It is just you and me

[mugge]

This gentleman was just sitting there with his performing cockatoo and getting a few dollars. He told me the bird would follow him around when at home. 'He just won't leave me alone.' I liked the interaction between the two and opted for a 'grungy look' to the image. C&C welcome

It's just you and me by Ole Hansen, on Flickr

[Shadowman]

Nice effort although barely any detail in the bird.

[mugge]

You are right John. Here is the edited version. I still want it grungy though

P3110024-Edit-Edit-2 by Ole Hansen, on Flickr

[Shadowman]

Ole,

Much better, if your software allows you can work solely on the bird and leave the surrounding areas as processed.

Ole,

Did you shoot this in B & W mode or manual? I know when I used to shoot in B & W mode the camera's sensor seemed to overexpose too, too far to the right.

[mugge]

Ole,

Much better, if your software allows you can work solely on the bird and leave the surrounding areas as processed.

Ole,

Did you shoot this in B & W mode or manual? I know when I used to shoot in B & W mode the camera's sensor seemed to overexpose too, too far to the right.

Thank you John. I shoot mostly in AP mode and when I'm in street scenes I normally have camera set at one stop underexposure. This time I did not

[Shadowman]

Thank you John. I shoot mostly in AP mode and when I'm in street scenes I normally have camera set at one stop underexposure. This time I did not

Ole,

Thanks for the follow up.

[Manfred M]

I normally have camera set at one stop underexposure

Why would you deliberately underexpose?

[mugge]

Why would you deliberately underexpose?

I started doing that a while ago as many of my images were in the 'overexposed region.' As I do love street scenes it is a safer option to have the camera set for underexposure at least I can recover a lot more images that way. No one likes blown images.

[Manfred M]

I started doing that a while ago as many of my images were in the 'overexposed region.' As I do love street scenes it is a safer option to have the camera set for underexposure at least I can recover a lot more images that way. No one likes blown images.

Clipped shadows are impossible to recover. I shoot to "correct" exposure all the time and very rarely clip the highlights. Underexposed areas tend to show digital noise, so that is something that needs to be avoided.

By correct exposure I mean I check the channel histograms and ensure that no channel is clipping. The camera bases the histogram on the JPEG file, so even if you show a bit of clipping, you still have about 1 to 1-1/2 stops of headroom on raw data.

[george013]

Clipped shadows are impossible to recover. I shoot to "correct" exposure all the time and very rarely clip the highlights. Underexposed areas tend to show digital noise, so that is something that needs to be avoided.

By correct exposure I mean I check the channel histograms and ensure that no channel is clipping. The camera bases the histogram on the JPEG file, so even if you show a bit of clipping, you still have about 1 to 1-1/2 stops of headroom on raw data.

That means you'll have a live histogram. And the time to use it which isn't always present in street photography.
Or you take a shot, examine the histogram and eventual take another corrected shot again.

George

[mugge]

Manfred and George. You are correct it is better to have the exposure spot on. I do not have the time to check histogram and ISO when moving through laneways in Melbourne. There may be a bright highlight and a shadow in one scene. What I do now is I set the ISO on auto and f8 minimum, sometimes f11 and at least 2/3 underexposure. I would prefer 'perfect exposure' but this way I may get some images. I don't think there is a 1 to 11/2 stop leeway (that I can achieve). I must point out I do a fair bit of 'navel photography.'

[xpatUSA]

I don't think there is a 1 to 1-1/2 stop leeway (that I can achieve).

I've never understood this claim, perhaps until now.

Taking the raw green channels as roughly representing exposure, one would expect a 'just clipped' value to be converted to 100% brightness in the output image. Turning that around, 'just clipped' image greens should represent 100% sensor exposure in the raw green channels. So something is missing from the claim.

Perhaps that missing factor is the setting of the clipping indication in-camera. If there is some leeway in that, then there would indeed be exposure headroom, but how much?

Again, taking greens and assuming an effective gamma of 2.2, for 1 EV exposure headroom we get a clipping warning level of:

0.5^(1/2.2) = 0.73 x 255 =186/255.

Seems low to me.

Let's say someone prefers a more normal clipping warning level of, say, 245. Then (245/255)^2.2 = 0.92, an exposure headroom of 0.12EV - nowhere near the expected 1EV. So something's amiss or I have missed something.

By 'exposure' here I mean the raw value expressed as a fraction of the saturated (clipped) sensor value - not the brightness of the output image.

Anybody?

[Manfred M]

I've never understood this claim, perhaps until now.

Taking the raw green channels as roughly representing exposure, one would expect a 'just clipped' value to be converted to 100% brightness in the output image. Turning that around, 'just clipped' image greens should represent 100% sensor exposure in the raw green channels. So something is missing from the claim.

Perhaps that missing factor is the setting of the clipping indication in-camera. If there is some leeway in that, then there would indeed be exposure headroom. How much?

Again, taking greens and assuming an effective gamma of 2.2, for 1 EV exposure headroom we get a clipping warning level of:

0.5^(1/2.2) = 0.73 x 255 =186/255.

Seems low to me.

Let's say someone prefers a clipping warning level of 245. Then (245/255)^2.2 = 0.92, an exposure headroom of 0.13EV - nowhere near the expected 1EV. So something's amiss or I have missed something.

By 'exposure' here I mean the raw value expressed as a fraction of the saturated (clipped) sensor value - not the brightness of the output image.

Anybody?

Ted - There are too many variables in play that we have no information on; specifically how the camera manufacturers set their histograms / clipping indicators. All I know is that if I see quite a bit of clipping on my camera display, when I get the raw file open in Photoshop the channels are not clipped.

My suspicion is that this could be a bit of leeway that the camera manufacturers have built into the in-camera display algorithms as we have no information as to what level those clipping indicator set points they are using. Nor do we know how the in-camera JPEGs are built from the raw data. That's the problem with empirical methods; we know that they work, but can't reverse engineer them as to why they work. Many photographers I have spoken with have noticed this with their own work, regardless of the make and model of camera that they use.

[george013]

Ted - There are too many variables in play that we have no information on; specifically how the camera manufacturers set their histograms / clipping indicators. All I know is that if I see quite a bit of clipping on my camera display, when I get the raw file open in Photoshop the channels are not clipped.

My suspicion is that this could be a bit of leeway that the camera manufacturers have built into the in-camera display algorithms as we have no information as to what level those clipping indicator set points they are using. Nor do we know how the in-camera JPEGs are built from the raw data. That's the problem with empirical methods; we know that they work, but can't reverse engineer them as to why they work. Many photographers I have spoken with have noticed this with their own work, regardless of the make and model of camera that they use.

You've a Nikon. Open a clipping raw in capturenx, save it as a jpg, load it again in capturenx and then compare that jpg with the "raw".
If done you can compare these images in capture with the image in your camera, compare the clipping fields.
By using capturenx you use the camera settings.
I know a simple diagram.....

George

[Manfred M]

You've a Nikon. Open a clipping raw in capturenx, save it as a jpg, load it again in capturenx and then compare that jpg with the "raw".
If done you can compare these images in capture with the image in your camera, compare the clipping fields.
By using capturenx you use the camera settings.
I know a simple diagram.....

George

This has nothing to do with Capture NX or any other software, but rather what the camera shows as being clipped versus what the raw data opens up as in the raw convertor.

When I set up a shot in the studio on a white, lit background, the camera typically shows "blinkies" in the positions and size as shown in this image. The aim is to make the background to be quite white. Too little light and it will appear gray and too much and there will be artifacts that wrap around the subject and cause the edges to wash out.




When I open the file in the raw converter, with the clipping indicator turned on, it looks like this:




That tells me that the camera is being very conservative when it indicates the highlights have been clipped. Canon, Sony, Fuji. Panasonic, etc. users tell me they find the same thing. Testing that I (and many others) have done suggests that one can overexpose by 1 to 1-1/2 stops in camera and still get a raw file that shows no clipping.

I know some photographers who will get a proper exposure reading (using an incident light meter) on the subject and who will then set the background lights 1 to 2 stops brighter than what the lighting on the subject is.

[xpatUSA]

Ted - There are too many variables in play that we have no information on; specifically how the camera manufacturers set their histograms / clipping indicators. All I know is that if I see quite a bit of clipping on my camera display, when I get the raw file open in Photoshop the channels are not clipped.

My suspicion is that this could be a bit of leeway that the camera manufacturers have built into the in-camera display algorithms as we have no information as to what level those clipping indicator set points they are using. Nor do we know how the in-camera JPEGs are built from the raw data.

I am lucky in that some of my cameras' metadata clearly shows the conversion from raw to XYZ (a 3x3 matrix) and, once you have that, conversion to any RGB space and JPEG is documented well enough in the literature. Just sayin'

That's the problem with empirical methods; we know that they work, but can't reverse engineer them as to why they work. Many photographers I have spoken with have noticed this with their own work, regardless of the make and model of camera that they use.

Thanks for the explanation, Manfred. So it would seem that I could do a practical test by simply 'chimping' until the camera shows 'just clipped' and then opening the raw file to examine the exposure headroom. Apparently, I should expect to observe significant headroom, 1EV or more, when so doing.

[george013]

This has nothing to do with Capture NX or any other software, but rather what the camera shows as being clipped versus what the raw data opens up as in the raw convertor.

When I set up a shot in the studio on a white, lit background, the camera typically shows "blinkies" in the positions and size as shown in this image. The aim is to make the background to be quite white. Too little light and it will appear gray and too much and there will be artifacts that wrap around the subject and cause the edges to wash out.




When I open the file in the raw converter, with the clipping indicator turned on, it looks like this:




That tells me that the camera is being very conservative when it indicates the highlights have been clipped. Canon, Sony, Fuji. Panasonic, etc. users tell me they find the same thing. Testing that I (and many others) have done suggests that one can overexpose by 1 to 1-1/2 stops in camera and still get a raw file that shows no clipping.

I know some photographers who will get a proper exposure reading (using an incident light meter) on the subject and who will then set the background lights 1 to 2 stops brighter than what the lighting on the subject is.

Using capturenx in the conversion will take in account the camerasettings of the used Nikon. When using a different converter you will compare 2 different images. That's what ypu're doing above. The images look alike but are the product of 2 different converters.
A small one


I see that the for years heard these that the raw data has more room in the exposure is left. Normal I would say that a jpg due to compression will have less blinkies as the raw.

George

[Manfred M]

Thanks for the explanation, Manfred. So it would seem that I could do a practical test by simply 'chimping' until the camera shows 'just clipped' and then opening the raw file to examine the exposure headroom. Apparently, I should expect to observe significant headroom, 1EV or more, when so doing.

Not quite Ted.

What we did was to set up the camera and set it to manual mode. We opened up the exposure until we just saw the first signs of clipping via the "blinkies" on the back of the camera. We then continued to open up the camera in 1/3 stop increments and took a shot.

Once the cards were downloaded we opened the files in our raw converter and opened each file and pulled back the highlights until we had a file where we could no longer recover the highlights. This varied a bit between subject matter and camera make and model, but in general the tipping point was somewhere between 1-1/3 ev and 1-2/3 ev above our starting value that the cameras first indicated that the image had been clipped.

[xpatUSA]

Not quite Ted.

What we did was to set up the camera and set it to manual mode. We opened up the exposure until we just saw the first signs of clipping via the "blinkies" on the back of the camera. We then continued to open up the camera in 1/3 stop increments and took a shot.

Once the cards were downloaded we opened the files in our raw converter and opened each file and pulled back the highlights until we had a file where we could no longer recover the highlights. This varied a bit between subject matter and camera make and model, but in general the tipping point was somewhere between 1-1/3 ev and 1-2/3 ev above our starting value that the cameras first indicated that the image had been clipped.

I understand, Manfred. Fortunately I have access to raw histograms and just did the following to measure the headroom directly without the need for iteration:

Shot a white card such that the live view histogram was just ETTR (cam does not have live blinkies).

Opened the raw in RawDigger and observed the histogram. Sure enough, there was 1EV of headroom below the known raw saturation level.

A quick test of one, with some room for experimental error, but I must now agree that between the raw capture and the sRGB JPEG there is far more "leeway" than simple math implies!

[george013]

I understand, Manfred. Fortunately I have access to raw histograms and just did the following to measure the headroom directly without the need for iteration:

Shot a white card such that the live view histogram was just ETTR (cam does not have live blinkies).

Opened the raw in RawDigger and observed the histogram. Sure enough, there was 1EV of headroom below the known raw saturation level.

A quick test of one, with some room for experimental error, but I must now agree that between the raw capture and the sRGB JPEG there is far more "leeway" than simple math implies!

I just converted several raw images in capture nx-d. And then compared the jpg with the "raw". All jpg's have LESS clipping.
I think you can use any converter to compare with the raw: just take care the jpg has been converted from the same picture in the same converter.

I don't know if you have seen those puzzles in the newspapers: 2 pictures with some slight differences. If you look careful you can see one of them is slightly darker: that's the one that was modified and saved again.

George