in camera sharpening

Does in camera sharpening generate additional noise? Also would reducing any color channels or any other settings help to reduce noise as well?

The best way to reduce noise is to not underexpose.

Any sharpening makes the existing noise more visible; and some noise always exists, even at ISO 100, especially in the shadows, although at low ISOs it is hardly visible. Whether the sharpening makes the noise visible and/or objectionable depends on how much noise there is without the sharpening and (more importantly) the standards of the viewer. (Personally, I like a bit of noise and often add a little film-like grain to images.) OTOH, noise reduction reduces sharpness - the two are opposite sides of the coin and finding the best balance between them is difficult.

Most of the noise is generally in the blue channel because increasing brightness digitally increases noise and since the sensor is least sensitive to blue, that channel is relatively underexposed and this is corrected by giving it a big boost during the application of WB. This is particularly true when the illumination is tungsten lighting which already is deficient in the blue end of the spectrum and the boost has to be correspondingly greater. In sunlight, where there is a lot of blue and less red, the WB will push the red channel a lot also, so the red channel will become noisier. The green channel is always cleanest. If you have the software to do NR to the channels separately, it can be worthwhile.

However, since you ask about in-camera settings there is nothing you can do to "reduce color channels". The best you can do is to experiment with various combinations of sharpening and NR to find what you like best. If I were a jpg shooter (I am not) I would set sharpening to the minimum possible and do my sharpening in post-processing where I could use sharpening methods, like through an edge mask, that increase noise less and NR methods, like through a surface mask, that reduce sharpness less.

Sharpening involves the contrast between the pixels as well as a the area around the pixel contrast area referred to as the sharpening halo. With in-camera sharpening you only have control over the contrast portion. You can't change the size of the halos.

In-camera noise reduction applies basically a blur across the image and doesn't separate the image into specific channels to apply the noise reduction where it is needed.

So when you use in-camera sharpening you have no way to reduce the noise before sharpening and the noise gets sharpened along with everything else unless you blur the image which defeats the purpose of sharpening.

The best way to do digital is to first apply noise reduction specifically to the channels where it is most evident. Then you can sharpen with control over the halo size and the contrast level. Even better, you can apply the sharpening through the unsharp mask which will apply the sharpening even more selectively to just the edges. Also, dedicated noise reduction software like "noiseware" does an even better job of eliminating noise and leaving the detail in place by using advanced algorithms which would be far too cumbersome to build into a camera.

You can automate this well with lightroom or photoshop so you aren't spending a lot of time on it.

calypsob wrote in post #15176687
Does in camera sharpening generate additional noise? Also would reducing any color channels or any other settings help to reduce noise as well?

Actual pictures from a Canon 60D that has enhanced sharpening applied in camera using Picture Styles.

Byte size: ZERO | Content warning: NOT AN IMAGE

Byte size: ZERO | Content warning: NOT AN IMAGE

Byte size: ZERO | Content warning: NOT AN IMAGE

Geez you guys hit the nail on the head. I need to ask questions in general photography more often. The reason I asked is because I have been doing widefield shots of the milky way and I do median stacks of multiple exposures to remove iso 1600 noise. Sharpening the night sky has always produced strange results for me, normally I do highpass sharpening but I am usually unsatisfied with the results when dealing with stars because it creates extremely sensitive and funky halos. I have been tweaking my WB, and color settings alot lately trying to dial in a good custom setup for stacking and I think I am going to try turning the sharpness down all the way after some of the things I just learned in this thread. It seems that median stacking pre-sharpened images with the objective of removing noise is counterintuitive. TZALMAN, you are right about the blue channel, it is always the noisiest channel when I edit RGB separately in CS5, red is a trouble maker as well.

calypsob wrote in post #15181268
Geez you guys hit the nail on the head. I need to ask questions in general photography more often. The reason I asked is because I have been doing widefield shots of the milky way and I do median stacks of multiple exposures to remove iso 1600 noise. Sharpening the night sky has always produced strange results for me, normally I do highpass sharpening but I am usually unsatisfied with the results when dealing with stars because it creates extremely sensitive and funky halos. I have been tweaking my WB, and color settings alot lately trying to dial in a good custom setup for stacking and I think I am going to try turning the sharpness down all the way after some of the things I just learned in this thread. It seems that median stacking pre-sharpened images with the objective of removing noise is counterintuitive. TZALMAN, you are right about the blue channel, it is always the noisiest channel when I edit RGB separately in CS5, red is a trouble maker as well.

why not use RAW, many stacking programs like Helicon Focus can use RAW files

watt100 wrote in post #15189896
why not use RAW, many stacking programs like Helicon Focus can use RAW files

I only use RAW

calypsob wrote in post #15210590
I only use RAW

OK, so why worry about in-camera sharpening

watt100 wrote in post #15210916
OK, so why worry about in-camera sharpening

Well I have been trying to get the least amount of noise possible so that when I make big prints like 20x30's there is as little grain as possible. When you shoot the milky way it gets tricky to decipher noise from the actual stars so I am just making as many precations as possible. RAW does offer alot of flexibility but it has its limits which is why I asked about in camera sharpening causing noise. At this point I have made a picture style with the sharpening set to zero when I plan to stack images. Now I just have to wait for all these clouds from Sandy to dissipate so I can test out my adjustment.

In-camera sharpening applies only to JPEGs produced by the camera. If you're working with the RAW files then the in-camera sharpening is irrelevant.

I'm particularly curious about this sentence:

In the world of image processing, there's nothing a camera can do that you can't do in post.

As previously pointed out, when working with a RAW file, you can do selective noise reduction first, then selective sharpening, with a level of flexibility that no camera preset can approach.

Another significant factor that hasn't been pointed out yet in this thread: Downsizing is an extremely effective noise reduction tool. If you have resolution to spare (more than maybe 240 pixels per inch at your intended print size), then you can apply noise reduction, then downsize, then sharpen. The order in which you perform those actions is of paramount importance.

Curtis N wrote in post #15216613
In-camera sharpening applies only to JPEGs produced by the camera. If you're working with the RAW files then the in-camera sharpening is irrelevant.

I'm particularly curious about this sentence:In the world of image processing, there's nothing a camera can do that you can't do in post.

As previously pointed out, when working with a RAW file, you can do selective noise reduction first, then selective sharpening, with a level of flexibility that no camera preset can approach.

Another significant factor that hasn't been pointed out yet in this thread: Downsizing is an extremely effective noise reduction tool. If you have resolution to spare (more than maybe 240 pixels per inch at your intended print size), then you can apply noise reduction, then downsize, then sharpen. The order in which you perform those actions is of paramount importance.

Curtis are you saying that picture styles do not work on .CR2 RAW files? Also I was not aware of downsizing to remove noise. Is there a specific procedure you could refer me to for downsizing?

The only camera settings that are going to directly affect what a RAW file records as image data are the three exposure setting values, Shutter Speed, Aperture and the ISO sensitivity. Having HTP set will also affect the image as it changes the way the in camera light metering sets the exposure for any given ISO value. HTP effectively causes the camera to expose to the left (appears darker than it needs to). Canons DPP software is able to deal with this but other RAW conversion software is either strictly dumb or only knows to boost the image brightness for you to compensate, if it knows to look for the HTP flag in the EXIF data. All other in camera processing options have NO effect on the raw image data in a RAW file. As all RAW files include a fully processed JPEG conversion as an embedded preview, the in camera settings do affect the preview. As with HTP that I have already mentioned Canon's DPP RAW converter is able to pick up the settings used to generate the JPEG preview in camera and apply the same settings as the default conversion. This is why DPP often "looks better" by default than using software by third parties. Third party software such as Adobe's LR or PS's ACR module also tend to use a much lower level of processing by default, as they assume that the photographer will be making their own adjustments to the image.

Alan

Alan's response was pretty comprehensive, and accurate.

For clarity, I would only add that a third-party RAW converter such as Adobe Lightroom will use the camera's white balance setting to render its preview by default, but the RAW data is the same, regardless of camera WB setting.

You could certainly test the broader concept of picture styles and RAW files. Take two identical images with very different picture styles, open both files with a non-Canon RAW converter and see if you notice any difference.

As far as downsizing to reduce noise, the concept is simple enough. When you downsize an image the computer has to interpolate between neighboring pixels and somehow average them to create a new pixel. If there is one "rogue" pixel in a group (digital noise), it becomes less noticeable when averaged with its neighbors.

Previously in this thread it was correctly pointed out that noise reduction can have a blurring effect with a loss of detail. But if you are going to lose detail via downsizing anyway, then there's no disadvantage to applying a bit of noise reduction first.

Since noise reduction targets the rogue pixels and renders them the color of their neighbors, it can help you achieve better color reproduction than downsizing alone. Since digital noise affects individual photo cells on the sensor, noise reduction should always be applied to the RAW file or as early as possible in the post-processing workflow.