ETTR (underexposure) help sought

[tjv]

Hi all,
I just wanted to apologise for not replying to the thread - being the original poster - I've been hectic at work and home. Thanks all for contributing knowledge and info. I will post some of the things I've learnt when the can steal some time...

 

[torger]

I'm no expert on Ligthroom but as far as I know the automatic black level adjustment is controlled via the DNG camera profile, via the DefaultBlackRender tag. If it exists and is set to "None", no automatic black level subtraction is made. This is the default for profiles made with my own DCamProf camera profile maker which makes shadows lighter. AFAIK all bundled profiles in Lightroom lacks it and thus get this adjustment.

I didn't think there was a white level adjustment, but I can't say I have verified it either.

I do know that over-exposed shots with lots of clipping is pre-processed in a special way to look more similar to overexposed film (brightened, desaturated), as over-exposed digital locks pretty ugly. This makes it a bit messy to make a HDR merge of say TIFFs output from Lightroom, it's then better to merge at the raw level as Lumariver HDR does for example.

In any case you can't really trust the big name raw converters to "not do anything" when sliders aren't moved. So tools like RawDigger or a "scientific style" raw converter like RawTherapee can be a good complement when one wants to really see what's in the file.

Just a note for those that might go through this thread in relation to Erik’s comparison of C1 vs LR defaults, one thing to be aware of (which can be good or bad) when Lightroom introduced it’s 2012 process, it added some preliminary “intelligence” for lack of a better term which analyzed your image and adjusted certain parameters. Lightroom tries to preset a white and black point based on the image. Unfortunately that adjustment is not reflected in the black and white sliders, which will always display 0, even if Lightroom has pulled those points in.

Here’s a before/after screen shot of an image, the left is the raw file in the 2010 process, the right the only change has been to move to the 2012 process. Note almost all clipping has been eliminated. the 2010 process was more like C1, meaning the user would deal with those end points through adjustment. as I mentioned, this can be good or bad depending on how you look at it.

 

[ErikKaffehr]

Hi,

Cannot give a good answer. If a subject has a large luminance range it can be very difficult to translate into a good on screen image.

From my experience I would suggest:

• Check out your raw files with a tool like raw digger that shows the real histogram of the image. Raw processors always show manipulated histograms.
• Something I tried with good results is make a flat image containing all the tonal information -> open in Photoshop ->duplicate it -> use (Image) HDR-processing on one of the images.
• Lightroom has content aware tools for shadows and highlights that may, or may not, give better results.

A trick I sometimes use in Photoshop is to fix bright/boring sky.

• Create a new layer
• Set blending mode to multiply
• Use the blue channel for luminance masking alt-cmd-5 (I think)
• Apply that channel as maks

This method does a pixel exact masking of the sky and avoids artefacts.

I do my processing in Lightroom, but this is what I would do on your image and on many of mine:

Use a graduated filter on the top part:

• After that I would reduce 'blacks' until some deep shadows are clipped.
• Use 'Shadows' to get good detail in darker parts
• Use 'Highlights' slider to handle sky, to get tonal separation in bright areas
• If I use 'Highlights' I may also need "Whites" to extend the brightness range.

Some clarity and vibrance to fix the picture.


The image below may not be the best example, it just happens to be the one that I checked out last yesterday evening…


Most of this can be done in Capture One, too, I am pretty sure. I don't think Capture One has the same content aware tone mapping capability Lightroom has.

Topaz Labs has a lot of tools to achieve better looks on images. Worth to try. Personally, I prefer to do as much as possible with parametric processing in Lightroom and I can achieve the rendering I want.

Best regards
Erik

Hi guys,

First up, please let me state the following:

I have absolutely zero interest a debate about the relative merits of CMOS vs. the CCD back I'm using. I'm well aware of them and have read a lot about the new products coming through, even demoed a few for myself. What I am interested in is discussing how to get the most out of what I do have (Credo 60 on a Linhof Techno platform) in terms of preserving important hightlight information in high contrast scenes and getting better shadow information. Like for slide film, perhaps this is somewhat of an oxymoron...

Anyay, a bit of context.

I've been trying to shoot a cliff face including a bit of surrounding contextual information for a long-term project I'm working on. I'm trying to shoot a two shot, 6x13 panoramic stitch and am using an average lens in the Rodenstock 55mm APO-SD with a centre filter (2.5 stops... ouch...). Movements are quite extreme, so I'm suffering a lot with sharpness fall off. +/- 17mm horizontal shift, AND 10-13mm fall of the back. Always between f11-16.

Anyway, the shots I've taken so far have been frustratingly bad for many reasons, not least because I simply haven't managed yet to catch the scene under interesting light. I've tried and I've been there at times when it's beautiful in the flesh, but the foreground is always dark and if I'm to retain detail in the highlights, i.e. expose to the right with only very minimal clipping in the sky, the foreground falls very dark. Today looking at the histogram on the back when in the field I thought I'd managed to reach a good compromise exposure (although I knew the resulting shot wouldn't be very good from the get go,) with what seemed a tiny bit of space at the foot of the histogram and only a tiny bit of highlight clipping (viewing histogram on the back). I was looking forward to editing the photo and seeing what I could do with it, but on the computer my enthusiasm quickly dissapated when I lifted the shadows. Total mush, not helped by the falloff of the lens...

Anyway, attached here is a screenshot of an LCC file (when applying it I turn the vignetting correction off and just remove the colour cast to try keep the noise levels down,) the flat file preview (right frame, 17mm horizontal shift plus 10mm fall of the back), and a preview after some very rough editing. For this example I've used the Product curve in C1, along with the Pro Photo colour profile (the product profile, which usually looks good if quite low saturation, looked horribly dead.)

My question really is what would others do in this situation? Would you expose the image differently, i.e. force the highlights to white to improve the shadows, or do you have any special tricks in C1 to magically clean things up? I must admit that I'm not a C1 native and hugely prefer the Lightroom interface, so admit to getting frustrated quite quickly when trying to get the results I want. Would you resort to exposure blending and if so, what is the best way to go about doing this? I.e. is there good software that can do it in an automated sense like the photomerge function in PS, even with images to stitch as a pano?

Seriously, any advise appreciated.

TJV

 

[Steve Hendrix]

Halo,

With the 3 100, is it safe to only watch the 2nd purple clipping warning and ignore the red clipping?

Thanks,
Dan

Dan, it depends on what you want the red clipping to tell you, since it represents the user definable warning. What the purple is telling you is, Man, you really messed up this part of the scene beyond all redemption. The red is allowing you to set the limit of what you might consider the edge up to that point of un-redeemable, or per your subjective setting, possibly a more conservative warning. Totally up to you. The purple is saying - you're hosed in the area you see me. The red is saying what you subjectively want it to tell you. The optimal use is deciding what you want the red to tell you.


Steve Hendrix/CI