enter image description here

I think this image taken at night or low Quality capture..

How I get clear effect to this image?

  • 2
    Simple answer: you probably can't do very much with it. You may be able to make it a bit better, but don't expect miracles. – Philip Kendall Apr 9 '14 at 10:50
  • @PhilipKendall We have Photoshop, But most customer came like this picture clarity photo. We want a perfect Solution. we mostly use Remove Grain for this like picture But it gives blur like effect. – sathya Apr 9 '14 at 10:52
  • 2
    You've pretty much got it. Noise reduction == blur. Nothing you can do about that. – Philip Kendall Apr 9 '14 at 10:59
  • See also: photo.stackexchange.com/q/49386/21 – Rowland Shaw Apr 9 '14 at 13:06
  • @sathya you can't really do much, indeed! Photoshop unfortunately cannot do miracles, like most claim. If the image quality is bad, you can d very little, compromising the sharpness to make up the grain. – Morpho Apr 10 '14 at 8:57

This is noise and the rule with photos is the same as most things, crap in, crap out. You can't magically generate information for an image that isn't present in the first place. You can reduce the quality, either through reducing resolution or averaging pixels (which effectively reduces resolution). This averaging tends to make random noise go down while actual signal remains constant, but you will never reduce the level of noise on the actual original resolution of the original picture without blur, simply because there is no way to tell what is noise and what is signal.

The only way to get rid of noise (after the fact anyway) is average it out and that means some signal will be averaged out as well. Some software may work off a model of the noise from a particular camera to try and do a slightly better job, but it's still going to be a big reduction in overall sharpness in exchange for the reduction of noise.

|improve this answer|||||
  • Some software might also work out a model for the content of the image (like a face). But I have no idea if this actually exists. – ziggystar Apr 11 '14 at 8:44

Whether this is better or not is highly subjective. It definitelly looks blurred compared to original, but we also have to be aware that sharp noise on top of a slightly blurred image makes the latter sharper.

This is what I've done:

  1. Copied image layer and applied High pass (1.0) on it
  2. Inverted the grey layer and then set it to overlay
  3. Copied the same layer and adjusted its opacity to make skin look as smooth as possible.
  4. Merged layers and applied Reduce noise filer on it.
  5. Created a new High pass (2.0) layer and Overlay-ed it to sharpen the image

The end result

We also have to be aware that this particular image has quite a bit of JPEG artifact noise which makes it even more difficult to reduce it.

enter image description here

Why the negative high pass filters?

Negative high pass overlays equalized sharp noise contrast on the image which made the skin tone and any other supposedly equally toned surface more... equal...

|improve this answer|||||

You do not say how you obtained this image at this pixel size or how it was generated. If you know you should say as it has a marked effect on the answer. The answer is "no, because..." but the "because" varies with the above answer.

This is NOT noise in the sense that it is usually meant. As presented the 'noisiness' of the image has been caused by massive lossy-compression artefacts. The image is comprised of a number of squares where detail is replaced with various pattern combinations.
As saved in Imgur (ie on this site) the image is 400 pixels tall BUT there are only about 55 vertical blocks. So detail in about 50 pixels total is being replaced by a single block with a pattern in it.

The compression effect MAY have been caused as the image was uploaded to the website BUT if so then it results in an image that does not convey the proper information.

These affects may be caused by

Excessive user compression of files either iun camera (unlikely to this extent) or on saving .

Cropping of a larger photo and then blowing it up (seems likely here),

Compression during processing or upload

The best way to get "clear effect" is to initially take the photo properly at required size.

If you REA::Y want a better result (eg subject has died and this is the best photo you have) then manual selective editing of the image at large blow up size may help. You are effectively recreating the image. This allows outlines and visual identifiers to be maintained while trying to deal with the very ugly compression artefacts.

|improve this answer|||||

Well, in general, if the noise distribution was completely unknown, then you would not be able to do too much.

Fortunately most of the noises of high ISO are of known types (see this link), and their stochastic process or random distribution is known, and this offers some ways to remove some of the noise.

Now, this is pretty much heavy math that you probably do not want to get involved in... :-)

From an user's perspective, your approach should be using Lightroom's noise and color noise reduction (in "Develop"). They do usually a good job, if you use them on the original, unaltered image, especially on RAW images.

Or Google for "noise reduction" plugins, and try some, e.g. Topaz's DeNoise.

You can do better than just resizing (an averaging process) or manually averaging pixel colors, as averaging in DSP is a low-pass filtering process, so on an image you do a spatial low-pass filtering on colors. However, the RAW image contains data per the sensor channels, and the color sensor (pixel) contributions to noise is different. So if you average, you are low-pass filtering colors but also mix the color errors with neighbouring pixels, if you are working on a not-RAW (cooked) image.

|improve this answer|||||

Of course you cant convert it to a quality image. All you can do is to retouch the image with best possible colors. It might take a while, but definitely looks better than the existing one.

|improve this answer|||||
  • To improve you answer can you please add some detail about the process. – damned truths Apr 11 '14 at 8:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.