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

[Skyblot36 9]

Of course not -- never said so.

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."

I think I understand now! I’ve read many of your posts mentioning exposing for the RAW file and never grasped the concept.

If I may paraphrase it back to you, the exposure metering and JPEG processing is biased towards creating a pleasing image using in camera processing that mimics the quality of film as far as ability to capture tonal range. But the image sensor has greater capabilities and biasing the exposure to best utilise this range will net us better images if we are prepared to post process the RAW file. Yes?

 

[Ysarex86 80]

I think I understand now! I’ve read many of your posts mentioning exposing for the RAW file and never grasped the concept.

If I may paraphrase it back to you, the exposure metering and JPEG processing is biased towards creating a pleasing image using in camera processing that mimics the quality of film as far as ability to capture tonal range. But the image sensor has greater capabilities and biasing the exposure to best utilise this range will net us better images if we are prepared to post process the RAW file. Yes?

Yes!
Back in the film era we had an old adage; same light, same exposure, and we used to use light meters that we held in the light at the subject and pointed back toward the camera. We measured the incident light on the subject. That worked because there was film in our cameras and the film's disproportional tone curve anticipated the lighting/scene contrast of our subject. Think Bell curve and look at this photo:

That's the camera JPEG and that lighting/scene contrast is centered under the Bell curve. (tech note: I exposed this at +.3 and so removed the normal degree of highlight protection the camera engineers built in. To get the photo you see here I put the raw file back in the camera and generated a new JPEG at -.3) That photo; blue sky, sunlit, normal to slightly high contrast is the target the camera engineers are using to adjust the camera's metering and JPEG processing software. They're assuming the center of the Bell curve -- of course, makes sense. And since the JPEG is a film simulation they're basically in the same place we used to be using film. That made the transition easier for lots of old timers who had learned to expose film. Just one catch: there's really no film in a digital camera and once you figure that out you can start exposing the sensor to best advantage.

Look at the photo I used in the example above of the potted plant. Compare the lighting/scene contrast with this one. I wanted to create an obvious example. When I walked out onto the front porch the late afternoon sun was shining. I deliberately looked for a lower contrast scene with no direct sun (1+ standard deviations off the Bell curve center) knowing what the camera would do and I'd get a good example. It's really very simple; there's no film in the camera, there's a digital sensor that behaves differently -- use it.

 

[Skyblot36 9]

Yes!
Back in the film era we had an old adage; same light, same exposure, and we used to use light meters that we held in the light at the subject and pointed back toward the camera. We measured the incident light on the subject. That worked because there was film in our cameras and the film's disproportional tone curve anticipated the lighting/scene contrast of our subject. Think Bell curve and look at this photo:

View attachment 295760

That's the camera JPEG and that lighting/scene contrast is centered under the Bell curve. (tech note: I exposed this at +.3 and so removed the normal degree of highlight protection the camera engineers built in. To get the photo you see here I put the raw file back in the camera and generated a new JPEG at -.3) That photo; blue sky, sunlit, normal to slightly high contrast is the target the camera engineers are using to adjust the camera's metering and JPEG processing software. They're assuming the center of the Bell curve -- of course, makes sense. And since the JPEG is a film simulation they're basically in the same place we used to be using film. That made the transition easier for lots of old timers who had learned to expose film. Just one catch: there's really no film in a digital camera and once you figure that out you can start exposing the sensor to best advantage.

Look at the photo I used in the example above of the potted plant. Compare the lighting/scene contrast with this one. I wanted to create an obvious example. When I walked out onto the front porch the late afternoon sun was shining. I deliberately looked for a lower contrast scene with no direct sun (1+ standard deviations off the Bell curve center) knowing what the camera would do and I'd get a good example. It's really very simple; there's no film in the camera, there's a digital sensor that behaves differently -- use it.

Thanks!
So now it becomes a learning exercise to find out what my camera is doing with respect to it’s default exposure (Canon 600d), and that sounds like recording RAW+JPEG and comparing the histograms to start with, maybe bracketing the exposure by 0.3 or 0.5.

 

[Ysarex86 80]

Thanks!
So now it becomes a learning exercise to find out what my camera is doing with respect to it’s default exposure (Canon 600d), and that sounds like recording RAW+JPEG and comparing the histograms to start with, maybe bracketing the exposure by 0.3 or 0.5.

Everything you need: RawDigger

 

[AlanKlein82 21]

Yes!
Back in the film era we had an old adage; same light, same exposure, and we used to use light meters that we held in the light at the subject and pointed back toward the camera. We measured the incident light on the subject. That worked because there was film in our cameras and the film's disproportional tone curve anticipated the lighting/scene contrast of our subject. Think Bell curve and look at this photo:

View attachment 295760

That's the camera JPEG and that lighting/scene contrast is centered under the Bell curve. (tech note: I exposed this at +.3 and so removed the normal degree of highlight protection the camera engineers built in. To get the photo you see here I put the raw file back in the camera and generated a new JPEG at -.3) That photo; blue sky, sunlit, normal to slightly high contrast is the target the camera engineers are using to adjust the camera's metering and JPEG processing software. They're assuming the center of the Bell curve -- of course, makes sense. And since the JPEG is a film simulation they're basically in the same place we used to be using film. That made the transition easier for lots of old timers who had learned to expose film. Just one catch: there's really no film in a digital camera and once you figure that out you can start exposing the sensor to best advantage.

Look at the photo I used in the example above of the potted plant. Compare the lighting/scene contrast with this one. I wanted to create an obvious example. When I walked out onto the front porch the late afternoon sun was shining. I deliberately looked for a lower contrast scene with no direct sun (1+ standard deviations off the Bell curve center) knowing what the camera would do and I'd get a good example. It's really very simple; there's no film in the camera, there's a digital sensor that behaves differently -- use it.

To someone like me who shot chromes for years before digital came out, there were no adjustments to the slides once they were shot. So being able to get a lot of editing room even from a digital jpeg is magic. Sure, there's more editing room with RAW. But you still have room with JPEG. Your method of raising the exposure eliminates the use of jpegs. That's OK if you want to edit every RAW picture. But if you happy with 95% of your jpegs, your method is too complicated for most. Shooting at the correct exposure determined by the camera works most of the time and if you shoot RAW + jpeg, you still get plenty of room for editing the RAW if the jpeg doesn't work.

 

[Ysarex86 80]

To someone like me who shot chromes for years before digital came out, there were no adjustments to the slides once they were shot. So being able to get a lot of editing room even from a digital jpeg is magic. Sure, there's more editing room with RAW. But you still have room with JPEG. Your method of raising the exposure eliminates the use of jpegs. That's OK if you want to edit every RAW picture. But if you happy with 95% of your jpegs, your method is too complicated for most. Shooting at the correct exposure determined by the camera works most of the time and if you shoot RAW + jpeg, you still get plenty of room for editing the RAW if the jpeg doesn't work.

I understand lots of folks are happy to edit JPEGs, I won't do it because of the additional image degradation that results when you make any substantial changes to tone/color in the JPEG. I'm happy to have my camera JPEGs trashed.
In fact my method behind the camera is less complicated and easier than trying to get a good JPEG from the camera. For 90% plus of the photos I take I set the same exposure = expose the sensor to capacity. I don't have to worry about any of the camera's features/functions to improve the JPEG result and in the 10% minus of photos where I can't fully expose the sensor I just default to a very simple expose as much as possible. I never have to second guess the exposure, I basically ignore the camera meter, I never have to bracket, I just trip the shutter and nail it every time. For me taking photos has never been easier.
I have only one exposure goal for every photo; place the brightest diffuse highlight at the sensor saturation threshold and click. I determine every exposure using the camera's live-view highlight alert which immediately finds the brightest highlight in the scene. I sometimes have to make the call whether that's a specular versus diffuse highlight, but that's as hard as it gets.

 

[AlanKlein82 21]

I understand lots of folks are happy to edit JPEGs, I won't do it because of the additional image degradation that results when you make any substantial changes to tone/color in the JPEG. I'm happy to have my camera JPEGs trashed.
In fact my method behind the camera is less complicated and easier than trying to get a good JPEG from the camera. For 90% plus of the photos I take I set the same exposure = expose the sensor to capacity. I don't have to worry about any of the camera's features/functions to improve the JPEG result and in the 10% minus of photos where I can't fully expose the sensor I just default to a very simple expose as much as possible. I never have to second guess the exposure, I basically ignore the camera meter, I never have to bracket, I just trip the shutter and nail it every time. For me taking photos has never been easier.
I have only one exposure goal for every photo; place the brightest diffuse highlight at the sensor saturation threshold and click. I determine every exposure using the camera's live-view highlight alert which immediately finds the brightest highlight in the scene. I sometimes have to make the call whether that's a specular versus diffuse highlight, but that's as hard as it gets.

Expose To The Right ETTR might give you more room to edit. But it also increases the chances you might clip the highlights and lose the picture entirely. Also, like you said, the brighter setting probably ruins all the JPEGs that many photographers would find perfectly acceptable for their purposes. But if it works for you, carry on.

 

[greybeard34 7]

The only camera settings I know of that affect RAW/NEF files are IMAGE SIZE, RAW COMPRESSION, and BIT DEPTH.

 

[Ysarex86 80]

Expose To The Right ETTR might give you more room to edit.

I don't do ETTR and what I am doing isn't because I want more room to edit. Very simply I recognize there's an imaging sensor in my camera and I use it -- all of it. Best image quality results from full utilization of that sensor and so I want that. I'm also aware that the metering system/image processor in all my cameras typically does not fully utilize the sensor. I understand why they do that -- I'd rather not so I've learned to get around them. Mostly my camera's metering systems/image processors leave some percentage (at times 50% +) of the sensor underutilized. But if I don't want to clip the highlights in my photos (I don't) then the first and best thing I can do is fully utilize the sensor rather than follow the camera's metering selection because when my camera's metering isn't underutilizing the sensor it tends to go the other way and obliterate the highlights by clipping the sensor. If I'm not setting a plus EC to override the camera then I'm likely setting a minus EC to override the camera.

Took this photo a couple months ago when it was still cold enough to freeze the pond in the Gardens -- early morning with a lovely blue sky and sunshine:

To record that sky in the raw file I had to set a -1 EC. Even with a -1 EC the camera image processor still obliterated the blue sky -- here's the SOOC JPEG (resized only). Can't edit that JPEG and get a blue sky out of it.

Imagine what the camera would have done without the -1 EC. The -1 EC was enough to record the blue sky in the raw file and it's really so simple for me because I just set the same exposure I use for every photo as long as I have enough light: Place the brightest diffuse highlight at the sensor's saturation threshold and click. It's been nearly 50 years now since I started working as a photographer and it has never been easier.

But it also increases the chances you might clip the highlights and lose the picture entirely. Also, like you said, the brighter setting probably ruins all the JPEGs that many photographers would find perfectly acceptable for their purposes. But if it works for you, carry on.

 

[AlanKlein82 21]

I don't do ETTR and what I am doing isn't because I want more room to edit. Very simply I recognize there's an imaging sensor in my camera and I use it -- all of it. Best image quality results from full utilization of that sensor and so I want that. I'm also aware that the metering system/image processor in all my cameras typically does not fully utilize the sensor. I understand why they do that -- I'd rather not so I've learned to get around them. Mostly my camera's metering systems/image processors leave some percentage (at times 50% +) of the sensor underutilized. But if I don't want to clip the highlights in my photos (I don't) then the first and best thing I can do is fully utilize the sensor rather than follow the camera's metering selection because when my camera's metering isn't underutilizing the sensor it tends to go the other way and obliterate the highlights by clipping the sensor. If I'm not setting a plus EC to override the camera then I'm likely setting a minus EC to override the camera.

Took this photo a couple months ago when it was still cold enough to freeze the pond in the Gardens -- early morning with a lovely blue sky and sunshine:

View attachment 295785

To record that sky in the raw file I had to set a -1 EC. Even with a -1 EC the camera image processor still obliterated the blue sky -- here's the SOOC JPEG (resized only). Can't edit that JPEG and get a blue sky out of it.

View attachment 295786

Imagine what the camera would have done without the -1 EC. The -1 EC was enough to record the blue sky in the raw file and it's really so simple for me because I just set the same exposure I use for every photo as long as I have enough light: Place the brightest diffuse highlight at the sensor's saturation threshold and click. It's been nearly 50 years now since I started working as a photographer and it has never been

It seems what you're doing is reducing the exposure by -1, so you don't blow out the sky. But that's nothing special for these type of pictures, if you're not using a graduated, neutral density filter. If you set it at average for the whole picture, the sky would have been overexplosed and blown out.

 

[Ysarex86 80]

It seems what you're doing is reducing the exposure by -1, so you don't blow out the sky. But that's nothing special for these type of pictures, if you're not using a graduated, neutral density filter. If you set it at average for the whole picture, the sky would have been overexplosed and blown out.

Yes, with the meter zeroed the camera would have blown the highlights in the raw file -- way over a full capacity sensor exposure -- raw file ruined so photo ruined. Even at -1 the camera image processor still clipped highlights in the sky and obliterated every trace of blue. But -1 gave me a full capacity sensor exposure with all highlights intact and so everything I needed.