Modern DSLR component question

[Abitconfused]

There are distinct internal components for white balance and exposure...correct? Or do some cameras use the sensor to obtain white balance information?

[Manfred M]

White balance is something that is assigned to the raw data at the time it is converted to an image file. A raw file does not have any WB information at all.

You can use one of the presets, manually dial in a Kelvin value or do a custom white balance by shooting a neutral target. You can also let the camera calculate one using the AWB setting. This value will be used when creating the jpeg file that you see on your camera's display.

If you shoot raw, you can select the WB setting using the raw conversation software. Both your camera AWB and the raw conversion software algorithms are proprietary.

[dje]

Ed I'm not aware of DSLR's having a separate sensor for Auto WB calculation but I could be wrong. My understanding is that Auto WB is calculated using the captured raw data from the main sensor and the calculated WB setting is then available for use in the raw conversion process. There is no need to determine a setting prior to the image being captured.

I suppose it's possible that some cameras may have a separate sensor to aid in the Auto WB calculation but I'm not familiar with that.

Dave

[Stagecoach]

From what I recall reading some cameras have a separate RGB sensor in the prism area that is used for the 'Auto' WB calculation.

[dje]

From what I recall reading some cameras have a separate RGB sensor in the prism area that is used for the 'Auto' WB calculation.

Grahame is this separate to the metering sensor ?

[george013]

Grahame is this separate to the metering sensor ?

From what I read from Nikon it's a combination of the image sensor and a special RGB sensor.

This link is from the D800 but other camera's have something like this too.

http://imaging.nikon.com/lineup/dslr...features01.htm
Somewher at the bottom

Advanced Scene Recognition System with 91K-pixel RGB sensor

there is a scheme.

George

[dje]

From what I read from Nikon it's a combination of the image sensor and a special RGB sensor.

This link is from the D800 but other camera's have something like this too.

http://imaging.nikon.com/lineup/dslr...features01.htm
Somewher at the bottom there is a scheme.

George

Interesting article George, thanks for the link. It seems to imply that the 91K RGB sensor is the metering sensor but it is also used in conjunction with the image from the main sensor to determine an Auto WB setting (amongst other things).

Dave

[Stagecoach]

Grahame is this separate to the metering sensor ?

Dave, George

It's a function of the metering sensor (as in the D800) and others I assume.

But back to the original question from Ed I suspect there are some cameras that use the main sensor for determining AWB.

[george013]

Dave, George

It's a function of the metering sensor (as in the D800) and others I assume.

The article shows it's a combination of the image sensor and the metering sensor

But back to the original question from Ed I suspect there are some cameras that use the main sensor for determining AWB.

I shoot Nikon and have CaptureNx. So it's said the same converter. When I open a raw-file in the converter I can adjust the wb. In the camera it's set to awb and in the converter I can recalculate the wb. It's always different.
If it's based on the sensor itself I wouldn't expect a difference. I can't check it. I only have a Nikon.
George

[Stagecoach]

I shoot Nikon and have CaptureNx. So it's said the same converter. When I open a raw-file in the converter I can adjust the wb. In the camera it's set to awb and in the converter I can recalculate the wb. It's always different.
If it's based on the sensor itself I wouldn't expect a difference. I can't check it. I only have a Nikon.
George

George,

Ed's original question refers to 'distinct' components within the camera with respect to WB. Suggesting this has nothing to do with what can be done later in post processing or RAW conversion.

There is something, component/s, adjustments, settings (manual WB) that will determine the WB of the output jpg and data that goes with the RAW.

[george013]

George,

Ed's original question refers to 'distinct' components within the camera with respect to WB. Suggesting this has nothing to do with what can be done later in post processing or RAW conversion.

There is something, component/s, adjustments, settings (manual WB) that will determine the WB of the output jpg and data that goes with the RAW.

You misunderstood my reaction.
The basic idea is that in the case of Nikon and CaptureNx the converters are the same. If I shoot raw and jpg and compare them in CaptureNx they're exactly the same. Both are converted to a RGB image. ( Yes, I stick to that). Both are using the same color temperature calculated by the camera. Where I'm surprised of is that recalculating the colour temp in the same converter gives me another value. My conclusion is that the in camera calculation isn't based on the sensor image alone. Which is stated in the article. I don't know if my argumentation is correct, but it is one. So probably someone with another brand camera can try this too.

The basic idea behind wb is a very grey world.

George

[Dave Humphries]

Hi Ed,

I'm wondering why you are asking the original question; technical curiosity or perhaps because things aren't coming out right in some of your shots?

I ask because in normal use, I don't see how it matters to you.

Nothing wrong with technical curiosity, I like to know how things work.
If the reason is the latter, it may help to show us a 'good' and 'bad' shot and explain a bit more.

Thanks, Dave

[Inkanyezi]

Different manufacturers have different approaches, and Nikon has claimed that a comparison of scenes is used for WB evaluation. This does make sense, as there are so far no really reliable calculations that can evaluate what white balance to use. After all, the camera cannot see the light source, only guess.

So AWB is mostly a guesstimate in the more advanced cameras, and the camera in most cases will choose among a set of preset WB settings; sunshine, cloudy, shade, tungsten, fluorescent etc, picking the one that seems most likely for the scene. In quite a few compact cameras, it will instead adjust the WB to make the world grey, which causes a complementary colour cast to any image shot with a prevalent colour (magenta cast in the park etc).

So it is indeed the little sensor in the viewfinder that sets the WB before processing the RAW image, and any software that is told to use the same setting will do so, regardless of image content. But of course users may alter the WB to their personal preferences. There is also software with AWB capability, and their approach may be an algorithm that tries to calculate the most likely WB for the scene, which is often wrong. The only way to assess a truly correct white balance is to measure the light source(s). The most convenient way to do this with multiple sources (sun and sky as well as clouds and maybe reflected light from buildings etc) is to use a white balance target, a neutral card, which can be shot in the same light and used for setting the software coefficients.

[dje]

And for a DSLR in Live View mode, it must use the main sensor to calculate an auto WB setting I assume. I must do a little test tomorrow taking two shots of the same scene with the same lighting, one in LV mode and the other normal, and then compare the WB multiplier values recorded in the raw files. They may be a bit different if two different sensors (ie the metering sensor and the main sensor) are used for the auto WB calculation.

Dave

[george013]

And for a DSLR in Live View mode, it must use the main sensor to calculate an auto WB setting I assume. I must do a little test tomorrow taking two shots of the same scene with the same lighting, one in LV mode and the other normal, and then compare the WB multiplier values recorded in the raw files. They may be a bit different if two different sensors (ie the metering sensor and the main sensor) are used for the auto WB calculation.

Dave

Pay attention to the mirror sound when shooting live view. The only sound should be the closing of the mirror. I've no experience with live view but I'm pretty sure to remember that with the first camera's the mirror closed and opened again. So doing its measurements.

George

[Manfred M]

The sensor in the pentaprism area is the light metering sensor. Nikon has used a RGB light meter for quite some time and Canon has now gone this way too in its latest generation of DSLRs, versus a system that measures luminosity only.

LiveView uses data from the sensor and the results of using either approach are quite similar.

[george013]

The sensor in the pentaprism area is the light metering sensor. Nikon has used a RGB light meter for quite some time and Canon has now gone this way too in its latest generation of DSLRs, versus a system that measures luminosity only.

LiveView uses data from the sensor and the results of using either approach are quite similar.

Not quite true if you mean with sensor the image sensor. I just checked the D700. Put the mode on M and live view on. First of all the screen doesn't change when I change the ss or aperture. This was subject in another thread not so long ago. I choice a long ss, let's say 1s. Live view on, clear image on the screen, press shutter button, mirror closing, measurements?, mirror up, curtain open, mirror closing. This is the D700. I don't know of newer ones.
I think, not sure, the only thing that can be done with live view is focusing. But again, I'm not sure.

George

[dje]

George my D610 has similar mirror behaviour in Live View mode ie the mirror closes momentarily before the shutter operates and also momentarily after the shutter exposure is complete. I've read somewhere that this is because the mirror mechanism is mechanically linked to the aperture stop down mechanism and doesn't necessarily have anything to do with exposing the metering sensor. Apparently though it's not like this on all Nikon models, particularly the higher end ones.

Dave

[Abitconfused]

Hi Ed,

I'm wondering why you are asking the original question; technical curiosity or perhaps because things aren't coming out right in some of your shots?

I ask because in normal use, I don't see how it matters to you.

Nothing wrong with technical curiosity, I like to know how things work.
If the reason is the latter, it may help to show us a 'good' and 'bad' shot and explain a bit more.

Thanks, Dave

Thank you. I want more technical insight into camera trends (component or software) in exposure and WB to see where the manufacturers are going technically, to have a better idea how their engineers heads work, and to better understand the input/output to leverage strengths and weaknesses for better results. Yes, there is no normal for me!

[Manfred M]

Not quite true if you mean with sensor the image sensor. I just checked the D700. Put the mode on M and live view on. First of all the screen doesn't change when I change the ss or aperture. This was subject in another thread not so long ago. I choice a long ss, let's say 1s. Live view on, clear image on the screen, press shutter button, mirror closing, measurements?, mirror up, curtain open, mirror closing. This is the D700. I don't know of newer ones.
I think, not sure, the only thing that can be done with live view is focusing. But again, I'm not sure.

George

With the D800 the mirror stays up. I did double check (firing a burst - the I could hear the shutter, but not the mirror slap), and of course in video mode the metering and autofocus all take place with the mirror up.