the complete guide from

camera to process

exposure blending

christopher odonnell

what is exposure blending?

how exposure blending works

5 when & why to auto-bracket

auto-bracketing overview

what is raw?

raw adjustments vs. auto-bracketing

the versatility of bracketing

10 how to auto-bracket

bracketing: step-by-step

manual bracketing

finaltipsonbracketing

14 editing in photoshop

evaluating the brackets

preparing to blend

theblendingworkflow

brushwork and halos

26 advanced blending

whentouserefinedmasks

ghosting & sensor bloom

creating the selection

refiningthemask

the overlay brush

sensor bloom & exposure blending

finaladjustmentstotheblend

48 introduction to luminosity masks

creating an alpha channel

highlight vs. shadow selections

refiningluminosityselections

creating the luminosity mask

transitions with luminosity masks

finishingtouches

finalthoughts

71 about christopher

72 other ebooks

73 copyright information

what is exposure blending?Exposure blending is one of the most powerful ways to create a stunning

photograph. As youve probably noticed, it can be difficult to capture an entire

landscape accurately (especially those with a sky) because your camera sensor

is limited to one aperture and one shutter speed at a time its not possible to

have two different settings within the same frame. Youll often note that your

photos rarely live up to what you actually saw exposure blending is one way

to overcome the limitations of your camera and photograph a landscape with

the full tonal range that you see in the field.

Simply put, in order to have a well-exposed landscape youll need one image

exposed for the brightest tones in your photo (typically the sky) and another

for the darkest tones (the ground, usually) when the lighting isnt suitable

to capture it all in one exposure think of sunsets and how differently the

natural light level is when you compare the sky to the ground. Once you have

your two extremes, youll blend them together to make one perfectly exposed

landscape.

Exposure blending can also be referred to as HDR blending in the sense that

were expanding the tonal range, and this method will create a more natural

looking image than tone-mapping with automated software. With exposure/

HDR blending, you have much more control over the end result by manually

choosing exactly where you want to blend your exposures together - and

at what strength you want the blend to be. Its a method that is entirely

customizable to your scene, which will produce a more pleasing result.

In addition to these two photos on the extreme ends, a middle exposure (one thats been light metered for the

entire image, finding a middle ground between your fastest and longest exposure) is good to have as well for

any middle ground elements you need to blend in. At times, the gap between your fast and slow exposure may

be so great, that you end up skipping the optimal exposure for those elements that fall in between the two. An

example of this would be the water in the below images. Notice how the image exposed for the ground (left)

has the water slightly overexposed, and the image meters for the sky (right) has some noticeable shadows. A

middle exposure was needed to balance this out. For intricate images with vast tonal needs - a scene with many

shadows and highlights - its not uncommon to have 5 or more photos of the same scene at your disposal.

The range of your exposures will depend entirely on how bright your highlights are and how dark the shadows

appear. For example, a cloudy day will have very little difference in exposure, but a golden hour image may have

5 or more stops to your complete tonal range.

The two different exposures to the left show how vast the

differences can be in order to capture the full tonal range of the

scene. As you can see, the setting sun caused some problems

in exposure. The image where the foreground is exposed well

(slower shutter speed) has an extremely overexposed sky, while

the photo exposed for the sky (faster shutter speed) has the

foreground in complete shadow. To get a proper exposure, I

needed to combine these two images (in addition to my base

exposure for the water) to get the result shown to the right. I

dont want you to be concerned with the how part just yet

- we will get to that later on. For now, the take-away point is

that to get a full tonal range, youll need to combine different

shutter speeds into one image.

how exposure blending works

1. when & why to auto-bracket

To get three different exposures of the same scene, you dont have to calculate your

settings and manually input a different shutter speed. Auto-bracketing is a common

feature with digital SLRs, which allows you to capture one photo exposed normally

(your base image), and then will automatically expose two more one directly after

the other - of a slower and faster shutter speed than your base. The difference in

exposure will depend on your camera and chosen settings, but usually 2 full stops

is the maximum increase/decrease from your base photo that you can use auto-

bracketing for.

This is where the term auto-bracketing comes from since youre bracketing your

images off automatically by capturing the same scene at different exposures within

X amount of stops.

This allows you to capture three sequential images of different exposures without

having to adjust your settings manually its very handy, and helps to streamline

your workflow when youre presented with a wide tonal range that one exposure

can not contain.

For more intense differences in lighting - as youll often find for sunrises and

sunsets - youll still have to adjust your settings manually to capture more of a

range in exposure than two stops (or whatever your cameras auto-bracketing limit

is). Ill explain how to do this soon, but first I want to show you why you should

auto-bracket instead of adjusting just one RAW file.

auto-bracketing overview

For brevity, a RAW file is the digital equivalent of a film

negative before it is printed onto paper - the raw state

of your image that you cant really use as a photo, but

all the information you need to create a photo is there. A

RAW file is just that - you cant upload a RAW file to your

website or email it to others without special viewing

software, but you can take that RAW file and turn it into

a universal image.

RAW is preferable to JPG in many ways, most notably

because of image quality. When you take a photo in JPG

format, it is instantly compressed - and this compression

will take detail away from your photo, thus limiting your

processing capabilities.

With a film negative, you can adjust the exposure by

exposing the negative onto your paper for longer or

shorter amounts of time. RAW allows you to do the same

- at least, for a limited amount of stops. Since youre

working with the unprocessed digital negative, you

can recover some blown highlights or blocked shadows

by adjusting the exposure of your RAW file before

converting it into a more universal file format for display

or print, such as JPG or TIFF.

You can artificially adjust the exposure of a JPG image in

post process, but this is not a true exposure adjustment

and will not recover any data for you. In the example

images, we can see the true benefit of RAW when trying

to recover these blown highlights. In Figure 1, you can

see that while the foreground is exposed accurately, the

sky is noticeably overexposed - a common occurrence.

Figure 2 shows how much detail is recovered by

reducing the exposure by two full stops in RAW editing.

In contrast, Figure 3 displays the same reduction in

exposure (two stops) for the JPG version. Since the JPG

file holds no additional data, the exposure is artificially

reduced simply by adding a layer of black tones - there

is no recovery of data, and the photo looks noticeably

worse than the RAW adjustment.

RAW does have its limitations however - it does not

give you the power to shoot blindly with no regard to

exposure, and it is always best to achieve a proper image

in-camera. A RAW file will typically give you two stops

of recovery capability - if you adjust the exposure more

than two stops beyond the original exposure, you most

likely will not recover any additional data and will start

to notice the quality to deteriorate.

what is raw?

With RAW being such a versatile format for editing,

why is there a need to auto-bracket when you can

simply adjust the exposure in RAW and import all

those exposures into one file? It would save you the

time of setting up a tripod, taking three (or more)

images, as well as saving on disk space I thought

RAW was supposed to be a magical exposure

recovery tool?

Its always best to capture the optimal exposure for

a scene in-camera simply because youll have more

data in an actual photo than an image adjusted

artificially even if its in RAW. The more data that

is retained, the more detailed your image will be.

Auto-bracketing is superior to RAW for the following

reasons:

1. You may need more range in your exposure than what RAW can offer you. Typically, you can only

adjust your exposure +/- two full stops in RAW

before you start to see damage depending on what

youre photographing, you may need 3, 5 or even 10

stops difference between your brightest and darkest

photos.

2. While RAW is a great, low-damage, accurate way to adjust your exposure, its still better to capture

the image with your camera to make sure you have

the highest quality photo.

Im going to compare a scene where I underexposed

my photo by two full stops using auto-bracketing

with my camera vs. dropping the exposure in RAW.

Taking the overexposed version that was adjusted

in RAW from the previous page and zooming in on

the recovered sky (Figure 1), you can see that there

is still some overexposed areas. When compared

to Figure 2 - the photo that I took two stops lower

with my auto-bracketing - you can see that more

detail was retained in the sky that RAW could not

recover.

So while RAW provides a great recovery tool, its

more of a last effort way to adjust your exposure

when you have no other alternative, or if you

only need a very minor exposure adjustment.

Adjusting your exposure in-camera will always give

you a superior quality when compared to RAW

adjustments, and is worth the extra effort of auto-

bracketing.

raw adjustments vs. auto-bracketing

the versatility of bracketingAlthough RAW can not replace auto-

bracketing, its still a file format that

has many benefits over JPG and should

be used when you bracket. Not only

is there more data in RAW format, but

youll greatly increase your tonal range

for editing. If you auto-bracket in JPG

format, youll typically be taking three

images your base image, one taken

two stops faster, and another two stops

slower - for a total range of four stops.

However, if you shoot in RAW format,

you can increase or decrease your

exposures by two stops in post process.

That safety net will add another two

full stops to your range on both ends of

your bracket, bringing your total tonal

range up to eight stops (assuming you

shoot three auto-bracket images in two-

stop increments).

Blending exposures using one RAW file

is best for those candid shots where

you didnt have enough time to set up a

tripod but need to correct some blown

highlights, or simply want to brighten/

darken parts of your image like in

wedding or action photography. Its

a fantastic way to save an otherwise

unusable photo, but if you have the

time and ability to auto-bracket your

landscape for exposure blending, this

is the method which will give you the

highest quality output with the largest

tonal range.

Exposure blending and bracketing is not

just for combining bright skies with

dark grounds you can use it in many

different ways. You can recover blown

highlights in water reflections, use it

as a no-damage way to dodge and burn

your landscapes, or to blend a brighter

foreground with a darker sky (like

under stormy, overcast skies).

Another popular use of this technique

is to blend different long exposures

together - for example, blending an

extended exposure that captures

moving clouds with an image that

shows stationary grass that would

otherwise be windswept if taken at

the same shutter speed as the sky: the

possibilities are endless.

Now that you know why we auto-bracket for exposure blending, its time to go over the in-camera workflow; the process I go through to capture those bracketed images to later blend in post process.

2. how to auto-bracket

The most important part to digital exposure blending is to auto-bracket the correct

way, giving you a solid base to edit your image in post process. You cant create a

stunning photo in the digital dark room you can only improve upon an already-

proper image.

With that in mind, Im going to show you how to execute an auto-bracket the correct

way so that you can capture images with an outstanding tonal range, like this example.

Ill also explain how to bracket your images manually for those instances where auto-

bracketing is too limiting for your environment.

Its very important that you have a sturdy tripod

one thats capable of withstanding the weight of your

camera, lens, and your tripod head. Solid legs with a

good grip are imperative to not only hold your camera

in place, but to make sure it stays there when you

change your settings, adjust your focus, or if a strong

wind comes along. Even the slightest movement can

alter your alignment, making it more difficult to

exposure blend in post process.

Place your camera on your tripod and adjust your

focus properly. Please read my focusing workflow as

this is a very important step to ensure that you have

consistent images.

Switch into aperture priority mode so that your

camera will only change your shutter speed to alter

the exposure and NOT your aperture otherwise youll

have inconsistent depths of field, making it nearly

impossible to exposure blend accurately.

Follow the instructions in your camera manual to

turn on your auto-bracketing. This is usually a simple

process with a few clicks and adjustments.

Press your shutter button and take three consecutive

images. I prefer to use a remote cable release so that

Im not actually touching my camera, which can

lead to camera shake. Out of habit, I also lock my

mirror which can also cause unwanted vibrations.

When finished, review the histogram to make sure

that you have the proper tonal range to work with

meaning that it indicates you captured one image

with no blown highlights and another with no blocked

shadows.

To the right, youll see the results of auto-bracketing

for one particular beach scene. The top image is my

base image one that has been metered on average

for the entire frame (evaluative).

The middle photo is my longest exposure horrible for

the sky, but it eliminated the blocked shadows from

my foreground rocks (as seen in the top image) nicely.

Finally, my least exposed image the blown highlights

in the first two photos have been eliminated (except,

of course, for the center of the sun) and has given a

nice starburst effect to the sun as well.

bracketing: step-by-step

At times, you may find that your cameras auto-bracketing

system is too limiting for the tonal range youre working with

usually only +/- two full stops. If youre shooting a high-

contrast scene with many highlights and dark shadows (such

as a sunset or sunrise), four stops (two in each direction) will

most likely not give you the full tonal range where all your

shadows and highlights are recovered.

When this happens, youll have to forgo the auto-bracketing

and do it manually, which is quite simple to do.

Auto-bracketing is a great benefit because it allows you to

capture three exposures without having to adjust your settings.

However, its only a time-saving feature auto-bracketing does

nothing more than what you can do manually.

Repeat the steps above regarding how to set up your tripod

and your focus. Instead of turning on your auto-bracketing, just

meter your image (in aperture priority mode) and make a note

of your settings. The aperture you choose to use should always

remain the same throughout your images instead, youll be

adjusting your shutter speed for your different exposures.

Switch into full manual mode (your instructions will tell you

how to do this) and set your aperture and shutter speed to

what it was when you metered. Take one image, then adjust

your shutter speed two stops faster (underexposed) to capture

another image.

Check your histogram for any blown highlights if you still

see some overexposed areas, keep increasing your shutter speed

until they disappear. Repeat this step for your shadows as well

continue to decrease your shutter speed (overexposed) and

check your histogram until you see that your unrecoverable

shadows are fully exposed.

During this process, never change your aperture or ISO to

ensure that you have smooth exposure blending.

manual bracketing

When capturing your photos, use a remote shutter release so that youre not actually

touching the camera. This will help minimize any blurring or camera movement.

Additionally, take special care when youre adjusting your settings as you may

accidentally re-adjust your tripod/camera in between photos.

Always shoot in RAW format, preferably RAW + a low-res JPG for quick preview

purposes. This will give you the most versatility in editing, and also an all-around

higher quality image.

Always set your ISO to 100. Youre using a tripod (no risk for camera shake) so theres

no need to sacrifice image quality for a faster shutter speed, unless theres a reason

for it (like a moving object you want to capture). The clarity and quality of your

image may suffer if you increase your ISO too much as explained here.

finaltipsonbracketing

3. editing in photoshop

Now I get to see the results of my careful preparation and

in-camera execution - blending the exposures together in

Adobe Photoshop software to make one complete image.

This can be achieved in any program that uses layers. For

this guide however, I use screenshots from Photoshop.

Once youve captured three or more RAW images of the

same scene, open all three files in the Photoshop RAW

editor so you can see them at once. For the example here,

I have stacked three brackets - the top image is my middle

(base) exposure, the next image is my overexposed bracket

(exposed for the ground), and finally my last image is my

underexposed bracket (exposed to retain sky detail).

In the top screenshot, Ive clicked on my base exposure

which shows you what is captured using evaluative

metering - and the reason why I exposure blend. The

foreground is in dark shadow, the sky is washed out, and its

an overall uninteresting and dull photograph that does not

represent the stunning scene my eyes witnessed.

By turning on the shadow clipping indicator (bottom image),

you can see that the areas marked in blue are underexposed-

in other words, there is no data to work with as-is so I need

to recover it through blending different exposures.

evaluating the brackets

The same method can be used for the highlights. If I

tick the highlight clipping indicator at the top right

corner of my histogram, the overexposed areas are

now shown in red.

This is a typical result when shooting scenes with

a high tonal range - a base image with blown

highlights and boxed shadows that have gone

beyond what your camera can capture in a single

frame.

With RAW editing, you can adjust the exposure

slider to recover these highlights and shadows, but

as you previously learned this is not ideal in terms

of quality - youll have more data to work with if

you use an image that had its exposure adjusted

in-camera than compared to the same adjustment

made in RAW.

Lets say you dont have a bracket to work with,

and just have one RAW file to adjust (like the image

above). You would still need to go through the

exposure blending workflow since your base image

has both over and underexposed pixels. If I tried to

expose this image more to recover the shadows, Id

overexpose the sky further, and vice versa the only

way to recover the full tonal range of this image is

to blend different exposures together.

Before I go any further with my RAW adjustments, I want to ensure that Ive

captured the full tonal range of my scene. By evaluating the outside brackets

- my fastest and slowest exposures - I can determine if I need to perform any

exposure adjustments in RAW to recover more detail.

With the clipping indicators still turned on, Ill first look at my longest exposure

(above), which is exposed to recover the shadows (blue). Except for a few rogue

pixels - which I can easily clone out later in Photoshop - my shadows have been

exposed enough to my liking. Im ignoring the blown out sky because that is

not the focus of this particular bracket.

I repeat the same process for my shortest exposure (fastest shutter speed),

which is exposed to recover the sky detail. There are no blown highlights in

the sky - indicated by red - so I have a solid tonal range to work with between

these two brackets. When you compare this image to the base exposure on

the previous page, its easy to see why exposure blending can greatly enhance

your photography. Much interest has been recovered, and you can begin to

visualize how your final image will look once these three exposures are blended

together.

I should note that these brackets are the result of a panostitch, and stitching

rarely produces an image with perfectly straight edges. The blue chunks you

see along the borders of the image are not part of the original scene - rather

just black areas filling the void where the edges arent flush. Ill end up warping

and cropping these sections out in process (shown later). My main concern

here are the under and overexposed areas within the scene itself, and I can tell

that Ive recovered all the data I can and am ready to proceed.

If I did have large areas of lost data, I would adjust the affected bracket in RAW

by sliding my exposure slider enough to bring back the details. Ideally, I would

have done a manual bracket in the field as described earlier, but at times you

still need to fine-tune your exposure in RAW.

From here, you should perform any RAW editing you

usually do, such as black and white conversion, white

balance adjustments, and so on. Click the Select All

icon to have your adjustments apply to all your images

except for any adjustments you make to your exposure.

This helps to cut down on your editing time and keeps all

your images uniform.

With all three images still selected, click Open to bring

them into Photoshop. Youll notice that your photos are

open in three separate windows - were going to combine

them into the same file for blending, with each image on

its own layer.

Select your base exposure file and click CTRL + A so that

you see the marching ants along the edge of your

frame. Press CTRL + C to copy the image and switch over

to your brightest photo (longest exposure). Create a new

layer and press CTRL + V to paste the middle exposure on

top. If you executed your bracketing properly and used a

tripod, you should have no alignment problems.

Repeat this step for your underexposed image (fastest

shutter speed) so that you have all three photos in one

file and on their own separate layers. Your top layer

should be your darkest image and the bottom layer

should be the brightest.

preparing to blend

theblendingworkflowFor the majority of this tutorial, Im going to be adding layer masks to create

one image using multiple exposures its much like HDR, only with a more

realistic appearance and completely customizable.

First, Ill add a layer mask to blend in my sky. Click the Add Layer Mask icon

located at the bottom of your layers palette.

Now select your brush tool by pressing B, and ensure that your foreground

color is set to black, and background set to white. With your new layer mask

selected, adjust your brush to a suitable size for this image, I want to combine

the water and ground exposure of the middle layer with the sky of my top layer.

Make sure your brush is set to 100% opacity, and with a 0% hardness level.

What Im basically doing here is creating an imaginary and reversible hole in

the top layer so that I can see the middle layer. Instead of applying it directly to

the layer, Im using a mask so that I can adjust or delete this hole later on.

Since I only want the ground and water to poke through, I just brush over

that particular area (i.e. I dont brush over the sky). While the base layer has

a brighter foreground than my top layer, its not bright enough to capture a

solid tonal range. My goal is to blend in the bottom layer for the foreground,

but I need to cut a hole in both my top and

middle layer in order to show the bottom

layer - remember, layers are stacked, so

your can not see anything beneath a non-

transparent layer unless its masked.

Also note that you can always switch your

brush color to white if you want to reverse

the brush work you previously did and

bring back the layer you are masking.

In the side-by-side comparison to the left, you can see how the base exposure

is now visible on the top layer. Im starting to see an overall improvement in

the tonal range, and how differently the image looks when blended with a

longer exposure.

By pressing the ALT key and clicking on the layer mask in your layers palette,

you can view the mask on your layer and see exactly where you brushed.

When looking at the bottom image, you can see where the brush strokes were

applied to create that hole, allowing the layer beneath to be visible. Since I

used a 0% hardness, the edges of the painted areas are soft and transitional -

depending on your image, this can be beneficial or counterproductive, so you

should adjust your hardness to accommodate your scene.

The brush opacity and hardness allows you to fully customize your layer mask

to fit your content. A brush with 100% hardness will give you a very sharp

(or hard) transition good for small areas or sharp horizons, but generally

doesnt look realistic when youre trying to blend tones gradually. I usually

start with 0% hardness for larger areas and readjust as necessary.

Another great tool you can adjust is your brush opacity or transparency.

When set at 100%, youre basically cutting a hole in your layer, allowing the

bottom layer to be seen completely. When you reduce the opacity lets say

to 50% youre only allowing 50% of your bottom layer to be seen and 50%

of your top layer to remain. This allows you to blend just the right amount

between exposures for the perfect balance of tones.

You can always reverse the masking you perform by inverting your

foreground brush color. Press X to switch your foreground and background

color - brush white to lessen your layer transparency, and black to increase it.

My seascape is starting to improve, but the foreground is still far

too underexposed for my liking. By selecting the base exposure, I add

another layer mask and brush the foreground rocks and tree line with

a 100% opacity brush set at 0% hardness (right).

In the end, I have three exposures in one to create a well-balanced

scene - an underexposure for the sky, the base exposure for the middle

ground water, and the overexposure for the foreground rocks and

shoreline. Note the changes in the histogram each time I blend a layer.

Often youll find that using 100% opacity may cause too sharp of a

transition on some areas and look rather unnatural. When I zoom in

on the shoreline (bottom right), you can see that the water and rocks

along the edge are a bit too dark and appear distracting.

If I use a 100% opacity brush, I have to choose between either the

base exposure (too dark) and the overexposed bracket I used for the

foreground (too bright). Rather than choose between the two, I can

adjust my brush opacity level to a lower percentage. This will allow me

to blend these two layers together gradually, displaying a bit of each

and create a more harmonious tonal range with a smoother transition

between layers.

To enhance my image and make the transition smoother,

I will mask the waters edge at a lower opacity level - in

other words, I will paint over my mask with a black brush,

but not at full strength.

In Photoshop, you can highlight where you have masked

by toggling the \ key, which will illuminate your brush

strokes in red. The top two images show that the transition

between layers is hard and abrupt (top right), which

explains the unnatural look to my photo (top left).

With a brush set at 30% opacity and set to black, I select

my overexposure mask and begin to paint the areas I think

are still too underexposed. If you look at the layer mask by

pressing ALT and clicking on it in the layers palette, youll

notice that this lower opacity brush will be varying shades

of grey, which indicates that your brush is not at 100%

strength.

By brushing over the area several times with a low-opacity

brush, I can layer my masking - gradually dissolving my

base layer so that the longer exposure underneath starts to

show for more accurate blending.

When you compare the left-side images, you can see how

natural and harmonious the layer blending appears now -

this change of appearance is mirrored when you compare

the two masks on the right-hand side. The bottom mask is

now gradual and soft, slowly fading away as you approach

the horizon line.

Using a low-opacity brush is not only useful for working

along edges, but for blending as a whole. If my foreground

had been exposed too much to my liking, I could have

masked it with a lower opacity brush - say 80% - and

retain a small bit of darkness from my base layer. You can

apply this method to any part of your image, and between

any pair of layers, for a truly customized blend.

brushwork and halosWhen blending your layers with the brush tool, it can be a tedious job to work around

structures and other prominent subjects that protrude between layers - such as trees

and buildings. Try to avoid the temptation to rush through your work. After all, one of

the top benefits of exposure blending over HDR software and using a GND filter is the

ability to selectively blend exposures together.

Now that I have generally blended our three exposures, its time to address the smaller

areas - such as the rocks in the water, and the treeline in the upper left corner where

it meets the sky. This requires a finer brush to mask these specific areas, and remove

what we call halos - identified by the red arrows in the top image.

Halos are the result of transitions that are too general - the brush I used was large and

soft in order to quickly blend areas together, as seen in the mask in the bottom image.

As a result, the blend wasnt customized to the distinct contours of my landscape,

which produced this unnatural glow - a halo. Youll usually find halos where the

sky exposure meets the ground exposure (as seen to the right), but they can appear

anywhere that youve added a mask to. This requires you to return to the smaller

details and edges of your image and work those for a flush and accurate blend.

This is one example of where manual blending outshines HDR software and GND

filters - you can take your brush and address halos individually and with 100% accuracy

rather than rely on the global adjustments of software or the fixed blend from filters.

By using layer masks, the blends are entirely reversible, and completely customizable.

To illustrate the importance of taking this extra

step, Im going to blend my overexposed layer with

my base layer to increase the exposure of this rock,

which is isolated in the water and looks unnaturally

dark for my landscape.

With my brush set to 100% opacity and 0%

hardness, I brush over the rock as carefully as I can

- zooming in only far enough to have my subject

fill the screen. Despite my best efforts, I have a

noticeable halo surrounding the edges (bottom left).

When I zoom in further to see this in greater detail

(right) and press the \ key, you can see how my

highlighted brush strokes spill over onto the water.

With this closer view, I can use a smaller brush

and remove this overflow, which will make my

mask flush to the rocks edge. Since Im removing

the layer mask of my overexposure, my foreground

brush color is set to white and with 100% opacity.

Since my brush size is small, the hardness level is

still set at 0%.

When compared side-by-side, you can see how a well-

placed mask affects the blending. The top two images

show a mask that was applied quickly, while the bottom

row displays more accurate brushwork that hugs the

contours of the rocks.

At times youll find that even the most accurate

brushwork can make your scene appear unnatural and

distracting as it creates a hard transition between tonal

ranges. Youll often see this along the horizon line where

the sky meets ground.

After you perform your detailed brushwork, create a

history snapshot by clicking on the camera icon on your

history palette. This will allow you to revert back to this

particular point in your editing in case you wish to start

over.

With a very low opacity brush (around 10% to start),

begin to brush around the area where the transition takes

place. Use 0% hardness, and experiment with different

brush sizes to make the transition appear smooth and

gradual - in other words, there should be no evidence of

brush strokes or sharp transition of tones. This will allow

you to blend these two exposures for a more gradual

appearance, much like I did for the shoreline previously.

advanced blending When blending multiple exposures together in Photoshop using layer masks, youll eventually run into a situation where refining your mask with freehand

brushwork becomes too difficult and time-consuming.

This is especially true for intricate, jagged horizons that

are interrupted by trees, rocks, and other elements

such as the trees to the left. Brushing over the detailed

contours of this tree line would require hours working

on a pixel-by-pixel basis, which is simply unnecessary if

you know how to refine a selection.

The following advanced blending workflow will show

you how to combine the longer exposure (+2 EV) of

the trees with the faster exposure (-2 EV) for the sky,

which eliminates all the blocked shadows and blown

highlights and recovers any lost data from this dynamic

tonal range. Most importantly, this step-by-step guide

will show you how to achieve this seamless result by

creating a refined selection that automatically masks

these two exposures together with great accuracy.

By exploring some of the advanced features of

Photoshop, you can create a customized layer mask that

instinctively hugs every contour for a truly seamless

HDR blend of exposures.

whentouserefinedmasks

Before I get into the specific workflow, I want

to highlight when and why you should forgo

the freehand brushwork method and rely

on Photoshop to select your masking. If you

would rather just skip ahead to the step-by-

step workflow, you can jump to Creating the

Selection. You can always come back and read

this section if you have any why and when

questions as you go through each step.

A large aspect to the freehand brush method

of blending exposures is what I call partial

blends, where you use a low-opacity brush and

paint over the areas you want to blend together

gradually or rather, not 100% transparent.

The result is a layer mask of different values,

with a certain percentage of your top exposure

transparent so that the exposure underneath

can be seen to a certain degree with the goal

being to find a happy medium between the two

for a natural appearance and a more complete

tonal range. You can also use the partial

blend method in order to create a gradient-

like appearance of tones where you gradually

transition from one exposure to the next

much like the effect achieved with a GND filter.

The example images from page 22 (also shown

to the right) show a partial blend executed well.

Partial blends are a fast and easy way to have

full control over how your exposures (layers)

interact with one another by being able to

manually adjust the opacity of the brush you

use to blend and for most situations, this

method works well. However, for jagged or

serrate horizons (such as mountain ranges,

tree lines, and other subjects that disrupt a flat

horizon line), partial blends can present several

side effects that detract from the success of

your final image.

When I do a partial blend of my example image below in an effort to

gradually combine the -2 EV sky and the +2 EV tree line, youll notice that

the tree line gradually gets darker as I approach the sky. This is where the

freehand brush method falls short, and more refined blending is necessary.

Since the trees are interwoven into the sky line, I have to brush over these

tree tops in order to blend in the -2 EV sky, which causes the darkened tree

line (top). My only other option here would be to do the reverse brush

over the tree line to blend in the +2 EV layer, which causes some unnatural

haloing around the trees (bottom).

When attempting to blend two exposures together with a large, low-

opacity brush instead of using a refined mask, you generally have to

choose between these two side-effects if your horizon line is not flat

either darkened tree tops to compensate for the -2 EV sky, or haloing to

compensate for the +2 EV trees. You cant blend the exposure of the trees

with the exposure for the sky seamlessly by using a partial blend, so this

is why pixel-by-pixel brushwork is often turned to in order to blend in

each tree branch individually incredibly time-consuming, but the only

way to create a truly customized blend. This is where advanced blending

techniques come in handy because rather than resorting to either a

partial blend (shown on the right) or tedious brushwork, you can render a

customized, contoured selection at full opacity.

Also, partial blends with a large brush present other issues besides this

gradient-like effect on tones. If your layers arent uniform in alignment,

then ghosting will appear the partial visibility of an element that is

present in one exposure, but not the other.

Even the most careful in-the-field execution of your

brackets can still produce inconsistencies between each

exposure. A common cause for ghosting is elements that

move within your frame such as the fluctuating tide,

swaying tree branches, and moving clouds.

When blending different exposures together, elements

may have shifted, so partial blending is something you

should avoid when alignment isnt uniform. By making

a custom selection, you can forgo a partial blend and

choose exactly what you want to mask, and at 100%

opacity.

Panoramic stitches are another example of when

images may not be uniform in alignment, and not

due to bracketing errors in the field. Even the most

sophisticated panostitching software with manual

links can produce minor discrepancies between each

exposure, and Adobes image alignment feature doesnt

always help either. In these instances, partial blends

would produce ghosting and thus, more specific

blending at 100% opacity is necessary.

ghosting & sensor bloom

Sensor bloom occurs when light bleeds or overlaps onto elements in front

of it, beyond where you would normally see the light, even with perfectly-

aligned layers. This is especially visible in high-contrast scenes, such as the

darkened pine trees against the sunset sky. This overexposed light spilling

around and in between the trees in my +2 EV image has given the illusion that

the trees are smaller, which is seen as ghosting for partial blends since the

underexposed trees were rendered larger.

When you compare the difference between -2 EV (exposed for the sky) and

+2 EV (exposed for the trees), you can see how sensor bloom affects the tree

line. The +2 EV trees were rendered considerably smaller in size due to the

overexposed sky, and this difference will be apparent when the exposures are

combined.

Sensor bloom is a common occurrence with exposure blending especially

with trees since many scenes that require blending are taken during the

golden hours with foreground subjects backlit by the sun. Even with a

perfectly-contoured selection with 100% opacity masking, Ill still have an

overlap to address as the trees in the -2 EV image are larger than they are

in +2 EV. Ill discuss how to correct this later (which is rather simple), but

for now I want you to keep in mind what sensor blooming is and why this

overlap would cause ghosting for partial blends.

The only way to rectify the blending issues with partial blends is to go in with

a very small brush and work on an almost pixel-by-pixel basis, selecting each

individual crevice to mask out. Admittedly, Ive done this in the past (with

great success by the way), but it took several hours of detailed work. Rather,

theres a more automated way to select your edges by working on specific

channels and using your overlay brush.

As a quick overview, Im going to create a high-contrast scene by selecting one

channel from the layer I wish to mask. The strong contrast will serve as a way

for Photoshop to detect a hard, contoured selection, which I can then use to

add a customized mask.

The typical approach for making a selection would be to use your magnetic

lasso tool or select a specific color/tonal range. Unfortunately with these

methods, too much guesswork is left to Photoshop which makes them highly

ineffective for refining a selection destined for exposure blending. Often, youll

have to re-visit the mask and contour it yourself with your brush tool, which

defeats the purpose of automated selections.

Once you have all your images aligned as individual layers in Photoshop (see

Preparing to Blend), select your top layer (fastest exposure) and click the

Channels tab to open up each individual channel for this particular layer.

Here we can see each channel on their own layer, making it easy to identify

which one to use for a selection. Since the goal here is to create the most

contrast between what we want to mask and what we dont for easy edge

detection, Im going to look for the channel that has the most contrast. For my

image here and for most cases its the blue channel as the value disparity

between tree and sky is higher here than the other channels.

creating the selection

Since Im going to be making adjustments onto my blue channel directly in

order to increase the contrast, I dont want that to be visible on the image

itself. I drag my blue channel down to the create new channel icon to

copy the channel (or just right-click and select duplicate channel). Now

that this channel is hidden underneath the others, the changes I make here

wont be apparent when I switch back to the Layers tab later.

Now Im ready to increase the contrast between sky and tree,

which will help Photoshop detect an accurate edge for selection.

There are several ways to go about increasing the contrast here, but

I like using the curves tool as it gives me direct control over my

tones. I click Image > Adjustments > Curves to bring up the curves

tool dialog box. Notice how the channel selected is Blue copy

the duplicate channel I just added.

What I want to do here is eliminate that grey from the sky and

slightly darken the tree line in other words, push the sky pixels

closer to pure white, and the tree pixels closer to pure black. I take

my white point and slide it over to 107 on my x-axis (input), and

slightly increase the input of my black point to 4. This compresses

my tonal range quite a bit, telling Photoshop to reassign most of

those grey pixels in my sky to pure white (255).

Whatever method you use to increase the contrast, make sure not

to compress the tonal range to the point where the edges of your

future selection are redefined. If you increase the contrast too

much, you can cause the black pixels to bleed onto white, or vice

versa. Well be refining the selection later so a perfect contrast of

black/white isnt necessary this step is to simply create a cleaner

canvas for easier edge detection.

When finished, click the load channel as selection icon to create

your selection.

What this will do is create a custom selection based on

the value of each pixel which we can then apply as a layer

mask; pixels closer to pure white (255) will have a stronger

selection, and pixels closer to pure black (0) will be selected

less. In other words, pure white will be 100% selected, pure

black will not be selected at all, and the opacity of each

selection for the varying shades of grey in between these

two points will depend on where they fall on the value

scale. This will make more sense visually, so Ill apply the

selection to a layer mask so we can see the result (right).

After the selection has been made, switch back to the

layers palette and make sure your top layer (or whichever

layer you want to apply your mask to) is active. Click the

add layer mask icon to apply the selection as a mask, and

ALT+click the actual mask to see it on your image.

Note: I have hidden my base layer here (+/-0 EV) in order to

access the longer exposure underneath (+2 EV) since that is

the image I want for my tree line. Im only incorporating

the middle ground from this base exposure (detailed here)

so once Ive finished my masking here, Ill go back and

activate this layer and paint in the middle ground water

using the freehand brushwork method described earlier.

Now we can see the result of this entire process a

completely customized and contoured layer mask for

incredibly accurate blending.

refiningthemask

Im going to zoom out a bit so you can see the entire frame, and

also see those shades of grey where the partial selections were

made. When blending exposures in this manner, grey translates to a

partial blend of exposures which is something I dont want here.

It means that Ive only partially blended my layers, which can cause

ghosting and an overall flattening of tones. The goal here is a pure

white-and-black mask that is contoured to the unique horizon line.

If I select the layer to see how this mask interacts with the blending

of exposures, you can see how flat the tones are towards the center

of my image, which is where the tones closest to middle grey are on

my mask.

While the general outline of the mask has been established, I still

need to refine it by reassigning all those grey tones on the mask to

either pure white or pure black. At this point, all the hard work has

been done the borders of the mask have been outlined. Now I just

need to fill those areas in.

Im going to go back and re-select my mask by

ALT+clicking it so I can see it on my photo. What Im

going to do here is quickly select most of the pixels I

want to be pure black basically, everything but the

sky and fill it with 100% black. Afterwards, Ill refine

the edges with my brush tool.

By selecting my lasso tool (not magnetic) and making

sure my mask is active instead of my layer, I trace

around the area where I want everything to be pure

black (fully masked). Its more important to only

select the pixels you absolutely want to mask as

opposed to getting as close to the edge as possible.

Well be refining the edge later (which is an easy

process), so dont worry about leaving some breathing

room.

Once selected, make sure your layer mask is still

active and click Edit > Fill. Select black as your fill

color with 100% opacity, and click OK. Youll end up

with a mask similar to the result below almost

perfect, but with a slight border of various grey tones

to refine.

As you can see, this is filled in quite nicely except for the top of the rock

down in the corner. Its various shades of grey, which translates to flatter

tones and possible ghosting in my image due to partial blending. What I want

to do is fill in this area with pure black. In other words, everything south of

the outline should be filled with black, and everything north should be white

all those grey values will be pushed to either pure black or pure white.

The easiest way to achieve this is by using the overlay brush. What this will do

is take any pixel that is a higher value than middle grey and lighten (screen) it

towards pure white, and inversely, it will reassign any pixel lower than middle

grey and darken (multiply) it towards to pure black. Now you can see why it

was important to create contrast in our channel not just to get a defined

edge for detection, but to also push those grey pixels that were on the

fence one way or the other, which will make the overlay brush all the more

effective.

I switch over to my brush tool and make sure my foreground is set to black,

and background set to white. On the brush toolbar, change your Mode to

Overlay from the drop-down menu and change your opacity to 50%. This

is where I typically start with in terms of brush strength, but feel free to

experiment with the opacity here to what suits your specific needs.

Now its just a matter of brushing over your canvas and pushing pixels

towards their extremes either pure white or pure black. A black overlay

brush will push grey pixels to black, and a white overlay brush will push those

pixels towards white. For this particular image, I paint over the sky with a

white overlay brush to make sure that all those pixels are 100% white (not

masked at all), and then focus in on the tree line to refine the mask selection

(image on next page).

I like to make several passes over the area, switching back and forth from

black to white as I brush to keep things on an even keel. Keep in mind though

that the more passes you make of the same color (black or white), the more

chance it will bleed onto the other. Even pixels that are near 100% white will

be affected by a black overlay brush if those pixels are on the border and vice

versa and this change becomes more dramatic the stronger your opacity is.

This can alter the size and contours of your newly-customized mask, so make

sure to keep an eye on your brush work and take note of any changes in size.

This is why I like to alternate back and forth from black to white as it helps to

make sure one color doesnt dominate over the other.

the overlay brush

After making several passes with an alternating black and

white overlay brush, my mask is looking much cleaner (top

image).

Now I can see the refined layer mask that Ive been

working towards completely customized to the unique

horizon line of my image, and something that would take

many hours to create by a freehand brush.

Lets take a look at the results to see how this mask

interacts with the layer exposed for the tree line (bottom

image).

While my mask is successful, this blend is not. Here you

can see a common issue with this method and exposure

blending in general inconsistent layers. However, since

we have a custom layer mask already in place, this sensor

bloom quite simple to fix.

As mentioned earlier in this tutorial, sensor bloom occurs

when the physical area of an object is reduced in apparent

size due to bright light sources pouring in from behind. This

transition between light and dark is overpowering, and the

overexposed light of the sky in my bottom layer (+2 EV) is

bleeding over onto the trees making them appear smaller.

The difference in size between the image exposed for the sky

(-2 EV, where I based my mask on) and the image exposed for

the trees (+2 EV) is quite different. Sensor bloom is why you

see the glowing edge of light around the trees the mask

is larger than the +2 EV trees, so its spilling over onto the

overexposed sky of this layer.

This can be seen when I highlight the layer mask. The trees

of +2 EV, which has all blocked shadows in the foreground

recovered, are masked in with a selection based on the -2 EV

trees, which is slightly larger.

Another issue we see here is simple inconsistencies between

panoramic stitches those ghosting areas identified earlier.

The software I used for panostitching rendered the -2 EV

stitch slightly different than the +2 EV stitch, and this

disparity affects how the mask interacts with these layers.

This ghosting can happen for various reasons, but usually

affects only a small area of your frame.

sensor bloom & exposure blending

To rectify both the ghosting and sensor bloom, Im going to duplicate the

ground layer underneath my mask (+2 EV) and with my clone tool, fill in to

cover the white areas from sensor bloom. Right now, the selection I created

for the sky layer mask (-2 EV) is too big when placed over the ground layer

exposed for the trees. However, some select clone stamping will correct this

unsightly distraction.

Making sure that the duplicate ground layer is on top of the original and

active, I select my clone stamp tool and choose a stamp source below the

sensor bloom by ALT + clicking on the trees. I then brush over the tree line

edges with a soft brush (0% hardness) and at 100% opacity, with a size large

enough to cover the areas I want to correct. Both the sourcing and stamping

are performed on the same layer, the +2 EV ground layer I just duplicated.

When you move about your frame, you may need to reselect new sources as you

go along to avoid repeating any patterns for the image above, notice how I

selected a new source area to correspond with the new stamping area. Since Im

working with pine trees, the pattern is very chaotic and random and the mask

is serrate, so it makes for easy stamping with no noticeable repeating patterns

or otherwise unnatural blending. However, for other scenes, you may need to be

more conscious on where you source your stamping and how much you perform

before selecting a new source.

Its also important to try and be conscious of your clone sources and where you

stamp for example, notice how I sourced a tree trunk and stamped it where the

tree trunk had been overwhelmed by sensor bloom as opposed to filling that area

in with random pine needles.

When finished, my sensor bloom and other inconsistencies between the layers have been removed for a seamless blend.

Another way to relate this process to something more familiar is to think of

filling in the edges around a stencil where the sky layer mask is the stencil,

and the trees are what fills it. Before the clone stamping, the stencil was too

big and only partly filled in, so the trees needed to be enlarged to occupy

those negative spaces (the sensor bloom).

When you take the stencil off (the layer mask) and look at the duplicate

ground layer where I did my stamping, the appearance is rather sloppy and

unrefined (below). However, when the stencil is put back on (previous page),

those sloppy edges dont matter much since you only see what the mask

(stencil) allows you to. The final result is a perfectly customized and highly

intricate blend of exposures, where the -2 EV sky is 100% visible around every

contour and through every gap along the horizon line.

finaladjustmentstotheblendWhen blending two different exposures

together in this manner, you can sometimes

encounter an unnatural transition where

the foreground looks washed out against

the background. Since the mask has already

been drawn, and any blending discrepancies

between each exposure corrected, I can now

adjust the tones of +2 EV to match the sky in

a more natural way.

Making sure to reactivate my -2 EV layer (sky)

that I just hid to show the stamped duplicate

layer, I add a curves adjustment directly on

top of the duplicate and underneath mid

(+/-0 EV) and shift the middle tones down

closer to pure black. The placement of the

curves layer here is important so that it only

affects my +2 EV image. Since Im working

on a curves adjustment layer, I can mask out

areas of this change where it is too drastic,

allowing for further refinement.

The best approach to this would be to actually take

an exposure at +1 EV (faster) rather than +2 EV so

that the curve adjustment step can be skipped, but

that can present its own unique set of issues. With

this particular image, the darker areas to the left of

the frame had blocked shadows at EV+1, so I instead

chose to use +2 EV and applied the curves adjustment

layer to reduce the tonal value closer to the horizon

line. I then masked out any areas that became too

underexposed such as the darker trees towards the

left of my frame. In my eBook The Art of Processing,

I discuss this method in further detail and explain

how to use layer masks to selectively apply your color

and tonal adjustments rather than settling for global

changes.

With this in mind, I always check my histogram in the

field to make sure that I have recovered all lost data

in both my highlights and shadows. In other words,

Id rather collect too much data and adjust it in

post as opposed to not enough and be left without a

choice. In theory, I could have captured both an +1 and

+2 EV frame and blended in the underexposed trees of

+2 EV later, but that seemed like an unnecessary extra

step since all the data is there in +2 EV.

I have one more step to perform before I can call

this blend complete. At times, the edges of your

masking can be rather harsh and unnatural, making

the transition between exposures appear pixelated. By

refining the mask edge, you can soften this transition

for a more pleasing appearance.

Before I edit my mask, I want to duplicate the entire

layer and hide it in case I need to refer back to it

later. The Refine Mask tool can directly manipulate

the pixels of your mask which means any changes

to those delicate adjustments performed to create

this mask will be permanent. By duplicating this layer

and the mask, I have the original to revert back to if

necessary.

Once Ive made my duplicate, I go back to my top

layer and right-click on the mask, which brings up

a menu. I select Refine Mask to open up several

powerful options most notably the Adjust Edge

section. Since my goal is to soften the edge of my

mask for smoother transition, I very slightly increase

the Feather slider to .6 px, which makes quite a

difference for this image. The amount of feathering

you perform will depend on the size of your canvas

since its a pixel-dependent adjustment, but a little

goes a long way to smooth out those jagged edges.

Its important to note that there are several other fantastic

options here for refining the mask most notably, the Shift

Edge feature will allows you to increase or decrease the size

of your mask edge. The only adjustment I need to perform to

this mask is feathering the edge, but I want to make sure that

you dont overlook the Refine Mask feature as it has several

versatile tools that you should explore for future reference.

When I compare the end result of this workflow with those

of the freehand brush method shown earlier, you can see

why a refined selection for my mask was preferable to simply

painting with a low-opacity brush for a partial blend.

While the refined mask here works much better for the unique

challenges this horizon line presents, that is not to say that

freehand brushwork is an inferior method to use its just

not suited well for this particular image. For example, if this

was a bracketed scene of a seascape with a flat horizon line,

partial blends with a low-opacity brush would be my method

of choice as it would create a gradient-like blending of tones

as opposed to the harsh, abrupt transition that comes with

a refined mask like this. In short, the method you use will

depend entirely on the contents of your photograph one is

not universally more effective than the other.

After my advanced brushwork along the sky and tree line, my layer

blending is now complete. When you compare the final result with

the base exposure - the image one would normally capture - you can

see a huge difference in the tonal range and detail. This recovery of

data is also reflected in the histograms - notice how I now have a

nice arc distribution of tones. In contrast, the U shape of my base

exposure indicates that I have a high amount of boxed shadows, blown

highlights, and very little mid tones to work with. While there is no

global definition of a good histogram, for this particular scene it has

been improved greatly.

The amount of blending you perform will depend entirely on your

scene, so there is no typical workflow to follow. Your blending may

be simpler or much more involved than this example, but thats the

beauty of exposure blending. Its entirely customizable to compliment

your photograph rather than trying to force your photograph to

compliment the confines of automated software.

When blending exposures, there are oftentimes situations where

gradual, soft blends are preferred to make the transition between

layers more natural for complex scenes - for example, the water middle

ground in the top image would benefit from a softer blend as the

transition is too abrupt. While you can soft-blend exposures together

using a low opacity brush, like explained on page 21, theres a more

advanced way to blend that will allow you make custom selections

like this for scenes with complicated or dramatic tonal shifts. For the

remainder of this eBook, Im going to introduce luminosity masks into

the workflow, which will give you the highest level of control over

your blends and transform the way you process your scene.

Luminosity masks are a powerful editing tool for

landscape photographers as the desire to capture

dynamic tonal ranges has become increasingly

popular, and this often requires blending different

exposures together in Photoshop in order to

recover any lost detail. In the previous section on

advanced blending, I explained how to create refined

selections for your layer mask to assist. That method

involves loading a custom selection based on pixels

that were either lighter or darker than middle grey

onto a layer mask, and then refining that selection

to the contours of your landscape.

An issue with this process is that it can be rather

rudimentary for more complex images that have

difficult transitions between exposures. To take your

mask refinement a step further, you can select more

specific groups of tones based on their luminosity

level rather than whether or not they fall before or

after 50% grey on the value scale.

Tony Kuyper first wrote about luminosity masks in

2006 as a way to make tone-based, local adjustments

to one exposure in Photoshop. However, these

masks can also be used to blend exposures together.

Luminosity masks are simply a unique and versatile

approach to creating a selection based on the tones

of your image and that selection can be used to

mask layers together in the digital darkroom.

Luminosity masks are created by loading a selection

that was created on a channel much like I did in

the previous section. The difference, however, is that

a luminosity mask does not simply create a black

and white mask, but one based on luminance for

gradual, natural transitions between exposures. A

black and white mask produces a hard transition;

when I showed you how to refine a mask previously,

I created a strong contrast directly on the channel

so the selection was well-defined. I then further

refined the mask so the trees from the +2 EV layer

were completely, 100% visible with the sky from -2

EV. In other words, all grey tones on the mask were

eliminated. However, this harsh transition may not

always appear natural, and can create an unpleasant

result when dealing with complicated, interwoven

transitions between light and shadow for example,

trees backlit by the sun with light rays pouring in

from behind. For situations like this, you can create

a soft, gradual transition between your exposures

by selecting more specific groups of tones that are

automatically feathered.

introduction to luminosity masks

creating an alpha channel

To begin, I want to first blend the sky of -2 EV with the foreground of +2 EV. This is basically the same thing I did in the

last tutorial, but instead Ill go through the luminosity mask creation so you can easier understand how this works.

Select your least exposed layer (-2 EV) and click on the Channels tab. Much like in the previous tutorial, select the

channel that provides the most contrast overall and load it as a selection by clicking on the selection icon at the bottom

of your palette (below) or simply CTRL/CMD + Click the thumbnail (CTRL for PCs, CMD for Macs). This will bring up the

marching ants around pixels which are brighter than middle grey, and the strength (opacity) of that selection depends

on how bright the pixels are.

Now that the selection is made, were going to save this selection

so we can refer back to later. The easiest way to do this is to

create an alpha channel its a duplicate of the channel youre

basing the selection off of (in this case, the blue channel) with

the selection applied to it. Alpha channels are simply a way

for you to save your tonal selections to use later since you

can always return to your Channels palette and select an alpha

channel to activate the selection. Dont worry, alpha channels

wont affect how your image appears since theyre hidden.

To create your alpha channel, make sure your selection is

still active (marching ants are still visible) and click the Save

selection as channel icon at the bottom of your channels palette.

This will automatically create a new alpha channel based on your

selection. Rename this layer Highlights for easy reference as

this is the channel youll base your future highlight selections off

of.

Next, I create the Shadows alpha channel by duplicating the Highlights channel I just created and inverting it by pressing

CTRL/CMD + I make sure that before you invert, no selections are active by pressing CTRL/CMD + D to deselect everything.

Otherwise, inverting will only affect the selection and not the image as a whole. For easy reference, rename this alpha channel

Shadows. Since the load channel as selection feature only selects pixels which are higher than 50% grey on the value scale

(closer to pure white), inverting the alpha channel means that the pixels which are brighter than middle grey are actually

your shadows.

What Ive done is split the two tonal ranges

the pixels which are brighter than middle

grey and the pixels that are darker into two

main selection pools: Highlights and Shadows.

These two selection groups, which are in the

form of alpha channels, will be the source of

all your future refined selections (at least for

this exposure). This is what luminosity masks

are based on making a selection on a group of

tones either highlights or shadows in order to

blend them with another layer. Since this method

creates a feathered selection, the transition from

one layer to another is almost undetectable for a

truly natural blend with minimal posterization.

This is where luminosity masks differ greatly from

simply creating a black and white mask you

can get very specific with your tonal selections

by narrowing the parameters which define

highlights or shadows. For example, if I only

wanted my very brightest pixels selected in order

to blend any blown highlights with a lesser-

exposed layer, I would select my Highlights alpha

channel and instruct it to only select pixels which

are completely white (255 on the value scale). This

is what Ill be discussing next how to take either

of these two alpha channels and narrow your

tonal selection to base your luminosity mask on

later.

Before I begin to create more refined alpha

channels, I should note that partial blending

should only be used when each layer (exposure)

you are blending together is uniform in alignment

otherwise, ghosting will be an issue. If youre

working with only one RAW file (not ideal, but

adequate enough if you have no other option) you

wouldnt have to worry about any misalignment

since youre working with the same image, so

theres no movement between frames. If you

bracketed off a scene and had no movement

between frames, youll be fine as well. The

big issue with ghosting comes when you are

blending panoramics together even the most

sophisticated software and seamless blend can

still produce inconsistencies between each layer.

So before proceeding, you should make sure that

your layers are uniform in alignment (by first

selecting all your layers and pressing Edit > Auto-

Align Layers) if they are still not aligned, make

sure you are comfortable with reversing any

potential ghosting manually before going through

this workflow.

highlight vs. shadow selections

The overall goal here is to create a luminosity mask to

combine the sky of EV -2 (sky) with the foreground of

EV +2 (ground). This is what I want to focus on now, so

Im going to deactivate my middle exposure (EV +/-0) and

hide the layer so that there is nothing between the sky

and ground layers (top right). Ill be bringing this layer

back in the next step in order to gradually blend in the

water since the middle layer has the best exposure for it.

Since the layer we are selecting is quite underexposed

(sky), there arent going to be any highlights for our

selection. This is confirmed when we look at the levels

histogram for this particular exposure (bottom right).

Almost all of the tones are to the left, which means the

Highlights alpha channel which only allows pixels

brighter than middle grey to be selected is useless.

Instead, what Ill do is base my selection on the Shadows

alpha channel which has a vast selection pool to choose

from and then invert my mask. If youre lost already,

dont worry visuals help, so Ill walk you through this

tutorial step-by-step.

refiningluminosityselections

To begin, Ill select the Shadows alpha channel by clicking

on the thumbnail, and then activate the selection by CTRL/

CMD + clicking the thumbnail again.

By looking at the selection here, you can see that while

all of the shadows are selected, theres still quite a bit of

sky detail that I dont want. This is where the true power

of the luminosity mask comes in, the ability to condense

and refine a selection based on the value of each pixel. In

other words, my selection is too general right now and has

overlapped into the sky. With a luminosity mask, I can tell

Photoshop to take this selection pool and narrow it down

to only include the darker shadows of this channel.

You can compare this process to wood carving or clay

sculpting I have the general outline of my figure, but

need to whittle away some of the rougher selections in

order to refine and mold it into what I want. Its been a

long explanation so far, but quite worth it when you see

the power of this process and why many photographers

use luminosity masks for their photography beyond

exposure blending.

What Im going to do here is take this alpha channel and intersect the selection with itself, which will tell Photoshop

to eliminate those pixels on the very edge of the selection the lighter shadows and only select darker tones: in

other words, reshape and refine. To do this, press CTRL/CMD + ALT + SHIFT all at once, and click on the thumbnail for

the Shadows alpha channel.

Here you can see the difference only one intersection made see how the selection now hugs the contours of the tree

line, and the sky is (almost) completely deselected?

The next step would be to save this selection to use

later by pressing the save selection as channel icon and

renaming the channel for easier future reference. I like to

simply add a roman numeral when I intersect channels, but

feel free to use your own labeling methods.

Looking at the refinement of this selection, youll notice

how similar it appears to the one I created in the previous

section on advanced blending, where I added contrast to

the channel directly in order to refine my mask. However,

theres one very important difference with luminosity

masks which isnt apparent by simply looking at the

selection here the selection is self-feathering, which

means that the transition from one exposure to another

will be very soft and gradual. The marching ants you see are

simply pixels that are selected with more than 50% opacity,

but pixels which are selected with less than 50% opacity

will not be outlined by the marching ants. So its possible to

have no marching ants on your alpha channel and still have

a selection.

I need to refine this selection a bit more in order to get the results I want, which is everything but the

sky. I intersect the selection once again, making sure

that my Shadows II alpha channel is active and the

selection is loaded. CTRL/CMD + ALT + SHIFT while

clicking on the layer thumbnail will again further

refine it to select even darker shadows.

Im going to stop intersecting here since the next

intersection starts to cut into the tree line a bit

too much. You can do this process as many times

as you need to, and can actually create your own

set of Photoshop actions to do this process for you

automatically.

You can do the same exact process for your

Highlights, gradually narrowing your selections

until only your very brightest areas are chosen a

good selection method for working on any blown

highlights. For this particular exposure, the scene is

quite underexposed so it made more sense to base

my selections off of the Shadows alpha channel.

However, if I was working with EV +2 or any other

overexposed layer, I would have created luminosity

masks based on the Highlights alpha channel and

selected the overexposed sky for my masking. So

the channel you choose to work with for blending

either Highlights or Shadows will depend on the

exposure of the image you are basing it off of, and

also your intent (what areas you want to select for

masking later).

My channel work is done; the selections have been created for blending. However, this is only half the process I need to transfer these selections over to my layer and apply them as a mask to blend the exposures

together. This process is rather simple, and there are several ways to apply your selection and create the mask.

Now that the selections have been made and saved as alpha channels, its time to turn them into a luminosity mask and

blend the exposures together for an HDR photo. Turning a selection into a mask is quite simple, and Ill go over two

different ways to approach this.

First make sure that your selection is active, and is based on the correct alpha channel for this image, its the Shadows III

channel. You can activate this by first clicking on the channel thumbnail to highlight it, and then CTRL/CMD + Click on the

thumbnail again to bring up the marching ants. Once the correct selection is active, switch back to your layers palette and

select the layer you wish to apply the mask to. Apply a new mask to your layer, and the selection will transfer over.

creating the luminosity mask

Alpha channels are based on selections of pixels that have higher concentrations

of white the closer a pixels value is to 100% white, the higher the opacity is

for that selection. When you apply an alpha channel selection to your mask,

youre transferring over the tonal ranges of that channel directly to the mask. In

other words, your channel looks exactly like what your mask is going to be. For

this particular image, since the mask was inverted for shadows, its doing the

exact opposite of what I want it to do the sky is predominantly black and the

foreground white, which means that the sky is actually getting masked out and

the foreground is still visible. This is a simple though as all I have to do is invert

my mask by pressing CTRL/CMD + I and since Im aware of this beforehand, Ill

hold down the ALT key when I apply the layer mask to instantly invert it.

This blend is starting to come together, but for this particular set of exposures

a gradual transition does not work well especially for the tree line. Theres

also some sensor bloom and ghosting issues to address. Referring back to the

previous advanced blending workflow, I refine my mask to eliminate all the

grey tones by filling the foreground with black and sky line with white first

generally with the lasso tool, and then more specifically with the overlay brush.

After refining the edge of my selection, the result is a fully

contoured mask.

As far as masking goes, this is exactly what I want the +2

EV tree line and foreground blended seamlessly with the -2

EV sky. However, this hard transition does not look natural

for my middle ground and horizon line. The water you see in

the bottom right image is from the +2 EV layer, which is quite

overexposed when compared to the -2 EV sky, so this dramatic

jump of 4 stops is too harsh.

Another option would be to blend in the water from -2 EV, but

this appears to be just as unnatural (inset). A gradual blend

is needed to transition softly from the -2 EV sky to the +2 EV

foreground by introducing a third exposure the base image

(+/-0 EV) which I hid at the beginning of this eBook.

transitions with luminosity masks

To begin, Ill unhide my base exposure and switch to the Channels palette. Here

youll notice something new all the alpha channels I created based on -2 EV are

still accessible. This is very useful since the alpha channel selections you create

can be applied as a luminosity mask to any layer (exposure).

For this image, however, I want to create a new alpha channel based on my

middle exposure as that has the greatest contrast between the water and the

surrounding land my goal here is to create a selection around the water and

isolate it, and then gradually blend in the water exposure of +/-0 EV with the +2

EV exposure to softly transition between these two extreme tonal groups.

Looking through my RGB channels, blue provides the greatest amount of contrast so I create an alpha channel just like I

did before load the selection, and then save that selection as a channel (Highlights I mid below).

Since my goal is to just select the water for blending, I intersect this selection once to further target the brighter

highlights: CTRL/CMD + ALT + SHIFT and click on the alpha channel thumbnail, and then save the selection as another

channel (Highlights II mid).

At this point I would normally just add a

layer mask and refine as necessary, like I did

previously. However, for gradual transitions,

you may want to apply your luminosity mask

in a different manner than simply loading

it directly onto a layer mask. I find it more

successful to transfer over just the selection

made with your alpha channel, and create your

luminosity mask by manually blending it in

with the brush tool at a low opacity level. Its

a lot like what I did earlier in the eBook on

manual blending, but now I have the confines

of a feathered selection that I can simply fill

in in other words, the selection I created acts

as a mold that I can then paint my luminosity

mask into, and at varying transparency levels.

To do this, make sure your alpha channel

selection is not active (no marching ants) by

pressing CTRL/CMD + D to deselect. Then switch

over to your layers palette, select the correct

layer and add a mask. The color of your mask

is an important consideration here depending

on what you want to accomplish remember,

white retains and black removes your layer.

Since my alpha channel selection is based on

what I want to keep the water the mask

will have to be black as I can only paint within

the confines of that selection. To add a black

mask, I hold down the ALT key while pressing

the add layer mask icon.

Now that the mask has been applied, switch back over to your

Channels palette and load the selection you just created. Once

the selection is active again, simply return to your layers palette,

select your mask, and begin to paint your luminance selection

onto the mask with a white brush.

I recommend you begin by using a very low opacity brush to

gradually apply your luminosity mask sele