One of my greatest pleasures is sharing knowledge with other people. Please enjoy this HDR walkthrough!
This walkthrough assumes that you are taking 3 exposures using auto bracketing of -2/0/+2 EV and that you will be starting the process with RAW files. I have attached the complete Photoshop file and three original RAW files used in this walkthrough here. Check out the finished image from this walkthrough.
Before you can combine your bracketed photos into a HDR image, you need to white balance each of your exposures. There are many ways to white balance images, but one of the easiest and quickest is to use the white balance tool in Camera Raw. You can see in the following screenshot the three exposures loaded as Digital Negative (.dng) files. Click on the white balance tool icon which looks like an eye dropper in the tool palette. Then choose the normal (0 EV) exposed photo and click on something in the image that is neutral in color. You may need to click a few times until the image looks right to you. In the following example, I clicked on the portion of the building that was grey as shown. You can see that the number 5100 was selected in the white balance display under the histogram in Camera Raw.
Once you have a proper white balance setting, you want to apply the same white balance setting to all three photos. If you do not do this, and especially if the white balances of your three exposures vary a great deal, you can see terrible colors in your HDR image. An easy way to apply your white balance setting to the other two photos is to click on the “Select All” button and then the “Synchronize” button as shown below. After you click on “Synchronize”, you will be presented with a dialog that will ask which settings to synchronize. In this case we will only synchronize the white balance settings as shown below. Click the “OK” button after you have checked white balance.
After you have synchronized the white balance settings to the other files, the next step is to export all three files to .jpg format. Click on the “Save Images…” button in the bottom left corner of Camera Raw. You will be presented with the dialog shown below. Make sure that the format is “JPEG” and the quality is set to 10 or “Maximum” as shown below. Click the “Ok” button to export.
This section is optional if you used a tripod to capture your three exposures. In some cases even with a tripod you might want to try it. In this example, the images were captured with a tripod but I noticed that a small amount of shift has occurred in the exposures. When I captured these exposures, I did not use a remote trigger and instead pressed the button on my camera to take each photo. This caused a small amount of shift. The best way I have found to auto align images is to use Photoshop. The fastest way to do this is to select your photos in Bridge, then under the “Tools” menu, choose “Photoshop” and then “Load files into Photoshop layers” as shown below.
You will see Photoshop loading the files in and creating new layers. When Photoshop is finished you will see your three exposures and three layers. Click on the middle exposure first, then hold down the “shift” key on click on the overexposed (+2 EV) and the underexposed (-2 EV) layers to select them all. The layer that you select first is the layer that Photoshop will use to align the other layers with. Once you have all layers selected, go to the “Edit” menu and select “Auto-Align Layers…”. You will be presented with a dialog that will ask you how you want the layers aligned. In this case we will use “Auto”.
Here is a screenshot of the “Auto-Align Layers” dialog.
Now that we have our layers aligned we need to save each layer out as a separate .jpg file so that we can load them into Photomatix for tonemapping. Under the “File” menu chose “Scripts” and then “Export Layers to Files…”.
Click on the “Run” button.
Photoshop uses a naming convention that I do not like when it creates the files. I use bridge to rename the files as shown below.
Create the HDR file
Now that we have our three jpeg images white balanced and straightened, the next step is to load them into Photomatix. I am using the new beta version of Photomatix 4.0 because it is very solid and has extra noise reduction features that are useful when creating HDR images. The HDR process creates a lot of noise. The amount of noise is compounded when the photos were captured at higher ISO settings. I captured the three example images at ISO 400 and you will see there is quite a bit of noise because of it. In general, try to use as low an ISO setting as possible to lower noise in your HDRs. If you are lucky enough to have a camera that will take more than three auto bracketed images, you can lower your noise as well. There are also ways to take more than three exposures when using a camera such as the Nikon D5000 that I use. I will not go into that here.
There are many ways to open the files. I selected the photos in Bridge and dragged them to the Photomatix window. After you load your jpeg images into Photomatix, you will be presented with the following dialog.
Specify the EV spacing at “2” in the drop down menu. Photomatix will then know to use the -2/0/+2 spacing that your images are captured with. Click the “Ok” button and after a minute or two you will see a really bad looking image. This is your monitor’s attempt to represent the HDR image on your screen. Your monitor does not have the dynamic range to do this, so it looks terrible.
Fortunately, we can fix this with a process called “Tone Mapping”. Before I perform the tone mapping, I usually save the hdr file at this point. There are a number of different formats to chose from. I usually chose the .hdr extension. The reason you save the .hdr file is that many times you will want to do more than one tone mapping. I usually do at least two with the .hdr file. In this example, we are going to tone map two versions. One for the main image and the other for the sky.
Now the fun starts! There are many combinations of the sliders in Photomatix. I recommend each time that you create an image, to start with the sliders in the middle and drag them first to the left and then to the right. You have to release the slider to apply the settings you have chosen. Keep dragging and releasing the sliders left and right until you get to a look that you like. Your style in HDR will come over time. For this example, you will can see the setting I used for the first image below.
For the second image, I concentrated only on what the sky looked like. The final product isn’t that different from my first image, but in many cases it will be quite a bit different.
For this example, we have one more image that we need to tone map. Because the cars are moving on the road, you may have noticed that there appear to be ghost images of the cars. I want a nice clean car on the road, so I am going to take the normal (0) exposure and tonemap it by itself. In order to do that, I need to use the RAW file rather than the jpeg file. The RAW file has much more information inside it. In fact, some photographers like to tone map their single image exposures to add some HDR flavor. Go to Bridge and drag the file “DSC_2740.NEF” to the Photomatix window. If you are using Photomatix 4.0 like I am in this tutorial, chose 100% noise reduction and the option of reducing chromatic aberrations. Leave the white balance setting alone in this case. Once you are ready to apply tone mapping, chose the following settings.
Feel free to create as many tone mapped images as you want. The next step of the process will be to combine all the images into one Photoshop image.
Putting it all together
Now it is time to put all the images into Photoshop. Select each of the tone mapped images you created and the straightened original jpegs in Bridge and load them into Photoshop layers as we did above. You may also want to use the “Auto Align” feature we used in the “Alignment” section above as well. Remember that the layer you select first will be used as the basis for the alignment. I have done all of that for you in the attached file at the top of this walkthrough. Below is a screenshot of what it looks like.
You can see that on the very bottom layer, I kept the original jpgs in case I wanted to mask in pieces of them. In this example, I did not wish to do that. You could of course delete the originals from your Photoshop file if you are not going to use them. I kept them here for completeness and so that you can experiment with masking in bits of them if you wish.
I started with the first tone mapped image in the first layer. I then added the sky tone mapped image on top of it with masking that brought in just the portions of the sky I was interested in. When using a mask, the white part keeps what is in the layer and the black part erases what is in the layer. In this example, you can see that I have painted white only in the sky portion of the image.
I created a group for the car. You can open up the group by clicking on the triangle next to the folder icon. You will see that I masked in a section of the road where the car is. I then duplicated that layer and applied a very generous denoise filter on it. I then masked in just the portions of the road that I wanted to denoise. I didn’t want to denoise the car, because I wanted all of the sharpness to remain.
Once I was happy with the portions of the sky that I had masked in and happy with the car, I merged all visible layers to the top layer you can do that in Photoshop by holding down the “shift/control/alt/e” keys on Windows machines and the “shift/control/option/e” keys on a Macintosh computer. I call these layers “Merge”. Then I add a new layer called “Cleanup”. This is where I do any cloning or healing that I may need. In this case, there were some pretty big dust spots. In the HDR process, dust spots are usually magnified. After the cleanup, I merged again and applied the “Topaz Adjust – Exposure Correction” filter.
I then made two copies of the “Merge and Topaz Exposure Correction” layer and performed a Denoise and a Topaz Photopop with the appropriate masks to isolate the sky and the rest of the image. Then I merged again and applied a unsharp mask filter to the merged result. You can see from the layer masks that I created a mask that separates the skyline from the sky. I used this mask or the inverse of the mask on each of these layers. For a Denoise filter, I used Imagenomic Noiseware Professional. Topaz Adjust was used for the Photopop filter. I used the built in Photoshop unsharp mask filter. For the unsharp mask settings, I used Amount of 99%, Radius of 3.2 pixels, and Threshold of 3 levels.
For the final processing, I added a curves layer to add contrast and applied it only to the luminosity layer. I then desaturated the car and the sign because they were too red. I added another adjustment layer to desaturate the yellow channel in the grass in the foreground. For one final adjustment, I upped the saturation in the sky, the buildings and portions of the grass.