OOC jpgs are sharper than RT processed RAWs?

Been working hard to see if I like RT…as compared to PSE14/NIK. Not unhappy with the latter, but curious to see what RT can do. I’m only a somewhat hobbyist using a Panasonic FZ1000 so I realize I may never get into the finer points of RT. Right now I just want to try some basic PP. Presently, when I compare jpg OOC with its RT PP RAW, I’m noticing that the jpgs are usually noticeably sharper than the RT RAW processed images??? My starting processing profile with RAW RT processing is a custom one: AutoMatching-ISO low/Noise reduction/Defringe/Capture Sharpening/White Balance…all left on their defaults. When doing Sharpening, I use the Contrast Threshold as a guide for mostly edge sharpening and have do quite a bit of trial and error with the Amount and Radius. I haven’t used Threshold as I really don’t understand how to use it. Anyway, I feel I must be missing something with regard to having PP’d images that are not as sharp as images OOC and welcome all suggestions. Thanks.

OOC images have sharpening applied. In some cameras you can control that sharpening.

You are applying capture sharpening, but may need some additional sharpening as well.

JPGs from the camera tend to be sharp, colourful and full of contrast. That is a part of the proprietary charm and great for selling gear.

As you learn more about raw and image processing, you would realize that the enhancement is crude and that you could do better with more advanced processors that give you lots of options. While PS/LR/etc. set their starting points as highly developed, these apps tend to do nothing to the image; so the starting point is super unappealing.

Yes, capture sharpening is good. In fact, I would encourage you to set it high because the operation is done early in the processing pipeline. Means that it will benefit other operations down the road. Sliders that have an auto check box underneath don’t usually need adjustment. (Well, I take that slightly back because I see that radius is the “strength setting” of the tool.) The only (other) slider that needs adjustment is the corner radius boost, and the setting depends on how soft the corners of your lens is at its particular zoom and aperture.

There are other goodies under the Raw tab; e.g., Chromatic Aberration Correction (Defringe is more of a last resort) and pixel repair tools if you have hot or dead pixels and other artifacts from the sensor.

Auto matching is great but you may want to refine the curves, etc. The reason is that it uses the OOC JPG as reference and you may not want that look. Noise reduction truly needs careful tweaking and masking. Try not to go overboard with it as noise is often mixed with real data. You wouldn’t want to remove detail just because you don’t like noise.

Lastly, if you want the end result to be sharp, don’t forget to use post-resize sharpening. Opposite of capture sharpening, post-resize sharpening happens at the end of the pipeline. This will recover the softening introduced by denoising and resizing.

1 Like

As the others say, we might sharpen:

  1. After any denoising and immediately after, or as part of, demosaicing. This is needed more in some cameras than others. Many cameras have a slight built-in blur to prevent aliasing, and we might want to remove that blur.

  2. As a final operation, after resizing for the final output, to make the image more readable or more appealing or whatever.

Perhaps it’s my age, but most images I see are over-sharpened. It’s a very blunt (pardon the pun) tool for improving images. I encourage people to consider masked sharpening. As a subtle effect, this can draw the eye into important features.

If you give us access to one of the raw files we might be able to work on it. You might be surprised by the results, even when you still think the sharpness could be better-

My sharpening has come from RT’s Sharpening and Capture Sharpening; however I must admit that I haven’t done any RT post-resizing sharpening. I didn’t think I needed to because the images were just test images and weren’t resized to begin with. Anyway, here are three images shot from my FZ1000’s (very few in-camera settings: a little sharpening and some reduced denoising)
jpg+Raw capture, the jpg, the RAW, and my processed RAW. Of course I am very interested in what you guys can do and can show me how to do it. BTW, as I hadn’t done much with these images (basically just WB, sharpening, denoising,) so I think it best that when you process, just keep it fairly simple. Thanks again for your help.

OOC jpg

RAW
P1120437.RW2 (22.5 MB)

RT processed RAW saved as jpg


P1120437.jpg.out.pp3 (13.0 KB)

A new fly in the ointment. I tried reprocessing following what I had did previously, but this time I added post-resizing sharpening. Don’t think it worked right. First, I didn’t resize and went straight to sharpening. However, Contrast Threshold didn’t change anything, nor did manipulating the radius/amount sliders. I then turned on Resize and did a slight upscale, and then went back to post-resize sharpening. Still no functioning of the Contrast Threshold, but I think there was slight sharpening with the sliders, but not sure. An image using post-resizing sharpening will have to wait until I learn how to use it.

You don’t need to upscale. Just set scale to 1.0 in resize if you want to use pr-sharpening

Also keep in mind that there is no preview for pr-sharpening

If there is no pre-view in pr-sharp, then how is one to"see" the changes?

Afre, thanks for the useful tips.

http://rawpedia.rawtherapee.com/Resize#Post-Resize_Sharpening

I followed the instructions to “correctly” use PR Sharpening, however didn’t seem to make any difference. Obviously, I’m missing something. Here are the instruction with my comments:

The default values work great, but if you want to change them, here’s how to preview what the image will look like:

  1. Tweak your image as you usually would and enable the Resize tool, e.g. downscale using the Lanczos method to a 900x900px bounding box.
    I found that if I leave Resize Tool as is, the saved image is 900x800…which I don’t like. So, if I change the size to the actual size of the image, I do get the right size. Is this OK to do?

  2. Save the image to a lossless format such as TIFF.
    I save the image as TIFF thusly: Send it over to PSE and then save. Is this OK?

  3. Open that saved TIFF in RawTherapee, apply the Neutral processing profile if that wasn’t done automatically, and enable the Sharpening tool in the Detail tab.
    TIFF image is then opened in RT using the Neutral processing profile.

  4. Zoom to 100% (1:1) and tweak the Sharpening tool’s parameters until you get a result that satisfies you. These are the values you should use in the Post-Resize Sharpening tool.
    TIFF image is tweaked in Sharpening–USM. I do know how to use Sharpening. I note the values from this "second sharpening and use them in PR Sharpening…OK?

  5. Go back to the raw image, enable the Post-Resize Sharpening tool and set it up with the values from the previous step.
    This last step was very confusing and may be where I erred??? So I have just sharpened the TIFF image, noted the values; do I now take this image and bring it into PR-Sharpening? I ask this because the instructions are telling me to “go back to the raw image.” Is this present image the raw image that is being referred to, or do I go back to the original raw image.emphasized text

The Post-Resize Sharpening tool is only available when you use the “Lanczos” resizing method.
Lanczos was left in place.

Gotta thank you guys for hanging in there with me. As I have said before, I’m pretty happy with PSE14+NIK, but always curious to see if I can do better. And, RT certainly has caught my eye.

If you don’t resize your image, you do not need to use pr-sharpening. Just use Capture-sharpening and perhaps on top of it a bit of sharpening in RT details tab (which you can preview)

Post resize sharpening is exactly that - post; viz: sharpening after something you have not done yet. You can not preview any form of ‘output sharpening’ in Ps,Lr,CapOne or anything else.

This is why the ‘step-down’ method of sharpening in Photoshop was developed.

2 Likes

Here’s the best I could come up with in trying to match the sharpness of RT processing vs that of the original jpg…and I don’t believe it is as good as that from the jpg OOC.

P1120437-1…jpg.out.pp3 (12.2 KB)

RT’s output is much cleaner. The saturation is higher than the OOC. The image is missing local contrast and fine detail enhancements. The former will lower saturation as a side effect.

I also did a try


P1120437.jpg.out.pp3 (13.2 KB)

@flycaster To follow up on @afre’s comment:

The image is missing local contrast and fine detail enhancements.

How about investigating more about the contrast enhancement modules that are available?

In my opinion, it is very easy to mistake sharpness for contrast, and vice versa.

Have fun!
Claes in Lund, Sweden

2 Likes