Ad
Ad
Ad
Pages: [1] 2 »   Bottom of Page
Print
Author Topic: PV 2010 /2012 new tonal controls  (Read 17455 times)
Peter_DL
Sr. Member
****
Offline Offline

Posts: 421


« on: May 15, 2012, 11:41:42 AM »
ReplyReply

  
Just started to try to compile some related information,
from current articles and earlier discussions (as referenced below).

From left to right along the tonal scale:
PV 2010:  Blacks – Fill Light – Brightness – Recovery – Exposure.
PV 2012:  Blacks – Shadows – Exposure – Highlights – Whites.

/>  The former Exposure slider was renamed to Whites.
Still, the slider moves the White point.  Not sure though, if it is really just a rename with regard to operations in terms of linear scaling, corresponding to +/- EV camera exposure (need to check).


/>  Recovery and Fill Light were replaced by improved functions called Highlights and Shadows.  
With the new Highlights control there is some automatic Highlight recovery going on, even without moving the slider away from zero, or using the Auto button.
With the new Shadows slider in PV 2012 its effect got more limited to the darker tones, while applying an improved algorithm to brighten without posterization.

/>  Blacks is called Blacks again, now featuring an auto-calculated Black point at default zero setting.
Still, the slider moves the Black point. Not sure though about possible changes under the hood with regard to operations in terms of applying a linear offset (need to check). Interestingly, the new Blacks slider was shown to bear a reduced side effect on color saturation.

/>  Former Brightness and Exposure sliders were hybridized to the new Exposure slider,
which now mainly controls the midtones, but also moves the White point.  It is to note that already at default zero setting* the highlights are compressed, and further with positive settings to brighten the image.

/>  The Contrast slider is now scene-dependent, offsetting its operational midpoint depending on whether you are editing a low key or high key image.


*PV 2010's default settings had Brightness +50, Contrast +25, Blacks +5, and a "medium contrast" Point curve applied.  With PV 2012, results are quite similar with all the sliders left at zero now, and Point curve = Linear.  The deviations are due to automatic Highlight recovery, and the auto-Black-point functionality, which are always enabled behind the scenes.
Anyway, the default settings impose an "S" curve on the tones, which already compresses highlight details [which is probably the main reason why some of us liked to start with a linear, flat, scene-referred rendition in the past. A subject on its own, I guess].

All six tonal controls in PV 2012 are image adaptive. They auto-adjust their behavior internally based on image content, just like Recovery and Fill Light (and also Clarity) already did with PV 2010. In PV 2012 this idea was extended to the rest of the central Basic controls from Exposure to Blacks.


Comments and additions are welcome.

Regards, Peter


--

Charles Cramer, Tonal adjustments in the age of Lightroom 4
http://www.luminous-landscape.com/techniques/tonal_adjustments_in_the_age_of_lightroom_4.shtml

Martin Evening, Extreme contrast edits in Lightroom 4 and ACR
http://www.dpreview.com/articles/1205103502/extreme-contrast-edits-in-lightroom-4-and-acr-7

Why "0" should be Zero
http://forums.adobe.com/message/4289492#4289492
http://forums.adobe.com/message/4128876#4128876
http://forums.adobe.com/thread/948951

ETTR, Blacks & Exposure settings vs. Point Curve controls (PV 2010)
http://www.luminous-landscape.com/forum/index.php?topic=57177.0

--
late edit: cancelled some lines for archival purposes
« Last Edit: June 04, 2012, 11:52:28 AM by Peter_DL » Logged
Tim Lookingbill
Sr. Member
****
Offline Offline

Posts: 1230



WWW
« Reply #1 on: May 18, 2012, 04:25:45 PM »
ReplyReply

Another thorough and concise tome of info on an important subject for digital photographers.

Nice write up, Peter. Thanks.

Took the entire afternoon to check out all those links but it was worth it because it tells me a wide range of cameras record scenes pretty much similar to one another no matter their price tag. I have to deal with same tonal mapping issues with my old $500 Pentax K100D PEFs whether with PV2003 and PV2010 in CS3 and CS5 respectively.

I used to use a Stouffer gauge back in prepress darkroom days in the '80's and I noticed in the Charles Cramer article the distribution differences from the spikes representing the mid to highlights compared to the shadows which show no separation between the steps (visually looking at the tones) even though it's there.

It would be great to understand why this is as a way to define some repeated pattern that could be compensated for to guide our adjustments without all the "zeroing" of default settings. PV2012 is a marked improvement especially evident in Martin Evening's backlit column demo. Those are some nasty halos I've become quite familiar with.

What I'm surprised is never covered in those articles is the mysterious nether regions of just exactly where absolute black resides (and shown on the display) and how the tones just next to it are affected by Fill, Black, Linear Point Curve slide to the right and most especially both Shadow and the far left triangle adjust in the Parametric Curve.

For instance in Charles Cramer's expose for clouds backlit forest shot all those brightened up trees could be even more defined by using the Parametric Shadow curve to isolate the edges of those trees and still keep the soft foggy look. The thing is similar images in my experience would show RGB readouts close to 5,5,5 on the edges of those trees but the Parametric Shadow curve seems to be able to pull incredible separation even from tones close to black I couldn't accomplish with any other tool.
« Last Edit: May 18, 2012, 04:30:08 PM by tlooknbill » Logged
Schewe
Sr. Member
****
Offline Offline

Posts: 5530


WWW
« Reply #2 on: May 18, 2012, 05:16:16 PM »
ReplyReply

/>  The former Exposure slider was renamed to Whites.

snip

/>  Former Brightness and Exposure sliders were hybridized to the new Exposure slider,
which now mainly controls the midtones, but also moves the White point.  It is to note that already at default zero setting* the highlights are compressed, and further with positive settings to brighten the image.

These two statements conflict...

In point of fact, Exposure was not renamed Whites...Whites is a new tone control specifically designed to fine tune the placement of the white clip point on top of Exposure. So, the first statement is wrong while the second is the correct description of the new Exposure.
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2861



« Reply #3 on: May 18, 2012, 08:53:10 PM »
ReplyReply

These two statements conflict...

In point of fact, Exposure was not renamed Whites...Whites is a new tone control specifically designed to fine tune the placement of the white clip point on top of Exposure. So, the first statement is wrong while the second is the correct description of the new Exposure.

I agree that Charles Cramer's post is masterful and very useful. In my own analysis of PV2012 I used a 40 step wedge, whereas Charles used a 20 step wedge. In the 40 step wedge the steps are in 0.3 optical density decrements, so 3 steps equals 1 f/stop. The values of the wedge are shown below, where step 1 has an OD of 0.05, placing it at the values shown. Step 8 is neutral gray.



I exposed so that step one was just short of clipping in the green channels as shown by Rawdigger.



On rendering the image in ACR 7 using the default settings, the highlights appear overexposed as shown in the ACR preview and histogram. Step 8 (neutral gray) is at 202 in ProPhotoRGB, whereas it should be around 98. As Charles suggests, one should set the midtones with the exposure slider, and this required a negative exposure value of -1.65 EV.



The highlights are too dark and the Whites slider can be used to set step 1 to the nominal value. The highlight control extends too far down the tonal scale and affects the midtones. This necessitates an adjustment of the midgray as shown.



The quarter tones are too light and these can be adjusted with the contrast control, which has minimal effects on the midtones and whites. Several iterations are necessary to get approximate values. This results in an approximately linear tone curve.



One can do further analysis with Imatest.



With PV2010 a linear curve is more easily obtained by setting the point curve to linear and the sliders on the main tab all to zero.



For ETTR exposures it is important that negative exposure to normalize the image are linear, and Eric Chan has stated in a previous post that this is the case. However, increasing the exposure rolls off the highlights so as to prevent clipping. This is shown in the composite graph below, using the same raw file as before. The adjustment is essentially linear except for the highlights.



Here are similar graphs for the highlight, whites and contrast controls.







Regards,

Bill





Logged
stamper
Sr. Member
****
Online Online

Posts: 2870


« Reply #4 on: May 19, 2012, 03:06:08 AM »
ReplyReply

I didn't think that the human eye could see 40 steps so 20 would be a more reasonable representation? The problem with all these charts is that you don't remember them when you are processing an image? Charles article would be a more practical workflow, imo. Smiley
Logged

BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3865


« Reply #5 on: May 19, 2012, 04:35:44 AM »
ReplyReply

I agree that Charles Cramer's post is masterful and very useful. In my own analysis of PV2012 I used a 40 step wedge, whereas Charles used a 20 step wedge.

Hi Bill,

Thanks for an excellent analysis and I'm glad you were also inspired by Charles Cramer's article. The added precision of finer step density increments allows to make a more accurate analysis of the 'cross-talk' between the sliders, requiring to go back and forth between controls for a more targeted result, and it allows a more detailed/accurate adjustment of the different controls.

It also demonstrates something that Christian Bloch also mentioned in the context of preparing HDR bracketed exposures, PV2012 requires a negative 'exposure' setting for starters for a more 'quasi-linear' tonecurve, but PV2010 allows a much easier route for that specific application (which requires reconstruction/calibration of the bracketed tonecurve with a glare component in each bracket). PV2012 is great for a creative approach (including exposure blending), but a nightmare for the more controlled types of photography, e.g. reproduction and somewhat scientific work which requires predictable (calibrated) response and linear tonecurves.

The only thing I missed in your post is which Camera calibration type your used. In LR I usually start with "Camera Neutral" instead of "Adobe Standard" because traditionally the color response for Canon cameras sucked. It would be interesting to find out how much of an influence, if any, that calibration choice has on overall tonality, besides the obvious effect it should have on color reproduction.

Again, thanks for a great contribution.

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

Posts: 2861



« Reply #6 on: May 19, 2012, 07:41:06 AM »
ReplyReply

I didn't think that the human eye could see 40 steps so 20 would be a more reasonable representation? The problem with all these charts is that you don't remember them when you are processing an image? Charles article would be a more practical workflow, imo. Smiley

The 40 steps with a Dmax of 4.0 are necessary to capture the full DR of current cameras, covering a DR of 13 stops. For rendering of an image with high DR one would have to use tone mapping to compress the image to something that could be visualized on screen or a print. One does not have to use all 40 steps. I used only the first 30 for my analysis. The 20 (actually 21) step wedge has a DMax of only 3.0 for a DR of 10 stops.

Charles' article and approach is better for visual editing of images in ACR/LR, but my intent was to quantify what he was seeing on the histogram.

Regards,

Bill
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2861



« Reply #7 on: May 19, 2012, 07:55:44 AM »
ReplyReply

Thanks for an excellent analysis and I'm glad you were also inspired by Charles Cramer's article. The added precision of finer step density increments allows to make a more accurate analysis of the 'cross-talk' between the sliders, requiring to go back and forth between controls for a more targeted result, and it allows a more detailed/accurate adjustment of the different controls.

Thanks for the kind words.

It also demonstrates something that Christian Bloch also mentioned in the context of preparing HDR bracketed exposures, PV2012 requires a negative 'exposure' setting for starters for a more 'quasi-linear' tonecurve, but PV2010 allows a much easier route for that specific application (which requires reconstruction/calibration of the bracketed tonecurve with a glare component in each bracket). PV2012 is great for a creative approach (including exposure blending), but a nightmare for the more controlled types of photography, e.g. reproduction and somewhat scientific work which requires predictable (calibrated) response and linear tonecurves.

With PV2010 and "linear" ACR settings one does get linear results and this enables one to produce a scene referred image using the method outlined by the ICC. It is really a nightmare to get such results in PV2012. Fortunately, Adobe was wise to retain PV2010 and Jeff Schewe indicated that future versions will also retain it.

The only thing I missed in your post is which Camera calibration type your used. In LR I usually start with "Camera Neutral" instead of "Adobe Standard" because traditionally the color response for Canon cameras sucked. It would be interesting to find out how much of an influence, if any, that calibration choice has on overall tonality, besides the obvious effect it should have on color reproduction.

I used the Adobe Standard calibration for my work. One could look at the tone curve in the DNG profile editor, but that is for another day.

Regards,

Bill
Logged
paul_o
Newbie
*
Offline Offline

Posts: 6


« Reply #8 on: May 19, 2012, 08:25:34 AM »
ReplyReply

I really like the tonal controls of the new ACR.  The one troubling area is when I want to use to new controls on images that previously worked on with PV 2010.  In some cases, when I update to PV 2012 the tonal changes are dramatic.  I’ve found it easier to use the original image and start from scratch. Except that some images have had significant spot removal, which now needs to be redone. It would be desirable to be able to retain the effect of some tools when converting to PV 2012.   Ideally (for me) it would be desirable to be able to have all the basic sliders set as close as possible to retain the tones that were set in PV 2010 and then be able to use the new PV 2012 to make additional fine-tuning adjustments.
Is anyone else having this issue? 
Regards,
Paul
Logged
Tim Lookingbill
Sr. Member
****
Offline Offline

Posts: 1230



WWW
« Reply #9 on: May 19, 2012, 10:36:17 AM »
ReplyReply

Will someone address the reasons behind the lack of definition and separation of the steps below middle gray compared to above middle gray on the Stouffer Gauge?
Logged
BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3865


« Reply #10 on: May 19, 2012, 10:38:44 AM »
ReplyReply

Will someone address the reasons behind the lack of definition and separation of the steps below middle gray compared to above middle gray on the Stouffer Gauge?

Hi,

Caused by gamma adjustment + noise?

Cheers,
Bart
« Last Edit: May 19, 2012, 10:42:35 AM by BartvanderWolf » Logged
Tim Lookingbill
Sr. Member
****
Offline Offline

Posts: 1230



WWW
« Reply #11 on: May 19, 2012, 10:40:10 AM »
ReplyReply

I really like the tonal controls of the new ACR.  The one troubling area is when I want to use to new controls on images that previously worked on with PV 2010.  In some cases, when I update to PV 2012 the tonal changes are dramatic.  I’ve found it easier to use the original image and start from scratch. Except that some images have had significant spot removal, which now needs to be redone. It would be desirable to be able to retain the effect of some tools when converting to PV 2012.   Ideally (for me) it would be desirable to be able to have all the basic sliders set as close as possible to retain the tones that were set in PV 2010 and then be able to use the new PV 2012 to make additional fine-tuning adjustments.
Is anyone else having this issue?  
Regards,
Paul

Copy the process of one and Paste to the other. When given the Settings dialog box listing all separate panel edits click off/on what you want. Spot Removal is selectable.

Do this on a copy of the image set to Default (no edits). When pasting the settings from the original only select Spot Removal.
« Last Edit: May 19, 2012, 10:45:07 AM by tlooknbill » Logged
Tim Lookingbill
Sr. Member
****
Offline Offline

Posts: 1230



WWW
« Reply #12 on: May 19, 2012, 10:41:35 AM »
ReplyReply

Hi,

Caused by noise?

Cheers,
Bart

I don't see noise in those shadow areas. Are you saying the physical electronic presence of noise is causing the recording of shadow detail to go flat?
Logged
Peter_DL
Sr. Member
****
Offline Offline

Posts: 421


« Reply #13 on: May 19, 2012, 10:45:39 AM »
ReplyReply

Post # 3 ff.
Charles' article and approach is better for visual editing of images in ACR/LR, but my intent was to quantify what he was seeing on the histogram.

Solid analysis, Bill,
many thanks for contributing it here!

If possible (with reasonably limited effort) I think it could be helpful for many of us to see the last 4 graphs (about PV 2012 Exposure, Highlights, Whites and Contrast)
in a non-log style, but gamma encoded, just like an applied Curve in a standard working space environment.

Regards,

Peter


--
« Last Edit: May 19, 2012, 11:01:44 AM by Peter_DL » Logged
BartvanderWolf
Sr. Member
****
Offline Offline

Posts: 3865


« Reply #14 on: May 19, 2012, 10:46:11 AM »
ReplyReply

I don't see noise in those shadow areas. Are you saying the physical electronic presence of noise is causing the recording of shadow detail to go flat?

Hi,

I added something to my post, gamma adjustment. When the gamma adjustment from linear is added, there will be a compression of shadow detail to the lesser significant bits, the noisy ones. For illustration one would also need to see the linear gamma histogram, where the separation would be better.

Cheers,
Bart
« Last Edit: May 19, 2012, 10:48:49 AM by BartvanderWolf » Logged
Tim Lookingbill
Sr. Member
****
Offline Offline

Posts: 1230



WWW
« Reply #15 on: May 19, 2012, 11:07:27 AM »
ReplyReply

Hi,

I added something to my post, gamma adjustment. When the gamma adjustment from linear is added, there will be a compression of shadow detail to the lesser significant bits, the noisy ones. For illustration one would also need to see the linear gamma, where the separation would be better.

Cheers,
Bart

That makes more sense, Bart. Thanks.

That also explains the behavior I mentioned previously of the Parametric Shadow slider adjust over just increasing the Black slider or trying to tweak the Point Curve. Wonder what new adjust in PV2012 offers a similar Clarity adjust only for shadows.

The Parametric Shadow adjust works great at separating and opening up tones below 20RGB +/- in ProPhotoRGB output space. But the effect varies image to image depending on quality of detail to noise influenced by varying scene light levels and exposures.

I'ld like to thank Jeff Schewe for the Parametric Shadow curve tip he suggested a while back. It took me a while to figure out this miracle tool because it wasn't working very well due to the fact that it often requires a hefty Fill adjust (along with its unwanted saturation boost: but fixed by Contrast reduction) and going back and adding Highlight tweak in the Parametric curve. And here I was all these years thinking the Point Curve was the miracle tool.
Logged
Peter_DL
Sr. Member
****
Offline Offline

Posts: 421


« Reply #16 on: May 19, 2012, 11:35:42 AM »
ReplyReply

In point of fact, Exposure was not renamed Whites...Whites is a new tone control specifically designed to fine tune the placement of the white clip point on top of Exposure. So, the first statement is wrong while the second is the correct description of the new Exposure.

Ah ok, I see,
many thanks for the pointer and clarification.

The more we find it hard to understand the philosophy behind the change and these particular new controls:

The former Exposure slider – which was a legit digital representation of camera exposure – got separated to two new controls, now with a higher and lower impact on the midtones (-> PV 2012 Exposure and Whites). Both moving the White point though, which also means less separation at this end, compared to former Exposure and Brightness.
In other words, why are PV 2012 Exposure & Whites considered to be superior to PV 2010 Exposure & Brightness ?

Regards,
Peter

--
« Last Edit: May 19, 2012, 12:10:49 PM by Peter_DL » Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2861



« Reply #17 on: May 19, 2012, 01:44:13 PM »
ReplyReply

If possible (with reasonably limited effort) I think it could be helpful for many of us to see the last 4 graphs (about PV 2012 Exposure, Highlights, Whites and Contrast) in a non-log style, but gamma encoded, just like an applied Curve in a standard working space environment.

Peter,

I think you are asking for something like this, which is a bit easier to interpret for photographers. Exposure is in f/stops (log scale) and the pixel value is linear. Middle gray is -2.5 EV for exposure and has a ProPhotoRGB value of 98. When I get time, I will revise the other graphs.



Regards,

Bill
Logged
mouse
Full Member
***
Offline Offline

Posts: 172


« Reply #18 on: May 19, 2012, 03:30:45 PM »
ReplyReply


Quote
With PV2010 and "linear" ACR settings one does get linear results and this enables one to produce a scene referred image using the method outlined by the ICC. It is really a nightmare to get such results in PV2012. Fortunately, Adobe was wise to retain PV2010 and Jeff Schewe indicated that future versions will also retain it.

Sorry to interject such a banal question into this most interesting and erudite thread.
Can I interpret this to mean that ACR7 will give one the option of working with PV2010 or PV2012?
Logged
bjanes
Sr. Member
****
Offline Offline

Posts: 2861



« Reply #19 on: May 19, 2012, 04:31:56 PM »
ReplyReply

Peter,

I think you are asking for something like this, which is a bit easier to interpret for photographers. Exposure is in f/stops (log scale) and the pixel value is linear. Middle gray is -2.5 EV for exposure and has a ProPhotoRGB value of 98. When I get time, I will revise the other graphs.

Replying to myself to post some more revised graphs in the format Peter requested. The contrast graph is illustrative of image adaptive editing. The contrast curves vary with exposure settings.







Bill
Logged
Pages: [1] 2 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad