Jump to content
zikronix

Sharpness and noise removal in cameras

Recommended Posts

This might not be relitive to all cameras but it is to the swann/lorex/hik cameras

 

inside the options they have a "noise" removal slider. its set at 50. with values from 0-100. I understand that cranking that to 100 can make the image softer in normal and low light. I also know if its too noisey you cant identify anything.

 

Additionally they have sharpness controls. Which literally adds data to a video to give the appearance of a cleaner sharper picture.

 

Why are their both of these settings, because they are counter intuitive each other. Once would think that you want sharpness at 100 and noise reduction at 100 which from a video standpoint is exactly what you dont want!

 

I guess my question is really to adjust these accordingly one needs to put a static image infront of the camera...especially in low light conditions and adjust accordingly correct?

Edited by Guest

Share this post


Link to post
Share on other sites

Even on high end cameras like Mobotix, it makes for pretty pictures to set sharpness high and you go wow, that's a crisp sharp image until one day there's a crime and you look at the images and you can't make out a persons face because the sharpening artifacts distorted the face when you dive deeper into the image. So I usually crank sharpness and noise reduction down, why, because I don't want the lousy in-camera post processing to mess with the image because I can do a better job in post-processing the image in photoshop to bring out detail.

Share this post


Link to post
Share on other sites

Well sharpness should defintily be cranked to 0. the noise reduction seems to do things different, at least in this camera...maybe its not as extreme as sharpness? idk

 

Sharpness in a video sense is left over from old low res analog bull ****

Share this post


Link to post
Share on other sites

Noise reduction and sharpening are not contrary each other. They are different procedures. There are a lot of noise reduction algorithms including 3D noise reduction. It is not simple blurring, but image blurring is a side effect of NR. General advantage of NR inside camera is traffic reduction in low light conditions.

 

Sharpening is image processing which can correct low resolution of lens in some degree. But noise increasing is a side effect of sharpening.

 

Thus you should try to use sharpening if you use a lens with not enough resolution in good light conditions. But you should not use sharpening with low light noise images. It makes the image more noise.

 

If your camera works in low light condition, using noise reduction can decrease traffic and disk space requirements. In case of limited bandwidth or disk space using noise reduction can increase image quality in whole.

Share this post


Link to post
Share on other sites
Even on high end cameras like Mobotix, .

Stop kidding me Mobotix is high cost, but not high end cameras

Share this post


Link to post
Share on other sites

Hey, Stanislav is back! Always good to see posts from an expert!

 

To expand a bit, sharpening is similar to edge detection, while noise reduction is closer to blurring or averaging. Most Chinese cameras oversharpen as the default these days, and some don't allow it to be adjusted.

 

This is a big problem in low light, as it brings up a bunch of rectangular artifacts around the normal camera noise, increasing bandwidth and image size. It also can trigger motion detection, which is a problem on one of my Dahuas. Around dusk, before it switches to night vision, it detects the oversharpened noise as motion for 15-20 minutes.

Share this post


Link to post
Share on other sites
Hey, Stanislav is back! Always good to see posts from an expert!

 

To expand a bit, sharpening is similar to edge detection, while noise reduction is closer to blurring or averaging. Most Chinese cameras oversharpen as the default these days, and some don't allow it to be adjusted.

 

This is a big problem in low light, as it brings up a bunch of rectangular artifacts around the normal camera noise, increasing bandwidth and image size. It also can trigger motion detection, which is a problem on one of my Dahuas. Around dusk, before it switches to night vision, it detects the oversharpened noise as motion for 15-20 minutes.

 

This is good to know i will look at my settings again for the cameras...and may do some fine tuning.

Share this post


Link to post
Share on other sites

Thank you, MaxIcon

Here are several images for better understanding sharpening. Level of one line from the images is shown as a curve.

 

Normal image with a little noise. No blurring and no sharpening. Pay attention on vertical curve without overshots at the line drop point.

 

219524_1.png

 

Blurring. Line at the drop point is not vertical. Any lens has limited resolution. If the lens resolution is unsufficient, it adds blurring. But lens blurring doesn't reduce noise because of the noise is added on the image sensor.

 

219524_2.png

 

Little oversharpening. Line at the drop point is strictly vertical as on the normal image. But overshots appear. There is optimal sharpening level which can correct lens blur and approximate normal image. It allows manufacturers to use cheaper lenses.

Images with little oversharpening look better. But sharpening as any other image processing can't restore detail lost because of limited lens resolution, it can make images which will look like a normal image but this corrected image will not contain all details.

Pay attention also on increased noise level in comparison with normal image. It is unnoticeable on image but it is visible on line curve.

 

219524_3.png

 

Noisy image without sharpening.

 

219524_4.png

 

The same noisy image with the same little oversharpening. With noisy image oversharpening corrupts it. To prevent this effect, many cameras turn off sharpening automatically when the image becomes noisy.

 

219524_5.png

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×