Editing rationale

A quick before and after with text caption to explain some of the reasoning behind selective editing. Mostly done with layer masks and blend if sliders.

It’s sometimes useful to ask “what do I want to achieve with this picture?” before delving into post processing. So here’s a personal Q&A I ran through during various stages of the post processing.

  1. Clearing Clutter : I thought some of the leaves at the bottom blurred by motion and the top left branch were a little too cluttery as was the sharp foilage in the background. I will probably clone the third remaining leaf out from bottom right but for some reason, I don’t mind that one as much
  2. Depth through lighting : I wanted to enhance the natural brights while leaving shadowed areas darker to maintain a sense of depth rather than lighten the whole image
  3. Depth through detail : Even though everything is in focus in the original shot, I sometimes like to sharpen different elements of the scene to exaggerate depth. In this case the foreground rocks and the immediate ferns in the backdrop while softening the light from the background.
  4. Care with contrast adjustment. In ‘normal’ mode, any contrast adjustment not only boosts contrast of lighting but contrast of colour. The browns really lit up in the foreground and the main branch in the background with contrast adjustment necessitating a dial back.
  5. Overall saturation. The final image is oversaturated beyond what I would print for myself at home as is the blurring. The 16 bit base edited file at home is a version I keep for myself to present in different forums differently depending on what I want to show.

purakaunui editing

Old dog new(ish) tricks!

In this post , I thought I would share with you some adaptations in work flow that have occurred for us in the last few years.

Some of them are extremely basic and applicable to most images, some are just a matter of preference, some may even be controversial! (well in a 1st world problem kind of way)

Hope you might pick up on something anyway!

1) Editing format: Many people ask us why we bother editing in LR, exporting as a 16 bit TIFF before working in photoshop instead of editing directly in photoshop. There are two main reasons we still do this. First of all, I consider our workflow in three separate phases .

  1. Phase 1 is the lightroom phase. Lightroom has been a life saver for us in terms of creating an easily searchable catalogue of images. It is also useful for ‘quick corrections of  the basics like white balance , exposure and profile corrections . Within lightroom , an edit requiring  more than 1 minute with sliders, I will leave to the finer controlsof photoshop. After this phase, images are ranked appropriately for future editing .
  2. Phase 2 is fine tuning the image in photoshop .This step often occurs literally years after being imported in lightroom!
  3. Phase 3  is output preparation.

The problem with our photoshop phase is that it can take quite some time depending on the image and in my current every day life, I can bet reasonable money on being interrupted at some stage either by work phone calls or children with various needs hence any work in progress in photoshop will need to be saved as an intermediary file anyway. We are comfortable with image quality as a 16 bit TIFF. The second reason is for organisation predominantly born out of our previous wedding experiences. Our habit is to bulk export a group of files into a  ‘to edit’ folder and then coming back to them at some stage when we have a spare moment. If you have a similar interruption filled editing experience, perhaps doing something like this might help but otherwise, there is no compelling reason to just edit straight out of lightroom or ACR! Oh, and I know people won’t like reading this but we work in sRGB right throughout for various reasons which are entirely dependent on our main output and preferred printer.

An example of images from the Great Ocean Road catalogued in lightroom. Most have been exported into a set folder waiting for editing!

An example of images from the Great Ocean Road catalogued in lightroom. Most have been exported into a set folder waiting for editing!

2) Sharpening: Once again I consider that there are two ‘phases’ to sharpening an image.

  1. The first is ‘artistic’ sharpening which is what I apply to the working 16 bit TIFF file. This is not severe sharpening and the method of choice I use is high pass filter. The reason for using high pass filter is that it is a transparent layer which even when sitting at the top of your ‘layer stack’ allows other adjustment layers below (and above) to be seen. If unsharp mask or smart sharpen filter was used , other adjustment layers would need to be placed on top as it is an opaque layer.  A typical setting I would use for a 22MP image is a radius of 1.5 and blend mode ‘overlay’ with appropriate masking.
  2. The second phase is ‘output’ sharpening. Typically for web output on an image 1024 pixels long edge, I would use smart sharpen with radius 0.2, 150%, lens blur. For print output , this is complicated and depends on print size, print medium and image type. For instance, stars need incredible degrees of sharpening to appear anything like stars on print media (vs faint white dust bunnies). For canvas prints, I like to take advantage of the textures it may produce and hence I tend to ‘excessively’ sharpen areas which have the appropriate texture.

Other tips for sharpening in general include minimising the sharpening of noise and minimising colour artefacts. Noise in dark areas can be accentuated by whatever sharpening process you use hence I either use the blend if slider (dragging in from the left mainly) or luminosity masks to exlude dark areas from the sharpening layer. In order to minimise colour artefacts, you can either work in Lab mode (which I haven’t done) or make sure your smart sharpen layer is in luminosity blend mode.

For an image like this to look similar on print, stars need heavy sharpening but without accentuating noise in other parts of the image

For an image like this to look similar on print, stars need heavy sharpening but without accentuating noise in other parts of the image

3) Correcting weird lens flares: I’ve found that oddly shaped lens flares can be made to look even weirder with cloning! Here’s a method I’ve had more success with lately.

  • Duplicate your background layer and change the duplicate layer blend mode to ‘colour’ .
  • Select the brush tool and select the colour on the image immediately adjacent to the flare in an unaffected part of the image.
  • Brush away at the flare carefully with a low opacity until the colours are equalised.
  • You can repeat the same process for luminosity of the affected area as well but having the duplicate layer in ‘luminosity’ blend mode and making sure your brush has colour selection is on the far left (with no ‘a’ or ‘b’ colour value).\
Correcting lens flare previously made me forget about editing this image. The flare is also an in field way to get 'light bleed' (see below)

Correcting lens flare previously made me forget about editing this image. The flare is also an in field way to get ‘light bleed’ (see below)

4) Light bleed: There are definitely way too many advanced tutorials on this subject for me to approach this in any detail. One crude method is as follows.

  • Duplicate your background image and change the blend mode to screen.
  • Introduce a gaussian blur to this layer (eg. 22 pixels for a 22 MP image) and then reduce the opacity until the bright areas are as blurred as you desire.
  • Increase the contrast of that layer through a simple ‘brightness/contrast’ adjustment. Use a luminosity mask to select brights only and reduce the opacity of the layer until the ‘bleed’ is of the appropriate intensity.
  • Now you can modify how you would like that bleed to occur whether it be through further gaussian blurring of the mask, radial zoom blur or hand brushing in or out other parts of the image you wish to include or exclude.

 

Light bleed tends to work visually when there is actually a source of light in the image (as opposed to inventing a light source)

Light bleed tends to work visually when there is actually a source of light in the image (as opposed to inventing a light source)

I hope that gives some pointers to people at any level!

 

-D

Advanced photoshop tutorial : What if?

The following tutorial outlines how to create a layer mask from a ‘blend if ‘ effect. If enough interest exists, a video tutorial could be arranged!

 

What am I using this technique for?

  • Exposure blending
  • Selective adjustments in areas of different luminosity /brightness

What are current alternatives? (pre-requisite knowledge)

  • Luminosity masks : Google search ‘Tony Kuyper luminosity masks’
  • Blend if sliders: Google search ‘Mark Metternich blend if’
  • Use of layer masks.

How applicable is this technique to you?

  • Limited applicability /Situational use only.
  • Likely for advanced users of photoshop who have encountered limitations with luminosity masks and blend if in isolation

What is the problem?

Luminosity masks are generated from ‘overlapping’ areas within the channels palette hence brightness and darkness masks are relatively fixed.  Various midtone masks can be generated based on overlapping , intersections and exclusions of various bright and dark masks. These likewise are only customisable after the mask has been created. The benefit of using luminosity masks is that once a mask is created, it can be subjected to all other mask refinement techniques.

Blend if sliders are excellent in that their effect can be viewed in real time. The other advantage is that they can target extremely specific and customisable areas of brightness. For instance, if I wanted to target midtones with luminance values of 100-130 alone, this can be done  easily ; I do not have this level of customisation with luminosity masks. It is easy to remove the effect of a blend if easily with a layer mask though ‘reincluding’ areas excluded by the blend if effect requires a work around.  The other limitation of the blend-if I have found , is that once I have achieved the effect in a target layer that I desire, I have never had a saved layer mask for future use. For instance, if I wanted to retarget the exact same areas with other adjustment layers or blurring /sharpening techniques/mask refinement it was always difficult to do with the flexibility of using layer masks.

After some reading , I found a way to create a layer mask (which can then be saved in the channels palette for future use) from the effects of a blend if layer.

Here is the way I have found to achieve this :

  1. The layer upon which you are utilising the ‘blend if’ effect has to be one with pixel data (ie. not an adjustment layer)
  2. Adjust the blend if sliders as you normally would as if you were blending or achieving selective adjustment of highlights/midtones/shadows

    Use the blend if sliders as per normal

    Use the blend if sliders as per normal

  3. With this layer selected, press Ctrl-G to create a group
  4. With this group selected, press Ctrl-E to flatten this group into a transparency (turn off the background layer to see what you have created)

    step 2

    After flattening, you should transparent areas where the blend if effect has excluded the brightness values you chose

  5. You should now see a layer created with only parts of the image selected by your ‘blend if’ effect
  6. Ctrl-left click that layer to select its contents
  7. Switch to the channels tab and save the current selection as a new channel  and there you have it, the blend if effect in the form of a layer mask.

    This is the mask you will now have for future use with your 'exact' prespecified blend if values

    This is the mask you will now have for future use with your ‘exact’ prespecified blend if values

 

** Alternative pathway suggested by Erin Babnik  (which is better and preserves your original blend if sliders) **

  • Follow step 1 and 2
  • At step 3, convert the blend-if layer into a smart object
  • Follow from step 6 onward

Potentially, this could be a very powerful way to develop very specific luminosity masks (for instance masks could be generated for ANY brightness value with any increment) but for the time being, manual adjustment of luminosity masks is still getting the job done 99% of the time for me.  It is however good to know that when blend if is working better for me in a certain situation, I can utilise the benefit of having this in the form of a layer mask.

Hope that was helpful and remember, these techniques are not mutually exclusive and often, combining will help you achieve what you need.

 

Follow

Get every new post delivered to your Inbox.

Join 195 other followers