and please use their services
Beginner White balance quandary.
- Thread starter Thatchapthere
- Start date
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
I think you are wrong there, especially if using a custom white balance.
Having done quite a bit of testing to achieve a white balance for shooting with welding glass as a neutral density filter, Adobe's RAW engine does not have enough latitude to correct the white balance and shows up its limitations.
Yes it is an extreme situation, but does show that there can be reasons for getting it right in camera, rather than relying on post processing after to correct.
Here's a comparison between using Canon's dedicated DPP software and ACR.
RAW or RAW ??? by Steve Bennett, on Flickr
Nope... I said white balance settings on camera will be transferred to the raw program via metadata, but white balance is not a fixed value in raw, and can be adjusted post shoot with impunity.
The OP was referring to cutom picture modes, like vivid, portrait etc... they have no affect on the raw. Nikon's raw programme may pass the settings through as metadata (never used it) but NOTHING you do to a raw file is permanent, and can be changed any way. You can "zero" a raw file with one click.
As for your welding glass shoot... take a test frame with a grey card next time, you'll be surprised.
Terrywoodenpic
Suspended / Banned
- Messages
- 8,172
- Name
- Terry
- Edit My Images
- Yes
I shoot only raw, and leave colour balance on auto except for when I shoot pans. It would probably make no difference if I always shot auto colour balance, but it is part of my routine to set everything to manual when working with pans.
When processing I like to find a mediun neutral grey and try the dropper on that. Some times that works well, if the light is constant over the entire scene. But even a known neutral grey pair of trousers can have a cast thrown into them by a near coloured object. It can be surprising how far out the colour of a scene can be when a grey is forced into being "numerically " neutral.
Provided you colour vision is good and your screen calibrated probably the nearest you ever get to "real" , is setting purely by eye perhaps helped by trying the dropper on a few greys.
Unfortunately there is no absolute way to measure colour balance. Even a grey card can give an unrepresentative result for the majority of the scene.
Autocolour balance with Jpegs usually gives near enough results to tweek slightly in pp. At least it does on my daughters camera.
When processing I like to find a mediun neutral grey and try the dropper on that. Some times that works well, if the light is constant over the entire scene. But even a known neutral grey pair of trousers can have a cast thrown into them by a near coloured object. It can be surprising how far out the colour of a scene can be when a grey is forced into being "numerically " neutral.
Provided you colour vision is good and your screen calibrated probably the nearest you ever get to "real" , is setting purely by eye perhaps helped by trying the dropper on a few greys.
Unfortunately there is no absolute way to measure colour balance. Even a grey card can give an unrepresentative result for the majority of the scene.
Autocolour balance with Jpegs usually gives near enough results to tweek slightly in pp. At least it does on my daughters camera.
boliston
Suspended / Banned
- Messages
- 3,959
- Name
- Adrian
- Edit My Images
- No
Does this not assume that "grey" stays "grey" in any lighting condition?
"grey" will be a lot warmer under incandescent light and a lot colder under shade or cloud - I think that 5600k seems to give the most "real" colours so that images are rendered how they actually appeared at the time.
Terrywoodenpic
Suspended / Banned
- Messages
- 8,172
- Name
- Terry
- Edit My Images
- Yes
That is of course what you are trying to do.. remove the influence of the light.
Your eyes might be able to adjust quite easily to make white, grey, and black, look neutral, under the influence of just about any lighting conditions, but for a camera that is quite difficult. Basically it does so by supposing that all the reflected light in a scene adds up to a neutral grey. Added to that, it makes adjustments using algorithms to compensate for the presence of major colour areas. But all told it is only an estimate. Only with an incident colour meter could the colour of the light illuminating the scene be measured, and even then scenes are very often illuminated with mixed lighting, so there are added difficulties.
If you want a grey to appear warm, it is far easier to set it neutral first, then add a set amount of warmth. You will then get a consistency between shots, much like a transparency shooter used to do by using a warm up filter.
HoppyUK
Suspended / Banned
- Messages
- 23,200
- Name
- Richard
- Edit My Images
- No
If you're saying that Raw files are actually free of software corrections, tweaks and other in-camera interference, then that's not true. It's never been completely true and today all kinds of modifications are made to Raws and baked in with no adjustment possible. Noise reduction at high ISO for example, and lens aberrations corrections.
Leica got a red face recently with the launch of the Leica T, claiming high optical performance from the lens without software assistance. Turns out it's just the opposite, with a lot of software help that's coded into the metadata and cannot be disabled. http://www.dpreview.com/previews/leica-t-typ701/7
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
I'm not saying that. I'm saying you can adjust white balance post shoot every bit as much as you can in camera, and there's no quality penalty for it. I've never once said you have carte blanche to adjust everything as much as you want.
- Messages
- 5,607
- Name
- Stephen
- Edit My Images
- Yes
AM I missing something here? If the corrections are hard coded into the raw data, how can RawTherapee access the data without these corrections being applied - or does it use the hard coded data and reverse the effect? Presumably it must, as otherwise the actual, unmodified raw file would be available - even if Adobe don't provide the means to access it.
Terrywoodenpic
Suspended / Banned
- Messages
- 8,172
- Name
- Terry
- Edit My Images
- Yes
I rather think that these corrections are more like addons, as the unadulterated data would be needded to know which and how much of an correction was needed for each camera set iso, and lens setting.
All processors have had trouble getting the best out of fuji raws even though Fuji shared their data with them. Likewise Leica must have shared this knowledge with Adobe and others as it would have been obviouse to them when they examined the Dng files.
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
Of course grey stays grey. Grey reflects all wavelengths equally.. which is why it's grey.
Just use a grey card.. white balance from it. You'll see.
Terrywoodenpic
Suspended / Banned
- Messages
- 8,172
- Name
- Terry
- Edit My Images
- Yes
You can test a lens with out a camera at all , or mount it on another camera. That will tell you what and how much is being corrected in firmware.
boliston
Suspended / Banned
- Messages
- 3,959
- Name
- Adrian
- Edit My Images
- No
If I were to look at a grey card under incandescent light I'm sure it would have a warmer appearance than if I were to view that same card outside on a cloudy day.
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
Yes, of course, but grey reflects everything that hits it neutrally. It takes on the appearance of the light temperature that hits it. However... as it's a neutral tone, you can white balance from it, accurately.
Just try it... LOL.
This is accepted best practice. That's why you can buy grey cards, and Q cards.. for this very purpose. (Yes, I know grey cards can also be used for light metering with reflective meters)
Terrywoodenpic
Suspended / Banned
- Messages
- 8,172
- Name
- Terry
- Edit My Images
- Yes
A grey card can appear any colour you like, depending on the colour illuminating it.
or even when illuminating it with white light and changing the colour of the light illuminating the background,
our senses are very easily fooled.
Sensors are not fooled and measure the light as reflected. So they would not report a change in colour of the grey when the back ground was changed. But would if the light illuminating the grey was.
A grey card is used by assuming that the light reflected from it should be numerically neutral. The numbers needed to bring the numerical balance of each colour equal, provides the data for calculating the colour balance for all the pixels in the rest of the image.
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
Why are so many having a problem with this.. LOL
I know.. that's the point. It will NEUTRALLY reflect the same spectrum spread that hits it without ADDING or SUBTRACTING anything from it. Hence it is a neutral white balancing point.
Instead of arguing with me.. just go and try it.
Terrywoodenpic
Suspended / Banned
- Messages
- 8,172
- Name
- Terry
- Edit My Images
- Yes
I am not arguing with you I agree with you...
I probably first used grey card before you were born.
But what confuses some people is what information a grey card gives when it might appear warm.
It was that, that I was attempting to add.
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
It doesn't give any information. The raw software will take a 3x3 pixel sample from the area you click. Even if the grey card appear warm, because the lighting is warm, because it is reflecting that light neutrally in simpatico with the light hitting the rest of the scene, when the software sets that warm tone on the grey card to neutral grey, because the card is reflecting everything equally, the rest of the scene is brought into neutrality along with it.
That was for the benefit of others.. not you Terry.
sk66
Suspended / Banned
- Messages
- 9,557
- Name
- Steven
- Edit My Images
- Yes
Some camera settings can/will affect the raw file. Things like Nikon's Active D-lighting will cause the exposure to shift affecting the raw file. And Dark Frame NR also affects the raw file. But most camera settings (camera profile, WB, NR, etc) do not affect the raw file.
However, they *do* affect the review image and the histogram. So if you use the review image (jpeg thumbnail) or histogram to judge/set your exposure, then what you see and what you get could be very different. And various raw editors will apply the jpegs settings as "default settings" to various degrees (and based on preferences). I.e. Nikon View NX will (can) apply all of them... but they are not "baked in" (LR only applies WB settings AFAIK).
However, they *do* affect the review image and the histogram. So if you use the review image (jpeg thumbnail) or histogram to judge/set your exposure, then what you see and what you get could be very different. And various raw editors will apply the jpegs settings as "default settings" to various degrees (and based on preferences). I.e. Nikon View NX will (can) apply all of them... but they are not "baked in" (LR only applies WB settings AFAIK).
dcash29
Suspended / Banned
- Messages
- 3,381
- Edit My Images
- Yes
I think you are wrong there, especially if using a custom white balance.
Having done quite a bit of testing to achieve a white balance for shooting with welding glass as a neutral density filter, Adobe's RAW engine does not have enough latitude to correct the white balance and shows up its limitations.
Yes it is an extreme situation, but does show that there can be reasons for getting it right in camera, rather than relying on post processing after to correct.
Here's a comparison between using Canon's dedicated DPP software and ACR.
If you want to achieve a better white balance with a welding glass, one click with levels will produce a more realistic sea colour
- Messages
- 5,607
- Name
- Stephen
- Edit My Images
- Yes
Yes, you can. But the point was that in the quoted dpreview article, the same raw file was put through Adobe and RawTherapee, and RawTherapee revealed the lens distorions without the software corrections being applied. Hence it appeared that the raw file really did contain the raw data without lens corrections, and it was simply that Adobe makes it impossible to ignore it - which is a different matter to saying that the raw file was permanently altered in camera. (And take "in camera" in both senses.) The point at issue in my mind isn't how good the lens is, it's whether you can access the "real" raw data - and it appeared that you could.
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
Nope... It is not adjusting the raw at all... it adjusts the metering. As the manual says, it "Adjusts exposure before shooting" (italics added by me).
In a way, yes... but it's actually retouching it... not altering it's properties. It maps hot pixels and effectively replaces them with values of adjacent pixels.
I know.
That's because the preview screen shows a JPEG preview... not the raw file, which is why you should never trust it.
Which is why I have my preview screen turned off. I trust my metering skills more than I trust a JPEG preview on a crappy 3 inch screen, or the histogram of a JPEG. However... if you have all that amateur b******s switched off, the JPEG histogram won't be far off.
I believe I said this already... some settings are pushed through as settings in the metadata, but nothing is permanent with a raw file... one click and you can zero it. None of that picture settings b******s is pushed through in Lightroom though... only in the crappy Nikon software.
Last edited:
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
sk66
Suspended / Banned
- Messages
- 9,557
- Name
- Steven
- Edit My Images
- Yes
And therefore the raw file will be recorded differently than it otherwise would have... the result has been affected (I didn't say "adjusted" if it makes a difference).
On my Nikons I use the camera neutral profile with all settings set at 0. I turn all other jpeg/noise settings off and turn off the automatic review. These settings make the review/histogram closer to the raw file, but they may still be notably different depending on the camera (my D4 is pretty accurate, my D800 was horrible, I haven't tested/checked my D810).
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
Huh? You can do the same by just stopping down in manual exposure.. it's not PROCESSING the raw in any way... it's just altering your exposure, just as exposure compensation will
It's not "recorded" differently... it's exposed differently.
Last edited:
dcash29
Suspended / Banned
- Messages
- 3,381
- Edit My Images
- Yes
Or use a bloody grey card!!
What and then batch process your shots?
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
If the lighting hasn't changed, yes. Be careful on location though, as lighting conditions can change rapidly, especially at the beginning and end of the day. You can get different colour temperatures literally minutes apart at sunset and sunrise.
The problem with long exposures though is that it can be a pain to do a grey card test if you're into minutes with a 10 stop... as the light can have changed between taking your grey card shot and the actual shot. However... if at any other time of day, or on overcast days, then yes, do a grey card test shot.
Last edited:
dcash29
Suspended / Banned
- Messages
- 3,381
- Edit My Images
- Yes
Yeah, I think I'll stick with my one click strategy for now.
sk66
Suspended / Banned
- Messages
- 9,557
- Name
- Steven
- Edit My Images
- Yes
Agreed. I'm not sure where the confusion is coming from...
Pookeyhead
Suspended / Banned
- Messages
- 11,746
- Name
- David
- Edit My Images
- No
Terrywoodenpic
Suspended / Banned
- Messages
- 8,172
- Name
- Terry
- Edit My Images
- Yes
I Think that there is some confusion being demonstrated between raw data and the saved raw file.
The unadulterated digital data gathered from the sensor does undergo some processing to create the rawfile.
This is not necessarily the same between cameras or between firmware updates.
There is no definitive standard for this, as all raw files are proprietary.
It is partly because of this that there is always a lag between the advent of a new camera , and the ability of the various raw processors to process them.
The camera manufacturer has at least two options and mostly takes advantage of both. The simple one is to add data in the same way it adds a reduced jpeg to the file. This data potentially instructs the raw processor what changes to make.
The second method is to partially process the raw data in camera, this is often done in response to setting a higher ISO. And can be done either as part of the capture on sensor or later or a mixture of the two.
My real point is that a raw file is rarely anything like an unadulterated raw, the processing done is entirely down to the whim of the manufacturer.
The unadulterated digital data gathered from the sensor does undergo some processing to create the rawfile.
This is not necessarily the same between cameras or between firmware updates.
There is no definitive standard for this, as all raw files are proprietary.
It is partly because of this that there is always a lag between the advent of a new camera , and the ability of the various raw processors to process them.
The camera manufacturer has at least two options and mostly takes advantage of both. The simple one is to add data in the same way it adds a reduced jpeg to the file. This data potentially instructs the raw processor what changes to make.
The second method is to partially process the raw data in camera, this is often done in response to setting a higher ISO. And can be done either as part of the capture on sensor or later or a mixture of the two.
My real point is that a raw file is rarely anything like an unadulterated raw, the processing done is entirely down to the whim of the manufacturer.
- Messages
- 5,607
- Name
- Stephen
- Edit My Images
- Yes
I'm still confused. I can understand that there is a difference between data as it comes in electrical signals and a file written out to some form of storage media. I can understand that the camera maker controls the format of that file (I believe at least one camera produces Adobe standard .dng files as its raw format). I'm confused as to whether this should be taken to imply that as well as controlling the format in which the data appears, the makers also "massage" the data, which seems to be a different issue.
On the quoted snipppet above, I can think of at least two different things that you could mean - that the signal gain is applied after the data is captured (in which case ISO could theoretically be increased in post processing if the full data available from the sensor were passed on) or that noise reduction is applied to the raw data before it's written. Or perhaps something else I haven't thought of.
Terrywoodenpic
Suspended / Banned
- Messages
- 8,172
- Name
- Terry
- Edit My Images
- Yes
It seems that variable signal gain can can now be applied on sensor, as well as during preprocessing in firmware, so I was purposely vague. I suppose anything that happens prior to the final file being written to the card is fair game. It is not necessarily true that ALL the data captured from the sensor is passed on to the raw file.
We will always be at the mercy of the manufacturer in this. The raw file we get is as un-processed as we are ever likely to get.
All cameras have their own colour characteristics. Even those that share the same sensor can have quite different "looks". It must be that part of the firmware function is to establish this look as a base.
When you open a Fuji raw in Photoshop raw processor, you have the option of using the Fuji Normal or Adobe normal. They are quite different. You also have the option of using any of the Fuji colour presets, like astia and velvia. So it seems that Fuji must be supplying a base data that is neither their "Normal" nor the Adobe "Normal" to which these preferences can be added. But there is no way for us to access this base data, nor would it necessarily be useful.
The Various raw processors all arrive at different interpretations of a raw file, in almost all respects.
So what is written in stone and what is down to a particular raw processor is open to question.
HoppyUK
Suspended / Banned
- Messages
- 23,200
- Name
- Richard
- Edit My Images
- No
Raw data is certainly 'massaged' and, increasingly it seems, massaged quite a lot! Detail are never published of course, but I can think of a couple of examples where ISO has obviously been tweaked (presumably to get a good figure for marketing) and heavy noise reduction applied at the same time.
Raw files are not data straight off the sensor, and have never been. At the very least, it needs some colour calibration to be usable, but it doesn't stop there. The real definition of a Raw file is that the data is completely fluid and can be manipulated in post-processing right down to pixel level if needs be, whereas with a JPEG pixels are locked together in clusters by the compression process, and then any data not required in that particular conversion is dumped.
sk66
Suspended / Banned
- Messages
- 9,557
- Name
- Steven
- Edit My Images
- Yes
I said it changes "the result"...The confusion comes from you saying that D Lighting "changes" the raw file.
Andy Taylor
Suspended / Banned
- Messages
- 423
- Name
- Andy
- Edit My Images
- No
It's no different to choosing a certain type of film for a particular look, or adjusting the developing process, except that you can try all the different looks and chose the one that looks best.
OP
Thatchapthere
Suspended / Banned
- Messages
- 223
- Name
- Don
- Edit My Images
- Yes
Crikey a lot to take in and ta for your input.
WB quandary indeed! hahah
WB quandary indeed! hahah
Garry Edwards
Moderator
- Messages
- 13,475
- Name
- Garry Edwards
- Edit My Images
- No
Agreed.
And to illustrate the subjective bit, Fuji cameras (S3 & S5) had a reputation for producing "good" skin tones in their jpeg files - but in reality, it wasn't that they were "better" than those of other makes, they were just warmer, which many people preferred.
juggler
Suspended / Banned
- Messages
- 5,059
- Name
- Simon
- Edit My Images
- No
Something I've never quite understood...
Why is it that "if you shoot RAW it doesn't matter", when you can change the WB of a JPEG just as easy? Is the outcome really that much better with RAW files?
'cos jpegs store colours in 8 bits rather than 12-16 and are compressed at that. When you tweak the colours - i.e. apply a curve to them - you'll lose still more information. It doesn't take much before you can see banding in one or more of the colour channels, especially if you do any further processing.
Applying a WB correction to a jpeg is a destructive operation; when you do it to a raw file it happens immediately after demosaicing.