I currently use Photoshop's merge to HDR pro but i don't always find that method very good. I wonder if there is a way to merge photos by using layers instead?
-
1\$\begingroup\$ Can you clarify what you don't find 'good' about it, so we may be able to offer suggestions to improve that aspect... \$\endgroup\$– Rory AlsopCommented Mar 21, 2014 at 16:24
-
\$\begingroup\$ See also photo.stackexchange.com/q/3973/15871 \$\endgroup\$– Michael CCommented Mar 23, 2014 at 2:22
-
1\$\begingroup\$ Here is an answer regarding manually merging exposures using layers: photo.stackexchange.com/questions/3394/… \$\endgroup\$– Matt GrumCommented Mar 23, 2014 at 20:04
2 Answers
There is a method that can yield some good results given good enough original photos. I don't have the original article on hand, but I can describe the method, but it's up to you to find the particular details that you need to get your desired results.
Let's start by merging just two images. Load them in a single Photoshop document's layers. It doesn't really matter which image you put on top, what really matters is that both images line up perfectly.
Next step is to add a layer transparency mask to the layer on top. Go to Layer > New Mask > Transparency Mask. (Sorry if this is not the exact name, My PS is in Spanish).
In case you don't know how transparency masks work: they are pretty simple, the white areas of the mask will show the layer, the black areas will hide layer contents, showing whatever is below. Gray values will partially show the layer, making it semi-transparent. The darkest the mask, the more transparent the layer becomes.
Now, If you have two layers at different exposures, one has shadow detail and the other has highlight detail. The core of the method is to make transparent the part of the image on top layer that doesn't have detail to let it show the corresponding detail on the bottom layer. You could paint the transparency mask manually, but that would work only on simple images that do not have intricate details, otherwise it would be painfully time consuming. The trick is to use one of the images to create a grayscale mask.
Just select one of the layers by clicking the corresponding icon on the layer palette. Select all and Edit > copy (Ctrl+C on Windows). Then Alt+click on the layer mask icon on layers palette. (When you created the Mask on step 2, a new icon appeared besides the layer's icon) the default transparency mask is all white, showing the complete layer. When you Alt+click the icon you will see the actual mask, thus your image will look white. Now paste the image into the mask (Ctrl+V). This will create a grayscale mask with your image as a pattern.
Remember that the mask should have dark areas in the parts of the top image that you need to hide, the ones that have no detail. If your top image has highlight detail then this step is not needed. If your top image has shadow detail then you need to make a negative of the mask. Simply hit Ctrl+I to invert the image while you are in mask edit mode (when you Alt+clicked the mask's icon).
Now click the layer icon to see the results. They will be disappointing at this time, but keep calm. Now you make several adjustments to fit your creative needs. One of main adjustments is to alter contrast, levels and curves of the mask. Simple click the mask icon and use the regular image adjustment dialogs. By altering these you change the tone of the gray areas of the mask, thus changing how much showing/hiding there is in each spot.
When you approximate the desired results, make fine adjustments to each source image. You can fine-tune the mask with the brush if needed. Finally you need to apply some degree of Gaussian blur to the mask, to create a softer transition between images.
That's the general method. Once you get the quirks of it, you'll learn your own tricks and shortcuts. It is not the same for all the images, some require certain adjustments and other nay require adjust in a different direction. The same idea is valid to use 3 o more images, you just need to work out which image areas are to be shown.
In regard to which image to use for the mask, you'll need to try both. Depending on the images one may work better than the other. Practice and be patient.
As the previous answer demonstrates, you can merge the different exposures using layers.
But it's a lot of work. I'd suggest trying enfuse, which I've been quite happy with. It produces realistic looking output, not the dreaded "HDR look", and does so with pretty much no work on your part.
It runs via the command line. However, if you have Lightroom there's a plugin called LR-enfuse (this is how I use enfuse. There is also a GUI available for enfuse if you don't have Lightroom.