HDR Tutorial 2

I was hoping to put this tutorial up a while ago, but never got the chance to complete it until now. I received a lot of requests from people who wanted to know the workflow on the Times Square image(below). It’s fairly similar to my previous tutorial (the first few stages relating to the camera setup and Photomatix processing are pretty much the same), although this one concentrates more on the post processing in Adobe Photoshop.

If anyone wants to practice with the original images, let me know and I’ll upload them. You can see my original HDR tutorial here.

* You can enlarge any of the screenshots below, by clicking on the image which will take you through to the larger image.

This is the final image.

Step 1 – Taking the Image

So here we go then…Firstly there are two main ways to create the source images needed for HDR.

- Autobracketing
- One RAW image

Autobracketing:

Most DSLR camera’s have the ability to bracket multiple exposures. This is simply a way to tell the camera to take several shots in quick succession, each at different exposures.

When deciding on how many auto-bracketing exposures to use, you will need to consider the contrast range of the scene. If there is a higher contrast range, consider using a wider bracket setting.

Now this may not be possible on some cameras. My own Canon 400D only has the ability to bracket 3 shots, but some cameras are able to go as high as nine. Normally 3 exposures are suffice for me when shooting in interior locations, such as churches and other buildings, but when shooting outdoors 5 or more may be required.

When I do need more than 3 exposures, I normally take auto-bracketing off, go into manual mode and fire out 5 or more shots by manually changing the shutter speed each time, but keeping the aperture the same.. For this image, 5 exposures were taken at -2, -1. 0. +2 and +2.

One RAW image:

The other way to produce the 3 shots needed is to take 1 photo and adjust it in a RAW editor such as Lightroom, Aperture or any other one you may use.

The main advantage with this is that you can produce a HDR shot with moving subjects such as cars and people. The downside from experience is that if an image contains dark shadows, the exposure adjusting followed by the HDR process has a tendency to create a lot of noise. The other thing that is debated quite often is that by taking 1 raw image, it doesn’t truly capture the full dynamic data range of a scene.

Camera Settings

-Shoot in RAW – it allows you to capture more dynamic range data than using JPEGs.
-Use a tripod. In some places this can be difficult, especially in churches. Alternatively, try and find some kind of ledge, bench, stack of books or take up an awkward stance and take the image handheld.
-Use Aperture Priority mode on your camera. As you will be combining several images into one, you don’t want your Depth of Field (DOF) to change between shots, as the final image will appear all out of focus.
-Keep the ISO setting to the minimum. The higher the ISO setting, the more noise it will generate. I try to keep mine at 100 whenever I can.

Step 2 – Software I use

I use a number of different software applications within my workflow.

Adobe Lightroom – Great piece of software. I categorize and organize all my images using this. Managing them becomes a piece of cake. You can download a trial here.

Photomatix Pro – Allows you to create a HDR image by either exposure blending or tone mapping different exposure images. You can download their free trial here.

As you can see, I am still stuck in the old age and using the older Photomatix version. I’ll have to upgrade it later this week to the latest one.
If you buy Photomatix, you can use the discount coupon code “sandmania”, which gives you 15% off the normal price.

Adobe Photoshop CS2 – This is used to final tune the image and to make any final adjustments. You can download a trial here.

Noiseware – As HDRs tend to create a lot of noise, this is an excellent plugin for Photoshop to help reduce it. You can download a free trial from their Noiseware website. You can use the discount coupon code “sandmania” if you wish, which will give 10% off if you wish.

Fluid Mask – This is an excellent piece of software which makes masking a whole lot easier. It’s pretty straight forward to use once you get the hang of it, even though it may look complicated.
You can get a free trial from their website, and if you wish to buy it, be sure to use the coupon code “sandmania”, for a 10% discount.

Step 2 – Load up images

First thing we will need to do is to load up our HDR images. With Photomatix opened, click HDR from the top menu and select Generate

This will bring up the following screen above screen.
Click on browse to navigate and select the 3 or more bracketed images.

Then press OK.

This next window lets you specify options for creating the HDR image.

The 5 images of Times Square were all taken handheld combined with balancing the camera on a dodgy bollard, as my tripod had been in an unfortunate accident the night before in Times Square. I wouldn’t recommend handheld exposures, but in certain circumstances it can’t be helped.

As they were taken handheld, I check the ‘align source images‘ option to correct for small alignment problems that may have happened.

I keep the other settings the same as they are recommended, so let the program take the standard tone curve. You also have an option here to let the program attempt to reduce ghosting artifacts such as moving objects or ripples. As the images on the Times Square screens are constantly changing, I put a check in the ‘attempt to reduce ghosting artifacts‘ option and change the detection to ‘high‘.

Press ‘OK‘.

Step 3 – Photomatix Merged

Photomatix will now merge the different exposed images into a single HDR image that represents the complete tonal range captured by the photos.

This is way off what the final image will look like, so first off select HDR from the menu and select Tonemapping.

Now this is where it starts to get interesting. This is the Photomatix settings screen that will allow you to tonemap the image. Now there is no general science to these sliders or some kind of magic setting. I have found each image requires something different and this really depends on your personal taste and the effect you want to achieve.

The trick is to adjust the settings to get a nice balanced image. You could be creative or go for something a little more fun. I wanted to depict somewhat of a visual feast with this image, as Times Square does have an array of mind boggling colours and I wanted to try and show that.

I always use the Details Enhancer option when tonemapping an image.

When experimenting with the sliders, these are some general rules which I normally follow;

Strength This controls the strength of local contrast enhancements. I am normally very flexible with this slider and just play around with it until I like what I see

Colour Saturation Controls the saturation of the RGB colour channels. The greater the saturation, the more intense the colour. I opted for a lower value in this instance to create a more darker atmosphere.
You can also create a monochrome image here by taking away saturation and increasing the strength.

Light Smoothing Controls smoothing of light variations throughout the image. I never set this value below 0 as the image starts to get really wacky and cartoony.

Luminosity Adjusts the brightness of the shadows. Moving the slider to the right has the effect of boosting shadow details and brightening the image. Moving it to the left gives a more natural look to the tone mapped image. Again this slider is flexible, but don’t take it over to the right as sometimes halos can start to appear.

Micro-smoothing I have this setting 70% of the time on 0. This setting really brings out the hidden details on the walls and floor that you didn’t know where there.

Once you have the sliders how you want them, press OK and wait for the image to process.

Once it has processed, click on File and then Save

Step 4 – Photoshop Processing.

Open up the newly created HDR image in Photoshop and we are now going to start tidying up the image and get it looking a little better.

The changes I am going to make include;

- Make the colours pop a little by going through LAB colour mode.
- Adjust the colour balance slightly to emphasis certain colours.
- Mask in the original HDR sky using the Fluid Mask plugin.

Step 5 – Lab Color Changes

This is a great technique to use to bring out the colours in an image, in effect making them ‘pop’ out.

Firstly change the mode to Lab Color (Image > Mode > Lab Color).

Then go into the curve settings via – Image > Adjustments > Curves

Within the Lightness channel, I made a slight curve adjustment in the shape of a ‘S’, as shown above.

I then changed the channel to ‘a’ and steepened the slope by moving the top control point to the left and the bottom control point to the right. The same was done to the ‘b’ channel.

Finally change the mode back to RGB.

Step 6 – Photoshop Layers

At this point, I select and copy this Lab Color layer and add it on top of the original HDR layer. I then duplicate the original HDR layer, as if any mistakes I make, I can always revert back to this instead of clicking UNDO each time. So the layers will be stacked up like this;

-Top layer – Lab Color layer
-Middle layer – Original HDR layer duplicate
-Bottom layer – Original HDR layer

Step 7 – Color Balance Processing

With the colors now popping a little, I want to bring out those extra hidden yellows dotted around the image. For this, I need to work on the yellow hues and can adjust these using the color balance settings.

The easiest way to do this is with the top layer selected, go to Image > Adjustments > Color Balance

This will open a new window, and you can play with the sliders to get the look you like. Since I was looking to bring out the yellow hues, I made sure yellow was the highest value, although I played around with the sliders until I got the right look for the image.

You can see difference above looking at the original HDR image and the processed (Lab Color and Color balance) image side by side. The colours are much more prominent and have a ‘pop’ feeling to them.

Step 8 – Fluid Mask Processing

Fluid Mask is a piece of masking software which I stumbled upon by accident one evening. It can either be used as a Photoshop plugin or standalone, and it makes masking a whole lot faster and easier. I’ll be writing a more detailed tutorial on this soon so watch out for that.

You can get a free trial from their website here, and if you wish to buy it, be sure to use the coupon code sandmania for a 10% discount.

As we have processed the top layer a little, the sky now looks kinda weird, so I wanted to mask it out and use the sky from the original HDR. I could have always masked the sky out the normal way in Photoshop using layer masks, but I wanted to give Fluid Mask a whirl.

So again with the top layer selected, we click on the Filters menu and select Vertus > Fluid Mask.

This minimizes Photoshop and Fluid Mask opens up independently.

With Fluid Mask open, we can now start masking and indicate which parts we want to keep and which parts to remove.

Fluid Mask uses three types of masks which are indicated by brush colour:

Green brush – Keep mask: indicates the part of the image to retain.
Red brush – Delete mask: indicates the part of the image to remove.
Blue brush – Blend mask: indicates the border of the cut-out across which Fluid Mask will create a blended transition.

The tool palette has a series of Delete (Red ) and Keep (Green) brushes. Take the centre red brush and dab into an area of background and it will be highlighted with a red mask.

Next use the Green Local brush to define the Keep areas.

Once finished and you are happy with the different masking areas you can complete the masking procedure. Looking at this image though, you can see that I have miscalculated some brush strokes.

Not a problem….by going into the segmented areas view, I was able to correct these mistakes and tidy up the image to what I wanted. This can be really good for images where masking can be difficult or tricky, such as portraits where the background needs to be seperated from the hair etc..

Once completed, Fluid Mask can be closed down and the image will automatically update on that particular Photoshop layer.

The whole Fluid Mask process took about 3 mins.

The final thing I did was a slight contrast adjustment to the middle layer to give a little oomph to the sky. Extra adjustments can also be done here such as cropping, sharpening and/or noise reduction.

I am sure as most of you know; processing HDR images tend to create a lot of noise….and with night time shots this is tenfold. Luckily there are plenty of good noise reduction plug-ins available such as Noise Ninja and Noiseware.

I used the Noiseware Plug-in to reduce some of the noise in the sky.

As you can see looking at the image, there are 3 layers.

- The top layer has gone through the Lab Colour and colour balance settings. The sky has been masked out to reveal the middle layer using Fluid Mask.

- Middle layer is the duplicated HDR layer with a slight contrast adjustment to give a little oomph to the sky. Also Noiseware used to reduce noise.

- Bottom layer is the original HDR processed in Photomatix (this can be deleted at this point)

In some cases you might want to retain the EXIF information on an image. Now Photomatix doesn’t transfer this info across when a HDR image is processed, so what you need to do is open up one of the original exposures and then drag the merged layer across to it.

Make any resizing adjustments if necessary and we now have the final image. All that needs to be done now is to merge the layers and save the image.

Stages

Here you can see how the image has changed from start to finish.

The Final Image

….and the final image….