Reduce Noise in a Chinese Lucky Bamboo

Total Views :4530

Chinese Bamboo Tree

When you take photo in raw format you will realize that the photo has a lot of editing to do. The first thing you will notice is that the picture is extremely noisy and will require editing most of the time. This tutorial I will show you some simple noise reduction tips for a photo taken on a smooth background. Here is the original photo of the detail blown up to show you the noise we will be dealing with.

First thing you want to do is duplicate the original file and edit on the duplicated layer only so that you can also look back at the original picture to see how much has changed.

Next I will apply a Surface Blur from the Filter – Blur menu. Face blur will take the detail using contrast detection into account. Only surface with low contrast will be blurred where Radius is how much to blur and Threshold is how much of the contrast(detail) is preserved.

A lot of the details are were destroyed even though surface blur attempts to preserve them so lets use Select – Color Range to isolate the bamboo area.

Create a mask using the selection by selecting the Mask icon down at the bottom on the layer pallet while the selection is still on, you will have to invert the mask to mask it correctly.

Now you have the background blurred but with the bamboo masked out. Since we still have the original layer right below this layer we see the original bamboo. Lets use Reduce Noise filter from Filter – Noise menu. Strength is how much noise to take out and Preserve Detail is how much contrast/detail to preserve. You can also Sharpen the contrast/detail.

Here is the Before and After Photo of the image. Surface blur attempt to smooth out most of the smooth surface without the bamboo bleeding into the background if you were to use Gaussian blur. Than we isolate the main subject and use reduce noise filter to reduce and preserve detail. reduce noise and filter were to be used on the entire photo, the background would not appear as smooth.

Leave a Reply

Your email address will not be published.