

So with a better understanding of what was done to this image I'm going to revise my earlier editorial comment.  Yes, having such a tiny control movement "blow up" an image is not a desirable behavior in a UI.  On the other hand, I know from experience developing such a UI with a lot of fairly complicated and interacting processing behind it is a task that rarely can be optimized over the entire range of theoretically possible inputs.  It is more valuable to make sure it is well behaved over the ranges of inputs that are practical.  I can see better now that since there are three control inputs with fairly extreme and "zig-zagging" settings that this case is probably well outside the typical or practical (even though at a particular setting acceptable output was generated).  So we've got pathological output, but it seems to be based on pathological control settings.  Sure, it would be nice if even pathological controls gave smooth and stable outputs but the reality is the control tuning should be best optimized for the "useful" settings.

This is often a pitfall of controls with interactions.  If the user isn't familiar enough with them, or the intended use, it is easy to start "turning the wrong knob".  The user is seeing an affect on what they are hoping to control, but the affect is indirect and may create pitfalls elsewhere.  So, answering my own question, yeah I think I can see how doing things "in the wrong order" are much more likely to lead you into a pathological case.

I've been really impressed with the PV2012 controls so far, even doing some impressive HDR work, but I've been using them "as recommended" which is probably why I haven't had problems.  I suspect one of the biggest causes of "problem" settings is trying to control the whites with the exposure slider - that's probably second nature to a lot of people these days from a lot of RAW converter workflows but isn't the way to do it with PV2012.

Thanks for taking the time!

Ken
Logged

luxborealis

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2798
    • luxBorealis.com - photography by Terry McDonald
Re: Interesting LR Threshold Behaviour
« Reply #17 on: June 27, 2012, 06:03:52 pm »

Hum...I just reread what I wrote and I fail to see any "tone" in my post. Exactly how did I offend you?

No offense – it's actually a small thing. I just took issue with your reference to me "flailing around" a term often  associated with someone who is clearly not in control of what they are doing. On the contrary, I knew exactly what I was doing – a very controlled progression from lesser to greater and vice versa – as I took the time to explain in my OP. We all approach things slightly differently – thinking outside the box may at times seem contrary but if we all thought the same then we wouldn't really be thinking at all.

Quote  luxborealis
Finally, when I decreased Highlights by –3 from "Normal" (from -48 to -51): the Highlights blew out
unquote

I am like the rest trying to get my head around these controls. My reading of the controls is if you go from -48 to -51 you aren't decreasing anything but increasing. You are adding illumination to the lighter tones. Decreasing, going from -51 to -48, would be decreasing and "darkening" the lighter tones and possibly adding detail. I am stating this in the context of clouds. There has been comment on other forums about users using the controls in reverse. Some of the controls in version 2012 should be used in reverse of 2010?

Stamper: -48 to -51 is a decrease in value (moving further away from zero in negative numbers) whether it is Highlights, Shadows or even temperature
Logged
Terry McDonald - luxBorealis.com

Schewe

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6229
    • http:www.schewephoto.com
Re: Interesting LR Threshold Behaviour
« Reply #18 on: June 27, 2012, 06:10:43 pm »

My interpretation (possibly flawed) is that the extreme exposure setting is pushing the midtones way up, but some adaptive fanciness is protecting the whites/highlights at the same time.  Next, the highlights setting is whipped rather negative and if I understand correctly this can affect a rather narrow range of tones and is perhaps stacked in a difficult to predict way on top of the highlight protection already in place with the big exposure push.  Then the whites are cranked the other way again, and if I understand correctly while this is focused on the extreme whites it does interact with other tones a bit (more so than the highlights possibly).  That fairly extreme zig-zag of plus, minus, plus on the mid-tones, highlights and whites is really creating a nearly pathological tone correction.

Yep...that's how I read the settings as well. There are two settings driving midtones, highlights and whites way up and a minus highlights driving it down and I think that's where the breakage happens and it certainly could be a failure in the math of the adaptive algorithms. You can see similar results using multiple extreme points on a curve. It's ok up to a certain point and then past that point it blows up...
Logged

luxborealis

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2798
    • luxBorealis.com - photography by Terry McDonald
Re: Interesting LR Threshold Behaviour
« Reply #19 on: June 27, 2012, 06:13:35 pm »

Another take on this...

So how did I start with an unmanipulated image that had bright, but not clipped highlights, and actually end up adding 2.5 stops of exposure?

I broke the rules (sorry Jeff and Eric). I know that with LR4 PV2012 I am supposed to start at the top (Exposure) and work my way downwards. However, I also know from experience and from other users (a few who have posted here) that improved highlight separation can be achieved by bumping the Whites adjustment high in the positive range and setting the Highlights adjustment low in the negatives. I didn’t quite need to do that here: the Whites were already high with the Exposure increase increase to 2.5 so all I needed was to tame the Highlights and that I did down to –48. Note that Highlights were literally blown out at everything up to –47, then at –48 they magically came into alignment, only to be blown out again as I dropped the Highlight adjustment lower (tell me that makes sense: lower Highlight value = brighter Highlights – go figure!).

Anyway – I’ve taking Jeff and Eric’s advice and went back to the drawing board and started with a Virtual Copy of the unmanipulated version from my OP…
  • Clearly, Exposure needs no increase, so I started by adding Contrast up to 50 to create better tonal separation.
  • The Highlights clearly need taming, especially after the Contrast increase, so I dialed them down to –20.
  • That had a darkening effect on Shadows so I upped them to +50.
  • Now, the mid-tones in the craters were still darker than I wanted, so I raised them using +1.00 Exposure (as that is the adjustment for mid-tones). I know, it’s out of order, but what to do. Any app must have the flexibility to allow for at least some back and forth.
  • So a final tweaking of Highlights down to –33 and a bit of “punch” added using Clarity and voila – a new and improved version. It’s not quite as contrasty in the highlights; i.e. highlight separation is not as good.

So, then i consider how to further increase highlight separation? Well, the trick in LR4 PV2012 is to increase the Whites and further decrease the Highlights adjustments. Doing just that (Whites to +30; Highlights down 11 to –44) did separate the highlights more, but also caused the mid-tones to drop. To bring mid-tomes back up again, I raised Exposure by 0.6 to 1.6 – visually speaking, we’re now getting somewhere. A final tweak of Highlights to –48, a brightening of Shadows to +100,  and Whites down to +25 and I have an image I can be happy with, without the bizarre, sudden and unexpected lurches experienced and described previously.

Now could I have gotten to this point in 5 or 6 deft adjustments; i.e. without “flailing around”? – probably not. I defy anyone to achieve such a feat (except perhaps Jeff and Eric). This is not a straightforward image and I had a pretty clear idea of how I wanted the tones in the Moon to be rendered. And now I’m there and significantly richer for the learning experience. Many thanks!

Logged
Terry McDonald - luxBorealis.com
Pages: [1] 2   Go Up