| Author | Thread |
|
|
04/02/2003 04:05:46 PM · #1 |
Sometimes with lower end cameras you end up with such noisy pictures that you think - well, this is completely unusable. Or you spend tons of time trying to get rid of the noise. I have just posted a 5 minute tutorial that uses Lab Color Mode in Photoshop as an alternative to cleaning the noise.
Let me know what you think.
Galina
Message edited by author 2003-04-02 16:06:03.
|
|
|
|
04/02/2003 04:13:33 PM · #2 |
| Thanks for sharing. I'll have to try this. |
|
|
|
04/02/2003 04:18:08 PM · #3 |
| very informative, especially since I have the same low-end camera as you. Any suggestions for reducing noise while keeping it in color (esp. when more than blue is showing a lot of noise -- even though its always the blue, always the blue) |
|
|
|
04/02/2003 04:38:43 PM · #4 |
| Trying this, but running into some practical problems: It seems that I'm unable to modify the channels seperately. What am I doing wrong, here? |
|
|
|
04/02/2003 04:41:34 PM · #5 |
Marco, at which step? The Color Mixer?
|
|
|
|
04/02/2003 04:51:01 PM · #6 |
Ah, It's OK now! I tried the gaussian on the noisiest channel, and the effect is nice. Thanks for the tip!
Is it possible to give some more info about the part where you use the 'find edges' filter? That's yet not really clear to me.
Thanks alot |
|
|
|
04/02/2003 05:19:39 PM · #7 |
The Find Edges method is a basis for several free and commercial Photoshop actions (you can write your own!). The success, as usual, largely depends on the particular image you are dealing with. The principle behind it is that noise is mostly visible in the large areas, so by running a Find Edges filter you isolate edges that you want to keep sharp and flat areas which you want to blur. Here is the method itself:
Step 1. Copy your image into the new channel.
Step 2. On this channel run the Find Edges filter. The edges will turn black, and flat areas will turn lightish.
Step 3. The reason you use this filter is because this channel is going to be loaded into our original image as a mask. So you want more contrast between edges and flat areas. To achieve this you need to use Levels or Curves to set the black and white points - manually or with 'point pickers' - so that your edges are really dark and the flat areas are very light (white? again it's a personal preference).
Step 4. Now run the Gaussian Blur filter on this channel so that the transitions in the resulting mask are smother.
Step 5. The Gaussian Blur filter might mess up your contrast, so you might need to finetune it with Levels or Curves again.
At this point there two different ways of proceeding. Some people prefer to create a snapshop of the original file at the beginning, then they load the alpha channel as a selection, and afterwards use the History brush to restore some detail. Some people prefer to load a selection as a mask and use the Brush tool afterwards for retouching. Anyway...
Step 6. Load a selection/or mask from your channel. Check what areas are selected and if you need to invert your selection(weird things happen sometimes!)
Step 7. Run the Gaussian Blur on the selection/non masked area. Be fairly conservative - the amount, radius and threshold settings depend on how much blur you need to get rid of the noise.
Step 8. After the Gaussian Blur you may discover that you lost some detail and texture. Either run Fade option from the Edit menu and use History Brush (if you worked on the selection) or, if you worked with the mask, use a soft brush at 30-40% strength to paint back original detail.
That's about it.
Message edited by author 2003-04-03 04:27:09. |
|
|
|
04/04/2003 02:34:50 AM · #8 |
| Thanks a lot Galina! Brings a bit clarity in the Photoshop jungle :) |
|
Home -
Challenges -
Community -
League -
Photos -
Cameras -
Lenses -
Learn -
Help -
Terms of Use -
Privacy -
Top ^
DPChallenge, and website content and design, Copyright © 2001-2025 Challenging Technologies, LLC.
All digital photo copyrights belong to the photographers and may not be used without permission.
Current Server Time: 08/27/2025 04:46:53 AM EDT.
