Photo-Merge - an HDR Technique.

There are times we have been out shooting and realised that we have left our filters at home, or that the filter adapter ring is the wrong one for the lens we are using.

Many of you will already be familiar with software packages like Photomatix Pro, which specialise in the merging of bracketed exposures, in order to give us an extended dynamic range over that of the cameras sensor.

Whilst this technique has its uses, the results generally don't look particularly natural as getting the tone mapping right is actually very difficult.

There is however an alternative that in many ways is far easier to capture (normally just two frames required) and equally the merge technique in CS3 is actually very easy (once you know how!)

Capture

This is the important part and you must get it right as otherwise the 2nd part will not work. Normally we would expose as far to the right as possible in order to maximise tonal range. However as we have already discussed, there are times when the dynamic range of our scene, exceeds the dynamic range of our sensor. If we do as above and do not blow the highlights, then we may find that our shadows are extremely dark to the point that recovery of tonal information means that we will also introduce noise.

To get round this we have two options, if possible, we can introduce an ND grad filter to allow us to reduce the exposure of the brightest part of our composition e.g. the sky, and then as a consequence, increase our exposure value, which in turn lightens our shadow or darker areas - nice and easy and all in one frame.

As an alternative, we can do the following:

1) Capture an image, exposing for the brightest part of the frame, ensuring that we do not blow out the highlight information.

2) Capture a 2nd image, ensuring that we do not clip the shadows (highlights will of course be blown as a consequence).

Our respective capture histograms could look something like this:

Image 1 above - no clipping in the highlights.

Image 2 above - no clipping in the shadows.

NOTE: You must use a tripod for the capture to avoid registration issues at the Post Production stage.

Post Production

With our images safely captured, we can open them up in our RAW software (in this case we are using Lightroom 2.1) and have a proper look at what we have captured. It is important at this stage to carry out all our normal basic adjustments ensuring that our capture WB on both frames is exactly the same. 

Once we are satisfied, then simply open both images into CS3. From Lightroom, go to the top menu and go photo>edit in> then choose CS3 (or whatever version of Photoshop you are using).

Both images will now be open in CS3 and we next want to overlay one on top of the other.  Select the move tool from the top of the LH tool box and then drag the lighter of the two images (while holding down the shift key)onto the darker one. Holding down the shift key will auto align your two images. If they are not aligned exactly, then align them manually.

The next part is crucial to getting a natural result. Select the lasso tool from the LH tool box and then draw just below the horizon (see video example below) and then all the way round to include the entire sky area.

Now we must make a smooth blend so that when we delete this part of the lighter image, the darker image below will blend smoothly, with the blend line being almost be invisible. To do this in CS2 you would choose select>feather> then choose the feathering amount (250 is normally ok). In CS3, choose refine edge which brings up the new blending options. (see the screen grab below for the recommended settings). The refine edge option can be found in the 'select' drop down menu at the top of the page, or just below the drop down menu on the tool bar.

Now press 'Q' for quick mask to view the blended area. Now press 'Q' again to exit quick mask. To delete the top layer of your image, now press the delete button on your keyboard. This will delete the selected area of your brighter image, allowing the sky to show through underneath.

Finally, before making any further adjustments, flatten the image by going layers>flatten image.

Your end result will look something like this.