Match levels

[Eikon]

hey all... trying to figure this out and i cant quite get it.

so lets say i have my image, bring it thru ACR and its all nice and properly exposed to my liking. i make a copy of that, go back into ACR and bring the exposure down so its much darker... more shadow.

if i arrange these layers side by side i can see the normal version and the dark version.

by adding a levels (or curvers, or whatever) adjustment, i was thinking if i sampled the black, mid, and white point of the dark shadow image, i could get the same look on the normal image. basically recreating the exposure and applying it to the normal image to get it to look dark. this doesnt work, or im doing it wrong.

any idea how i could do about setting up some sort of adjustment layer that would mimic the exposure of the darker version of the image and make the normal exposure look the same?

thanks

[skoobey]

You might try to get some of that effect through combination of techniques.

Exposure uses an algorithm which is useful for keeping the dynamic range, yet introduces more noise that when applying curves or levels.

Try combo of curves and selective color. That should resemble the exposure of the initial conversion. Curves are way more precise than levels.

[Eikon]

exposure doesnt work. i find that adjustment pretty useless unless you have a 32bit image where there is actually enough dynamic range for it to be usefull. using it on 16bit just makes the image look grey.

i tried curves as well but also no success... though im not the best at curves in general.

im sure i could get there with a few adjustments stacked up, and doing it by eye till it was very close. the idea was seeing if there was a way that it was easily reproducable with a few clicks... hence the white, grey, and black sample points via layers (or curves).

[skoobey]

I rarely take images that are more than half aa stop over or under exposed, and even then I do a few separate conversions. Exposure does introduce an insane amount of nosie, so I preffer using curves than and there combined with blacks fill light and recovery.

[Tony W]

It is not exactly crystal clear to me what you are trying to achieve.

Have you considered using Smart Objects to alter ACR data on different layers?

In case you need it:
Open your file in PS via Camera Raw as a Smart Object.
After adjusting to get good image. Right click original Smart Object and select New Smart Object by copy.
You will now have two adjustable images on seperate layers that can be edited independently in ACR by double clicking the layer.
Fine tune the effect by changing opacity, blend modes and or masking may give you what you are looking for.

[John Wheeler]

Hi Eikon

I agree with Tony that it would help to know exactly what you would like to achieve. Did you want to replicate a specific exposure with curves in Photoshop? Is it curiosity. Smart Objects were already mentioned yet you could just doulbe click into the Smart Object and change the exposure from there.

Here is an example of a pretty good match. I took a random pic of a stream as an example.
-The one on the left is brought into Photoshop without exposure adjustment
- The one in the middle is just through ACR with a -1.0 exposure adjusment
- The one on the right is a curves adjustment layer with the curve as shown in the image. This gives a pretty good match.

Screen-shot-2012-01-15-at-2.45-SFW.jpg

First, this curve will not work the same if you bring in from ACR in different color spaces (e.g. ProPhoto vs sRGB) because the black levels don't come in exactly the same. Also, I am not convinced that the exposure level is just an adjustment layer equivalent. Finally, if you want a different exposure, changing the curves would not be real easy.

There is a process to determine a match very similar to what Tony mentioned.
- I take the original ACR image with a curves adjustment layer above it.
- Above that I put an ACR version with a -1 EV adjustment and set the blend mode to difference
- above this final layer I add one more adjustment layer to amplify the difference result (so it is not so dark)
- Go back to adjustment curves layer (2nd Layer) and adjust until you can get the darkest final image you can. The darker it is the better match you have.

Hope the above is useful.

[Eikon]

hmm, i thought the explanation was pretty clear. copy/mimic the exposure accurately from 1 layer/image and apply it, via levels/curves, to another layer/image.

the point here is not to recreate something by eye, cause if i wanted to do that id just do it. the point was to see if there was a way to recreate the high/mid/low tones/values of an image (my guess was thru the eyedrop sampler), load those values onto an adjustment layer so that that adjustment layers gives those same values to whatever layer is above or clipped to.

maybe im looking for something that doesnt exist and ill just continue to do it by eye.

[Tony W]

Originally Posted by Eikon

hmm, i thought the explanation was pretty clear. copy/mimic the exposure accurately from 1 layer/image and apply it, via levels/curves, to another layer/image.
...
maybe im looking for something that doesnt exist and ill just continue to do it by eye.

any idea how i could do about setting up some sort of adjustment layer that would mimic the exposure of the darker version of the image and make the normal exposure look the same?

It appeared to be quite unclear what you wanted due to the fact that you stated you are copying an image via ACR and on the second one you are changing exposure so that you have 2 different images one perhaps normal and your adjusted darker image. So you already have the darker copy that you seem to be striving for!

Then you want to just ignore the dark version and for some reason try and copy the effect using levels or curves on your original lighter version in PS. As you have found out you cannot do this accurately just selecting white, grey, and black points. ACR works quite differently than your method in PS.

The question still remains why on earth would you want to do this when you already have a superior image from your ACR conversion?

If you are talking about applying changes you make to one image to a series of different images then have you looked at the option of applying changes via Batch processing either LR or ACR or PS?

[Eikon]

if i didnt have a reason i wouldnt have asked. and thats all i wanted to know... since you said its not realy possible to do accurately, well then there we are! [sarcasm]case closed[/sarcasm].

edit: for those who cant read between the lines.

[Tony W]

Fair enough, perhaps next time you could explain your reason for needing to do something you have actually already accomplished by another method. This may illicit a better response with more options - now case closed?

[Eikon]

the reason was experimentation and trying new methods and workflows. case closed if you want... if anyone else can offer up a method, theyr free to do so.

[John Wheeler]

Originally Posted by Eikon

the reason was experimentation and trying new methods and workflows.

Bingo! That short statement said it all to me for what you are looking for. Not necessarily an exact bit for bit match yet something that is close to the same functionality yet still maintains some sort of slider control for adjustments as opposed to matching an exact EV change to be applied to a number of images.

I pulled out my black box reverse engineering tools and here is what I came up with to simulate the exposure slider in ACR from within Photoshop proper.

- Start with Original image
- Add Invert Adjustment Layer next layer up
- Add Levels Adjustment Layer next layer up
- Add another Invert Adjustment Layer next layer up

Use the gamma slider (middle slide) of Levels adjustment Layer. In this case to darken image similar to negative EV move the slider to the left. If you want to brighten image move the Levels gamma slider to the right.

For reference, a -1.0EV setting on the Exposure ACR setting is an almost identical match with the Levels Adjustment gamma setting of 1.98 (moved to the left). For a +1EV setting on the exposure ACR setting is very close to a Levels gamma setting between .65 and .7 (moved to the right). This brightness setting is not an exact match because the ACR exposure setting actually clips the highest brights where as it is a smooth transition with the Photoshop Levels gamma slider.

One may think that putting two Invert Adjustment Layers would not make a difference. That is not the case. The gamma slider has larger percentage changes in the lower part of the luminosity range than the higher part. Doing the inversion makes the gamma slider have a larger impact in the higher parts of luminosity just like the ACR exposure slider.

Let us know how this works for you (I too do some similar experimentation looking for alternate workflows)

[Tony W]

John an interesting solution to aproximate ACR control and one that I have to say that I would never have considered. I suspect you like to 'tinker' and take things apart somewhat - so do I .

I had a quick play with it and can see how it works (I think ) but cannot see at this time what benefits this may offer over just using ACR. Can you see a scenario where this would offer any benefit or was this merely an excercise?

[Eikon]

john, thats great... arent we glad we didnt throw in the towel right away! while its not perfect, i think its a good approach and gives a pretty close match. on the picture i used, reds were still a bit more saturated with this method as opposed to the exposure in ACR, but nothing that couldnt be fixed with another color adjustment. it works well in reverse too, making the exposure brighter.

ill play around with this method and see if it gets me where i want to be. i appreciate the insight!

oh... care to offer any wisdom and maths on why this works different than straight up levels?

[John Wheeler]

Originally Posted by Tony W

Can you see a scenario where this would offer any benefit or was this merely an excercise?

The exercise was to provide an answer for Eikon. I can imagine some uses yet the ACR exposure tool is best especially for Camera Raw in pulling out details at the high or low end. I asked Eikon what the need was for and he said sure I can tell you but I will have to kill you if I do (so I declined). So Tony, I bet Eikon would tell you if your really really want to know

I do a lot of black box under the hood evaluation of PS to both find out new workflows yet also to find out when PS tools are not good to use (if that is tinkering then I guess I do).

Originally Posted by Eikon

john, thats great...oh... care to offer any wisdom and maths on why this works different than straight up levels?

Your welcome Eikon. I came up with this not by equations yet by taking a look at the transfer curve for the ACR exposure tool. Here is link to an image of the regular gamma curve in Levels in Photoshop:
http://upload.wikimedia.org/wikipedia/commons/7/74/GammaFunctionGraph.svg

The ACR exposure transfer curve is apparently very close to a curve that would closely match if you flipped the X axis and the Y axis both between 0 and 1. Since doing an inversion in Photoshop is the same as Output = 1-Input you just invert the Input before going through the gamma equation (Input ^ gamma) and flipping the Y axis is just inverting the result of the Levels gamma output. The result of the Y axis flipping also inverts the direction of the control on the Levels gamma slider (what was making dark now makes light). Hard to explain in text yet I did this all visually without equations (though that could be done too). Hope that gives you a feel for the approach I took. Graphs just compute in my head - what can I say

[Eikon]

lol, you lost me at "the". as long as it works!

[Tony W]

Originally Posted by John Wheeler

The exercise was to provide an answer for Eikon. I can imagine some uses yet the ACR exposure tool is best especially for Camera Raw in pulling out details at the high or low end. I asked Eikon what the need was for and he said sure I can tell you but I will have to kill you if I do (so I declined). So Tony, I bet Eikon would tell you if your really really want to know

I do a lot of black box under the hood evaluation of PS to both find out new workflows yet also to find out when PS tools are not good to use (if that is tinkering then I guess I do)....

Thanks John as I suspected and tried to point out in earlier posts there is little point in trying to re invent the wheel, without at least good reason. There seems little point in asking any further questions due to the fact that previous questions have been totally ignored. This of course is the OP's perogative but comes across at best as bad manners.
Yes what you do is tinkering in my book (using this definition One who enjoys experimenting with and repairing machine parts) - and long may you tinker well

[Eikon]

tony, which of your questions did i not answer that you want cleared up.

[John Wheeler]

Hi Tony
Base on Eikon's inputs I just saw him as a fellow tinkerer and that's what triggered me to offer another type of solution. Sometimes the intermediate steps are just part of the journey

Originally Posted by Eikon

the reason was experimentation and trying new methods and workflows....

[tom60634]

Load 2 files into Photoshop.

Then go to Image>Adjustments>MatchColor.

Choose 1 file as destination the other as source.

You then have a minimum of 4 adjustment levels to fine tune your result.

This is useful when,for instance, you want to match the colors found in a photo of an athletic match to the colors found in an outdoor recital, or if you want to match the colors taken with different cameras of the same event.

[Eikon]

thanks tom, but match color is "destructive". i was looking for a way to do this with adjustment layers.

[Tony W]

Originally Posted by Eikon

tony, which of your questions did i not answer that you want cleared up.

It is not just questions being unanswered (and not just my own!) it is the fact that in this thread you appear to be ignoring help and advice by not having the courtesy of posting a reply.

As your objective was not clear to me at least I tentatively suggested, have you considered Smart Objects. Guessing that you may want to combine exposures with masking hence making a copy using New Smart Object by Copy. Which would give you the opportunity to adjust the second copy independently.

John also posted an answer and went to some trouble to provide examples and plenty of detail.

Both posts ignored!

While I can understand that the answers may not be what you are looking for it is common courtesy to at least acknowledge the posters contribution and perhaps consider thanking them for their efforts, with explanation why they may not be appropriate.

Forum members here always try to be helpful and offer opinion, help and where possible advice. It should be automatic to acknowledge and thank them for their efforts (however feeble you may consider them to be) and IMO is good netiquette.

It is sometimes difficult to estimate the level of knowledge and experience someone has from their posts. In your case you stated in Post #3 ‘i tried curves as well but also no success... though im not the best at curves in general’. The inference being that perhaps you are not an experienced user. The statement further amplified by what seemed to be reinvention of the wheel which is ACR. Do not misinterpret this I doubt there is anyone here no matter what their level will claim they know it all and there is nothing new to be learned, even from PS beginners

For completeness and to directly address your query the questions I referred to include the following:

• Have you considered using Smart Objects to alter ACR data on different layers?

• The question still remains why on earth would you want to do this when you already have a superior image from your ACR conversion? Not answered until Post #11 after further probing your Post #9 and trying to ignore what appeared to be your curt answer ‘if i didnt have a reason i wouldnt have asked’

Johns questions

• Did you want to replicate a specific exposure with curves in Photoshop?
• Is it curiosity

It has just occured to me that it is possible that there is a misunderstanding of meaning and intent. Maybe like other forum members your first language is not English. If this is the case then I can appreciate that the way things are put can lead to an incorrect interpretation of the message and the intent of the poster.

[Eikon]

well tony, sorry you feel im not hand holding enough here. my 2nd response reiterated what i was trying to accomplish, so that kinda means that above responses were not what i was looking for. and unfortunately i dont have the free time every day to respond to every post on every forum im a member of.

johns 2nd response was great and exactly what i was looking, after explaining i was experimenting with different methods, and as he put it ' not looking for a bit for bit copy', we arrived at a solution. if no one was interested in reinventing the wheel, as you put it, we'd all still be using microsoft paint. i came here to learn something and i did... and i think you did as well. isnt that the point of this forum? i dont think any of us would get very far if we all told each other stuff isnt possible just cause its something we dont know.

my intent was simply to get some information. the fact that i was asking should be proof enough i have a reason to inquire. and its not my fault you didnt understand what i was trying to achieve based on my first posting. i always make an effort to make my original questions as clear as possible, not the standard "which button do i press to get the dave hill effect" question that pops up every other day. rereading it, it seems pretty clear to me. even if you read it and thought 'why the hell would someone wanna do that', thats not your concern, and shouldnt lead you to believe i somehow didnt know how to ask what i wanted to know.

and anyway, wasnt this case closed for you on post 10?

i do appreciate the fact that you went digging through all my other posts here to try and determine how much of a noob i am. i guess you have more time than i do. but i didnt join this forum to prove whos the better photochopper.

the bottom line is you didnt have a solution and john did. so thanks again john, and thanks tony for you initial response even though it wasnt the answer i needed. but now we've learned it was indeed possible, despite your belief that it couldnt be done... or didnt need to be done.

and if my english aint not to great for ya... so sowwy.

[Tony W]

Sadly the inane and immature comments carry on ,I only half expected this, but at least I tried to give you a chance to explain your rudeness and inability to answer questions posed. FYI it was you that said case closed in #9 mine was in response - see it is possible for anyone to descend to the level you appear to work in.

There seems to be no point that can be agreed upon therefore as far as I am concerned case closed!

[Eikon]

its about time.

[Der_W]

Just in case someone is still interested in a little black boxing, I've set up a little PSD file that helped me a lot in black boxing curves for blending modes: http://dl.dropbox.com/u/7124285/Fore.../Blendmodi.psd
Basically what you need to do is first set up the "Posterize" layers in the "Stripes" group. The higher the values, the more accurate your final curves will be.
Then create a new contrast change (if you want to change contrast in ACR, use Russell Brown's Script).
Now you can use your info palette to see, how the numbers changed.
For the 9x9 grid the values are written down at the borders, so you just have to see where the values were and where they are now.
Now inside the curves adj. layer (which really should be set to 0% so you don't get any changes from it and still can see the values of all the layers), set the input to the old value and output to the new value.
Copy this curve over any image you like with 100% opacity and it should be pretty much the same effect as using the adjustment you previously used.
I also show the process in a video with blending modes here: https://www.youtube.com/watch?v=_uu0XIbNAxY

Might make the black boxing a little easier :-).

[Eikon]

Cool video jonas. Thanks for the link