Photoshop Tutorial : Low and High Frequency Editing
Low and High Frequency editing has been around for a long time in Photoshop. It can be used for precise sharpening of images, and skin smoothing for portraits.
So what is the meaning of Low and High Frequency editing. You are basically splitting your image into two layers. One layer holds the texture and fine detail of the image, or high frequency information – and One layer holds the tones and colour, or low frequency information. The High Frequency layer is Sharp, the Low Frequency layer is Blurred and you use the sliders on both layers to obtain a very precise degree of editing.
The numbers used on the practice image are going to be relevant to these image only.
The first image is brought to us by nancynator over at the Stock Exchange Site I have adjusted the size of the image for the purpose of this tutorial – 800px X 800px @72dpi
Click through for larger images
The second image is brought to us by boeinglove over at the Stock Exchange Site I have adjusted the size of the image for the purpose of this tutorial – 533px X 800px @72dpi
Click through for larger image
Open the car image and duplicate the b/ground twice. You should now have 3 layers. Name the topmost layer High and the middle layer Low. Turn off the visibility of the top layer. Go to Filter>Blur>Gaussian Blur and use a Radius of 10px. Remember that the purpose of the Low Frequency layer is to retain the tones and colours, but lose the detail. You will have to experiment to see what numbers you like the best on other images. Click OK to accept, and then turn back on the top layer and click on it to make sure you are working on that layer.
Go to Image>Apply Image and when the dialogue box opens Low from the Layer drop down box. Change the blending to Subtract and then Scale 2 and Offset 128. For those that use the High Pass Filter to Sharpen their images, the result should remind you that. Click OK to accept and then change the blend mode of the layer to Linear Light. If you have done everything correctly, you should end up with an image exactly the same as you set off with 🙂
Click on the top layer and then shift + click on the middle layer and Ctrl + G to group the 2 layers. Add a layer mask to the group by clicking on the add layer mask icon at the bottom of the layers palette. Click on the little arrow to open the group so you can edit the layers. Click on the Low Frequency layer and take the opacity of the layer down to around 50% – click on the High frequency layer and adjust it to around 75% …. or whatever numbers you want. The idea is to show you another way to sharpen images, so choose your own numbers on your images. If you want to target the sharpening to a specific area of the image, click on the layer mask you added to the Group, Ctrl + i to invert the mask then use a soft edged white brush to brush the effect back in.
What you’ll find using this method is that you don’t get that overly contrasty look between the pixels where the transition is being exaggerated. You know this better as a Halo.
In the image below, I have reduced the Low frequency layer to 50% and below that image is a comparison image.
Try this on your images where you can zoom in to see that there is little haloing produced from this method of sharpening. Remember that you can play with the sliders on both the Low and High layers to get the effect you want. Also remember that you can Blur more or less on the Low frequency layer to achieve the effect you want.
Open the portrait image and duplicate the layer twice. You should now have 3 layers. Name the top layer High and the middle layer Low, then turn off the top most layer. Go to Filter>Blur>Surface Blur (if you have an earlier version of Photoshop, use Median Blur) and you can punch in the same numbers as I have in the image below. Remember that the numbers are relevant to this image, you will have to play around with them for other images
Make the top most layer active and then go to Image>Apply Image and follow the same procedure as you did earlier. Change the blend mode the layer to Linear Light and then group the 2 layers together as you did earlier and add a layer mask. Open the group and click on the High layer to make sure it is active and reduce the layer opacity to around 50%. Alt + click on the eyeball of the b/ground layer to see the changes so far – or click the eyeball of the Group Layer on and off. What you’ll notice is that a few other things beside the girls skin have been blurred – paint them back in on the mask with a soft edged brush set to black. See image of mask below
Click on the Low layer and grab a soft edged brush. Reduce the opacity of the brush to around 40% in the top toolbar. You might be able to make this number a bit higher, but becareful that you don’t go too high. You are going to fix some of the blemishes, although you may have done quite a bit of cleaning up the image before you started the whole process, this is to show you another way. Remember we spoke about what a Low frequency layer was at the start of the tutorial – It holds the tones and the colours and loses the detail. Editing can be done with a paint brush on the Low frequency layer by picking up a colour close by and then painting over the blemish (whatever) that needs to be removed. This isn’t generally for large faults and it is probably best to have the skin set at how you want it before you start the process. You should also take a snapshot of the file before you start so you have a point to get to in case it all goes horribly wrong.
OK, Take a snapshot, grab a soft edged brush and lower the opacity to around 40%, zoom in on the nose area of the image (300%) and then Alt + click on the image next to a freckle. They eyedropper tool will make that colour the f/ground colour and all you have to do is paint over the freckle. You can do that a couple of times and sample the colour from a different area each time. You can go over the entire face in this manner, not removing everything to make the poor girl into a Barbie Doll (Plastic).
The most important part of this technique is that there is still quite a lot of texture left in the skin. In the comparison image below you can see my start and final image. I didn’t do any healing before applying this method. The final image retains some of the imperfections of the girl that make her uniquely who she is, whilst still covering up a couple of them and giving her a more pleasing (god, that sounds horrible) appearance. There is so much more that could be done to this image to improve it, but is a job for you to do 🙂