Understanding DSLR's - Settings that effect only the JPeg and not the RAW File

[RScottie127 41]

I am starting to get the hang of making my JPeg's look pretty good right out of Camera but still feel like I could get better results if I started editing RAW shots.

My understanding of the RAW image is that it incorporates all the information provided by the light that hits the pixels. As such, this tells me that only a few settings, such as focus, shutter speed and aperture as any filters used would have an effect on the RAW image.

Am I leaving anything out?

If I am thinking correctly, all the other settings are there to allow the camera to edit the data to make the JPeg.

Am I on the right track?

 

[wfooshee1]

All of the color settings and scene presets will only affect images stored by the camera as JPGs.

 

[480sparky52 52]

Not much affects the NEF data. Most settings are applied to any JPEGs that are generated from the raw data, either in-camera or in post.

 

[Ysarex98 92]

I am starting to get the hang of making my JPeg's look pretty good right out of Camera but still feel like I could get better results if I started editing RAW shots.

Yes. and you are on the right track.

By hand processing a raw file I can NEVER FAIL to do better than the software in the camera. So were I to chose to use the camera's picture controls to make the best possible JPEG the camera can produce I would be choosing to degrade all of my images and produce inferior results.

The software in the camera has a couple handicaps that are ultimately determinant.
1. The engineers designing the camera's software work under threat that you at some point will decide to press and hold the shutter release. When you do that you expect the camera to take a burst of images in rapid succession. The camera processing software then has to keep up. In order to keep up the engineers have to compromise. And that's a period. Working at my desk with a raw file I have the option to use better tools that they can't possibly put in the cameras because those better tools would slow the cameras down too much. So this isn't even an issue of skill versus mediocre automation -- better tools produce better results, simple and straightforward.
2. The cameras all function as my old Latin teacher would say bas ackwards. The camera JPEG is a film simulation. A disproportional tone curve has been applied in processing the raw data and the camera's metering system has been adjusted to expose to that curve. That standard curve will occasionally match the lighting/scene contrast in your subject but mostly it won't. The picture controls in the camera can help here but it's a learning curve and the capacity to adjust the output is limited. The camera is pretending it's loaded with a roll of film. Stop pretending there's film in the camera and everything changes. Practically what the cameras end up doing when they pretend they're using film is they frequently expose the sensor in the camera less or more than would be optimal. So, in your first sentence, as you make the JPEGs look pretty good you're often compromising the exposure of the sensor. Learn to expose the sensor optimally and the camera JPEGs will look like cr*p and you'll get much better photo IQ.

My understanding of the RAW image is that it incorporates all the information provided by the light that hits the pixels. As such, this tells me that only a few settings, such as focus, shutter speed and aperture as any filters used would have an effect on the RAW image.

Am I leaving anything out?

If I am thinking correctly, all the other settings are there to allow the camera to edit the data to make the JPeg.

Am I on the right track?

 

[RScottie127 41]

I guess I am just trying to have the best of both worlds.

RAW pics I can edit as I learn how to do that (which is entirely another new adventure for me), and JPeg pictures I can immediately show to others.

 

[480sparky52 52]

If you want the best of both world, shoot raw+JPEG.

 

[RScottie127 41]

If you want the best of both world, shoot raw+JPEG.

That is what I have started to do. Saving the JPeg's on one SD Card and the RAWs on the other.

Still looking at editing software. Thinking about getting Corel's Paintshop software.

 

[Ysarex98 92]

I guess I am just trying to have the best of both worlds.

RAW pics I can edit as I learn how to do that (which is entirely another new adventure for me), and JPeg pictures I can immediately show to others.

While you're learning you can shoot raw + JPEG. When you do that the appropriate action is to try and make as good a JPEG as you can. This will often result in a compromised raw file. You can still work with the raw file and learn to process it, just know that it could often be better.

I ran out on the front porch and grabbed a snapshot for you. One of my wife's potted plants seems to have an oak tree growing in the pot. When we discovered it my wife decided to leave it alone for now in case she wants to transplant the tree.

First I set the camera to pattern metering (P mode), and zeroed the meter allowing the camera to select the exposure. Other settings were auto-WB, noise reduction off and sharpness +2. Arguably the camera made a decent JPEG. I made one minor edit and leveled it (it was hand-held) Here it is and it's full-res:

Next I took control of the exposure and saved a raw file in which I exposed to utilize all (100%) of the recording capacity of the sensor. I processed that raw file and here it is also in full-res:

You should be able to see the EXIF data for both photos. What you'll find I did was I used the camera's EC control to set an exposure +1.6 stops more than the camera's metered exposure. A subtle metering difference probably because I hand-held the camera and the difference between the two photos is 1.3 stops. That is in one sense staggering. That means the camera, in exposing for the JPEG, left more than 1/2 of the sensor completely unused -- 1 stop is 1/2 the sensor!

When I save raw files I use the sensor in my cameras -- all of the sensor. Increasing exposure improves signal to noise ratio and overall image quality. It's there in your camera, why would you not chose to use it fully? Why the difference? Because the camera software is pretending there's film in the camera. It's exposing for the film simulation tone curve of the output JPEG. There's a range in operation here (think bell curve) where the camera will in an ideal lighting/scene-contrast condition deliver both a good JPEG and a good raw file (center of the bell curve) and then as that lighting/scene-contrast condition changes the camera will go to both sides of the bell curve center overexposing the raw file on one side and underexposing it on the other. Most of the raw files the camera will save will not be optimally exposed even though the JPEG looks good.

There's an inexpensive software program called RawDigger that you can use to examine your camera's raw files. And some good articles on their website.

 

[Ysarex98 92]

That is what I have started to do. Saving the JPeg's on one SD Card and the RAWs on the other.

Still looking at editing software. Thinking about getting Corel's Paintshop software.

If you're going to process raw files you want to consider a parametric raw processor and not a raster editor.

When you need a raster editor there's a good free option in Affinity Photo from Canava, better than Paintshop.

Parametric raw processors run the gamut from expensive to free. The best are a bit costly:
Capture One
Lightroom
DXO PL-9
Some middle of the road options at modest prices (watch for sales)
ACDSee
ON1
And some free options:
DarkTable
RawTherapee

 

[480sparky52 52]

While you're learning you can shoot raw + JPEG. When you do that the appropriate action is to try and make as good a JPEG as you can. This will often result in a compromised raw file.

Huh? Improving JPEG results makes the raw file worse? What exactly is the process? Doing what, exactly, makes the JPEG better but 'compromises' the raw data?

 

[Ysarex98 92]

Huh? Improving JPEG results makes the raw file worse? What exactly is the process? Doing what, exactly, makes the JPEG better but 'compromises' the raw data?

If you want a good camera output image (JPEG) you have to expose for that camera JPEG. That exposure is often not an optimum exposure for the sensor and subsequent raw file -- in the example I just posted the camera exposed to create a good JPEG and left over 1/2 the sensor unused -- the raw data then could have been much better.

 

[480sparky52 52]

If you want a good camera output image (JPEG) you have to expose for that camera JPEG. That exposure is often not an optimum exposure for the sensor and subsequent raw file -- in the example I just posted the camera exposed to create a good JPEG and left over 1/2 the sensor unused -- the raw data then could have been much better.

I hope you do realize that ALL digital cameras shoot in raw. And in monochrome. This is why they have a bayer array in front of them. No camera sensor can 'shoot in JPEG' as they are all returning luminance values only. All JPEGS are the result of the raw file data being processed. The sensor has zero ability to determine whether the processing is for raw or JPEG.

 

[Ysarex98 92]

I hope you do realize that ALL digital cameras shoot in raw. And in monochrome. This is why they have a bayer array in front of them. No camera sensor can 'shoot in JPEG' as they are all returning luminance values only. All JPEGS are the result of the raw file data being processed. The sensor has zero ability to determine whether the processing is for raw or JPEG.

Yep, I understand all that. You the user know if you're exposing for the camera JPEG or raw file. I put the camera away that I used to make the example above and didn't save the files. But here's an example using a recent photo I took.

First here's the raw file displayed in RawDigger:

The camera is an Oly E-PL9 and is equipped with a 12 bit ADC. You can see in the two green channels in the histogram the slight nick up indicating the sensor saturation threshold has been reached. That little nick up represents some specular highlights in the scene. The exposure for the sensor & raw file is perfect. 1/3 stop more would blow diffuse highlights and any less exposure would unnecessarily reduce SNR and so compromise the raw file. Note the +1.7 EC that was set when the photo was taken.

Here's the camera JPEG (reduced in size for forum display, otherwise untouched):

That JPEG sucks. It should with an EC set to +1.7. It's badly overexposed. So to improve the JPEG results what do we do? How about reduce the exposure. And since the raw file is already perfectly exposed, what's reducing exposure going to do to the raw file? Make it worse in order to get an improved JPEG. Reducing exposure for the JPEG would reduce SNR in the raw file = worse.

 

[480sparky52 52]

You're missing the point:

ALL digital images start out as raw files. Every single one. There is not a single digital camera that has a JPEG sensor. Or a sensor even capable of taking a JPEG image.

They are all raw files. Every bit of data coming from a sensor is just numerical values assigned to an x-y coordinate. JPEGs are processed raw files. Whether the camera or the post-editing app processes them, it's still the same data coming off the sensor. Maybe the reason the JPEG 'sucks' isn't because it's a JPEG... it's because the processing needed to take raw data and convert it into a JPEG isn't performed right.

 

[Ysarex98 92]

Maybe the reason the JPEG 'sucks' isn't because it's a JPEG...

Of course not -- never said so.

it's because the processing needed to take raw data and convert it into a JPEG isn't performed right.

And that is what I said, it's in the post that you quoted, and that was the WHOLE POINT, did you miss it? "Why the difference? Because the camera software is pretending there's film in the camera. It's exposing for the film simulation tone curve of the output JPEG. There's a range in operation here (think bell curve) where the camera will in an ideal lighting/scene-contrast condition deliver both a good JPEG and a good raw file (center of the bell curve) and then as that lighting/scene-contrast condition changes the camera will go to both sides of the bell curve center overexposing the raw file on one side and underexposing it on the other. Most of the raw files the camera will save will not be optimally exposed even though the JPEG looks good."