RAW vs White balance

Started Nov 20, 2006 | Discussions
vortout Regular Member • Posts: 211
RAW vs White balance

Hi Everybody

There is very interesting results for RAW files done with different presets

One fellow on a Russian Pentax forum made a very simple program reading Pentax RAW file and showing histogram for all 3 channels. ( http://www.penta-club.ru/forum/index.php?showtopic=19120&st=0 )

X – axis has maximum 4095 and Y one shows number of pixels accordingly.

Test shots was done on a tripod with using of incandescent lamp

He used different white balance presets

So enjoy with samples:

AWB:

Daylight

Tungsten:

Manual WB

As you see files are different not only on the header level but also it is different physically for red channels.

Conclusions from forum:

1. Your WB settings can lead to overexposure for RED channel but it is not proved yet since may be it just a way to turn down 3rd party converters.

2. It looks like that 3rd Party converters except SilkyPix do not know about WB behavior and that’s why there is some strange color development results.

Also there are some examples for the histogram for different ISO settings.

ISO200 1/5

ISO400 1/10

ISO800 1/20

ISO1600 1/40

ISO3200 1/80

Conclusions:
1. There is not 1 stop between ISO200 and ISO400

2. ISO 1600 and ISO3200 are rather software made and shall not be used in RAW mode since you can fix exposure in during RAW conversion.

So folk think about it. Any suggestions?

-- hide signature --

Never Say Never

blende8 Senior Member • Posts: 1,488
Re: RAW vs White balance

Regarding the first test on the WB:

I don't understand what has been done. Do you want to say that the RAW file looks different depending on the WB preset?
What exactly do the histograms represent?

OP vortout Regular Member • Posts: 211
Re: RAW vs White balance

Oops sorry, i will put it in not so technical form. People who always use JPG mode are better just to ignore this post.

These 4 histograms represent the structure of each resulting RAW file.

1. On X axis we have level of signal for every pixel on the sensor. It has value from 0 to 4095 what actually represents a 12bit nature of Pentax RAW file.
2. On Y axis we have total number of pixels for corresponding level of signals.

3. Every color has an own representation on the resulting histogram accordingly how it is in the resulting RAW file

4. Actually what we see is a “pure” 12bit histogram without any post processing. Remember that sensor has some number of green pixels, and less number of red and blue ones. Raw processing software interpolates it to normal RGB pixels later on.

Ok. Now I will try to explain what the results mean.

The standard axiom is that RAW file structure does not depend on a WB preset inside of a camera and affects only its header. Actually it means that you can ignore WB settings during the shooting and can adjust it later on in post processing software.

However the differences between histograms proves that it is not so for RED part of Pentax sensor and actually brightness level of all RED photo-sites depends on WB preset even if we use RAW mode.

It actually affects post processing and for that all software which does not know about such behavior (Adobe as an instance) resulting colors will be different from expecting results.
---------------------------------------

The bigger treat is another. It seems that under certain conditions RED channel can be overexposed in case you have wrong settings in your camera for WB even if you use RAW mode!!!! This shall be checked. I did not do any tests myself. I just posted what I found to be quite interesting to share.
---------------------------------------

The ISO test showing that in camera processing actually uses software amplifying instead of hardware amplifying to get ISO higher than 800. Obviously it means that there is no need to use ISO 1600 and 3200 if you shoot in RAW mode. You can get better results in your 3rd party converter.

I hope this time I explained everything

-- hide signature --

Never Say Never

w Regular Member • Posts: 138
Re: RAW vs White balance

Thanks for sharing, this is very interesting. Are these tests from a K100D? I thought only Nikons did color preconditioning but that looks like what is happening here. I guess they're the only ones to advertise it.

I'm surprised by the ISO 1600/3200 results, but it does look like they're just doubling the x-values without any interpolation. If they have some preconditioning circuits before the analog to digital conversion (as the first set of graphs tends to imply), it seems like they would be able to do the boosting there. I'd try to test this conclusion with a real world example but I don't have a pentax yet.

btw, wouldn't this also imply that the dynamic range at ISO 3200 is 1/4 of what it is at ISO 800?

koka Senior Member • Posts: 1,186
interesting, thanks for posting (nt)

nt = no text

RoscoT Senior Member • Posts: 2,943
Dod you know which camera was used? [nt]
-- hide signature --

Rosco
Terminal Stage LBA
http://www.pbase.com/roscot

 RoscoT's gear list:RoscoT's gear list
Nikon D300 Nikon D700 Nikon AF-S Nikkor 70-200mm f/2.8G ED VR Nikon AF Fisheye-Nikkor 16mm f/2.8D Nikon AF Nikkor 50mm f/1.4D +10 more
Jonas B Forum Pro • Posts: 14,596
Re: Dod you know which camera was used?

"Итак, *ist DS на штативе"

I understand nothing of the text but the part *ist DS is repeated.

-- hide signature --

Jonas

Jonas B Forum Pro • Posts: 14,596
Re: RAW vs White balance

Very interesting. I have considered the whibal adjustment being a weather control, I have it at Sunshine, since Day 1 and still... I wonder if this affect the whimsy exposure meter as well?

We seem to have to rethink somewhat with regards to white balancing.

What was the test target, or picture, for the series of pictures, and what lamps where used?

-- hide signature --

Jonas

distudio Veteran Member • Posts: 3,913
Re: RAW vs White balance

w wrote:

btw, wouldn't this also imply that the dynamic range at ISO 3200 is
1/4 of what it is at ISO 800?

That's the implication and expected outcome regardless of method used to achieve the higher ISO settings, it's going to either be clipped or contain noise. However the notion of shooting at ISO800 and pushing in post processing later disregards the fact that you would then need to compensate by underexposing two stops to attain higher shutter speeds for a given aperture and lighting, frankly I'd rather dial in ISO3200.

-- hide signature --

Rob

 distudio's gear list:distudio's gear list
Sony RX100 II Pentax K-5 IIs Pentax K-3 Pentax 645Z Nikon D750 +19 more
OP vortout Regular Member • Posts: 211
Re: RAW vs White balance

I will answer in one post

1. It is DS on tripod

2. The scene is nothing special but light is incandescent lamp was used. I think that subject is not much interesting for such experiments.
3. It affects also Blue part as well. It is confirmed from the forum

4. It looks like confirmed that overexposure can be a problem. I did not see yet any real example

5. I do not think that high ISO (1600+)is a good idea in anyway after seeing pure RAW histogram.

-- hide signature --

Never Say Never

OP vortout Regular Member • Posts: 211
Re: RAW vs White balance

you can use this simple software to experement yourself

http://malykh.com/temp/2006-11-19/malykh-pef-2006-11-19.zip

-- hide signature --

Never Say Never

ptodd
ptodd Senior Member • Posts: 1,327
Re: RAW vs White balance

vortout wrote:

4. It looks like confirmed that overexposure can be a problem. I
did not see yet any real example
5. I do not think that high ISO (1600+)is a good idea in anyway
after seeing pure RAW histogram.

вы, this has a direct impact on quite a lot of my shooting.

-- hide signature --

 ptodd's gear list:ptodd's gear list
Pentax K-30 Samsung D-Xenon 50-200 F4-5.6 ED Tamron SP AF 17-50mm F/2.8 XR Di II LD Aspherical (IF) Tamron SP AF 90mm F/2.8 Di Macro Google Nexus 4 +1 more
Foxbat121 Senior Member • Posts: 1,577
I don't believe it unless you tell me...

How do we know the histogram is from RAW not from interoperated RGB from RAW where WB already applied?

 Foxbat121's gear list:Foxbat121's gear list
Sony a6000 Nikon 1 J4 Nikon 1 J5 Nikon AF-S DX Nikkor 35mm F1.8G Nikon AF Nikkor 180mm f/2.8D ED-IF +15 more
Jonas B Forum Pro • Posts: 14,596
Re: RAW vs White balance

distudio wrote:

w wrote:

btw, wouldn't this also imply that the dynamic range at ISO 3200 is
1/4 of what it is at ISO 800?

The part above I don't understand... I think you have to test this first, I don't dare to guess what that is retained in the shadows in an underexposed ISO800 picture. I'm not at all sure the DR is shrunken to "1/4". What does that mean btw? Usually we count DR in exposure stops... Hey w, can you elaborate a little? I can imagine the opposite: In practise you might get better DR as underexposing probably holds the highlights better. What do you think?

Do remember though the earlier posts on this topic where we learned that ISO1600 and ISO3200 is nothing but software stretched products. To my eyes it seemed as the EV compensation in the raw processor made a better job than the camera.

That's the implication and expected outcome regardless of method
used to achieve the higher ISO settings, it's going to either be
clipped or contain noise. However the notion of shooting at ISO800
and pushing in post processing later disregards the fact that you
would then need to compensate by underexposing two stops to attain
higher shutter speeds for a given aperture and lighting, frankly
I'd rather dial in ISO3200.

I wouldn't. See above. As a sidenote I don't really want to use ISO1600 or ISO3200 (or ISO800 underexposed) at all... but when forced I can just as well stay with ISO800 and underexpose to retain highlights and get (maybe, possible) less banding in the resulting picture.

-- hide signature --

regards,

Jonas

Jonas B Forum Pro • Posts: 14,596
Re: RAW vs White balance

vortout wrote:

I will answer in one post

1. It is DS on tripod
2. The scene is nothing special but light is incandescent lamp was
used. I think that subject is not much interesting for such
experiments.

I would like to think he took pictures of a grey card in controlled light... Ordinary bulbs or "daylight" bulbs? No colors in the subject please if we should look at a histogram...

3. It affects also Blue part as well. It is confirmed from the forum
4. It looks like confirmed that overexposure can be a problem. I
did not see yet any real example
5. I do not think that high ISO (1600+)is a good idea in anyway
after seeing pure RAW histogram.

Thank you for posting!

Jonas

OP vortout Regular Member • Posts: 211
Re: I don't believe it unless you tell me...

Because it is from RAW. you can get software and check with your examples
--
Never Say Never

OP vortout Regular Member • Posts: 211
Re: RAW vs White balance

May be it is so. Or it was jus a wall. But anyway fact is the fact. In camera WB affects RAW file on deep enough level.

I would like to think he took pictures of a grey card in controlled light... > Ordinary bulbs or "daylight" bulbs? No colors in the subject please if we > should look at a histogram...

3. It affects also Blue part as well. It is confirmed from the forum
4. It looks like confirmed that overexposure can be a problem. I
did not see yet any real example
5. I do not think that high ISO (1600+)is a good idea in anyway
after seeing pure RAW histogram.

You are welcome

Thank you for posting!

-- hide signature --

Never Say Never

blende8 Senior Member • Posts: 1,488
Re: RAW vs White balance

Can you explain how to operate this software?

Amazi New Member • Posts: 14
Re: Dod you know which camera was used?

I understand nothing of the text but the part *ist DS is repeated.

Yes, also confirmed on Dl and K100D.

w Regular Member • Posts: 138
Re: RAW vs White balance

Jonas B wrote:

The part above I don't understand... I think you have to test this
first, I don't dare to guess what that is retained in the shadows
in an underexposed ISO800 picture. I'm not at all sure the DR is
shrunken to "1/4". What does that mean btw? Usually we count DR in
exposure stops... Hey w, can you elaborate a little? I can imagine
the opposite: In practise you might get better DR as underexposing
probably holds the highlights better. What do you think?
Do remember though the earlier posts on this topic where we learned
that ISO1600 and ISO3200 is nothing but software stretched
products. To my eyes it seemed as the EV compensation in the raw
processor made a better job than the camera.

That's true, you can theorize all you want, but what really matter is the results. I also would not assume either of the two of vortout's conclusions are true of the K10D until it is tested (completely new sensor and processing). Until I get my K10D though, I can only guess and hope it helps to pass the time

I shouldn't have said the dynamic range is shrunken to 1/4. Keep in mind that I'm only guessing based on what I see in the graphs, I don't actually know anything, and I have never used a pentax. It looks to me like the upper 3/4 of the histogram at ISO3200 is discarded. This is a 12-bit linear histogram though so in a normal histogram (the kind in photoshop), that would correspond to losing two stops dynamic range in the highlights.

This would imply to me that (and I'm still guessing here) underexposing by two stops (set -2ev) and using ISO 800 will give you an extra two stops of highlight detail without increasing the amount of noise. A RAW converter is also likely to do a better job of interpolating. Shooting at ISO3200 wouldn't give you any less noise than ISO800 since the sensor doesn't "expose" these two cases differently.

I didn't see the earlier post you referred to. Like so many others, I came to this forum recently when I found out about the K10D. Then I found out about limited lenses, pancake lenses, and now I have LBA and no camera. Pretty pathetic.

Keyboard shortcuts:
FForum MMy threads