Near UniWB for D3

Started Jan 18, 2008 | Discussions
Raul Veteran Member • Posts: 8,409
And why leave the poor D300 so lonely

when will you start working on a little uniWB for it so that I can use it ? jijijiji

Regards.

Raul

 Raul's gear list:Raul's gear list
Nikon Df
OP Psweetie Regular Member • Posts: 453
Re: And why leave the poor D300 so lonely

Hi Raul

Only have a D3 I am afraid so someone else will need to play on that one!
--
Philip Sweetman

Zewt Regular Member • Posts: 437
Re: Near UniWB for D3

I was using UniWB in my D200. And it did seem to give me more accurate in-camera histogram readings.

I always had good results as long as I made sure to shoot a gray card for referencing in post processing.

When I would forget to shoot the gray card, it was more difficult to post-process.

Am I not understanding how to carry the use of the UniWB to the post processing? I would hate to think I have been doing it wrong for 6 months. :(.

Raul Veteran Member • Posts: 8,409
Re: And why leave the poor D300 so lonely

Well, hmmm... er... grrr.....

Someone should start working (jijijijijijiji)

Kindest regards

Raul

Psweetie wrote:

Hi Raul

Only have a D3 I am afraid so someone else will need to play on that
one!
--
Philip Sweetman

 Raul's gear list:Raul's gear list
Nikon Df
Raul Veteran Member • Posts: 8,409
Hello Zewt

Zewt wrote:

I was using UniWB in my D200. And it did seem to give me more
accurate in-camera histogram readings.

it WAS

I always had good results as long as I made sure to shoot a gray card
for referencing in post processing.

Pretty good.

When I would forget to shoot the gray card, it was more difficult to
post-process.

If you use RawMagick Lite it comes out looking great. For others you can use auto WB or the presets - depends which one you use. Gray is the best.

Regards

Raul

 Raul's gear list:Raul's gear list
Nikon Df
Raul Veteran Member • Posts: 8,409
Also please read this
 Raul's gear list:Raul's gear list
Nikon Df
OP Psweetie Regular Member • Posts: 453
Re: Near UniWB for D3

Raul mentioned it I think. You can either take a shot using a gray card in the same light and use that to set WB in PP or you can remember conditions and set it manually in PP (i.e. sunny, cloudy etc. and refine for the look you find best)
--
Philip Sweetman

auhopu New Member • Posts: 16
Re: Near UniWB for D3

I am new to digital and have spent the last couple of months reading. I understand that white balance is done by applying coefficients on the Red and Blue channels. I understand the purpose of uniWB to undo in-camera WB and optimize exposure for SNR.

However it is still unclear to me why the Green channel (e.g. in a neutral scene in daylight) would end up higher than the others before WB. I assume that the Bayer array 1:2:1 ratio between R:G:B would result in more information for the Green channel - more data to use in interpolation - not higher values. Where does the Green channel gain originate from?

I believe I miss a fundamental piece in this puzzle.

Elias

-- hide signature --
Bill Janes Senior Member • Posts: 1,848
Re: Spectral response function

auhopu wrote:

I am new to digital and have spent the last couple of months reading.
I understand that white balance is done by applying coefficients on
the Red and Blue channels. I understand the purpose of uniWB to undo
in-camera WB and optimize exposure for SNR.

However it is still unclear to me why the Green channel (e.g. in a
neutral scene in daylight) would end up higher than the others before
WB. I assume that the Bayer array 1:2:1 ratio between R:G:B would
result in more information for the Green channel - more data to use
in interpolation - not higher values. Where does the Green channel
gain originate from?

I believe I miss a fundamental piece in this puzzle.

This link contains most of the information you need. Look at the java script animation showing the sensor response to various wavelengths of light. The short wavelengths (blues) have higher energies, but are scattered and do not reach the electron collecting part of the chip. Sensitivity is maximal in the greens and reds. The green has higher energy, but the reds penetrate more deeply into the chip. Finally, in the deep infrared, the photons pass through the chip and are not detected.

http://www.olympusmicro.com/primer/digitalimaging/concepts/quantumefficiency.html

Digital cameras usually have an infrared cut filter to take care of the infrared response. Finally each pixel has its own filter (red, green, or blue) and the final signal is a combination of the silicon response, the infrared filter, and the Bayer RGB filter as shown below.

http://scien.stanford.edu/class/psych221/projects/05/joanmoh/spectral.html

The Bayer array does have twice as many green sensors as red or blue sensors, but since the demosaicing algorithm fills in the missing colors for each pixel, I do not this affects the final color balance.
--
Bill Janes

auhopu New Member • Posts: 16
Converter WB and exposure compensation order

Thank you, Bill. It makes sense now.

Ok, I shall take the opportunity to ask something further regarding the uniWB chain, also to confirm that I get it right...

To my understanding the steps involved are roughly:

  • camera WB set to uniWB

  • camera exposure can be set for maximum level recording (by shifting the histograms without clipping them)

  • image is imported to the RAW converter

  • converter WB set as desired (based on grey card, preset, or manually)

  • converter exposure compensation set (towards underexposure)

a) If WB and exposure compensation are applied in a converter, does the order matter? or is this
FINAL_LEVEL = LEVEL * WB * EXPC = LEVEL * EXPC * WB
what happens for each channel?

(as I look again, the above seems oversimplified... but doesn't the converter apply each correction to the raw linear levels?)

b) In case the order does matter, what happens if WB is done first and either the Red or Blue or both coefficients cause clipping in Red or Blue channel? Will a negative exposure compensation recover the "clipped" data?

c) If yes, will the data be saved due to the extra space of the 16bit length of the channel (assuming the RAW file is 12 or 14bit) or for another reason?

d) If yes, does this mean that recovery caused be converter-applied WB is possible for up to 2bits (--> 2 stops) in 14bit RAW?

It took me seven question marks to pose "one" question. I hope I am not repeating issues that have already been resolved in previous posts. I did search around, but I still miss this information.

Thank you for your patience
Elias

-- hide signature --
Raul Veteran Member • Posts: 8,409
Hello Au Hopu

Do you use Mac or Pc?

Raul

auhopu wrote:

Thank you, Bill. It makes sense now.

Ok, I shall take the opportunity to ask something further regarding
the uniWB chain, also to confirm that I get it right...

To my understanding the steps involved are roughly:

  • camera WB set to uniWB

  • camera exposure can be set for maximum level recording (by shifting

the histograms without clipping them)

  • image is imported to the RAW converter

  • converter WB set as desired (based on grey card, preset, or manually)

  • converter exposure compensation set (towards underexposure)

a) If WB and exposure compensation are applied in a converter, does
the order matter? or is this
FINAL_LEVEL = LEVEL * WB * EXPC = LEVEL * EXPC * WB
what happens for each channel?

(as I look again, the above seems oversimplified... but doesn't the
converter apply each correction to the raw linear levels?)

b) In case the order does matter, what happens if WB is done first
and either the Red or Blue or both coefficients cause clipping in Red
or Blue channel? Will a negative exposure compensation recover the
"clipped" data?

c) If yes, will the data be saved due to the extra space of the 16bit
length of the channel (assuming the RAW file is 12 or 14bit) or for
another reason?

d) If yes, does this mean that recovery caused be converter-applied
WB is possible for up to 2bits (--> 2 stops) in 14bit RAW?

It took me seven question marks to pose "one" question. I hope I am
not repeating issues that have already been resolved in previous
posts. I did search around, but I still miss this information.

Thank you for your patience
Elias

 Raul's gear list:Raul's gear list
Nikon Df
auhopu New Member • Posts: 16
Re: Hello Au Hopu

Hi Raul.

I use a PC at the moment, but plan an upgrade sometime soon. The 4GB RAM limit of 32bit OS's makes me consider the Mac option too. Linux is also a candidate. I have played with Gimp but haven't experimented with any RAW converter or import/tagging tool. Rawtherapee is on my to-do list. In case I find and like the tools, why not.

I am an amateur, I do have the luxury of waiting. At the present stage, I try to learn how things work

Elias

-- hide signature --
Raul Veteran Member • Posts: 8,409
Hello Elias

Great subject uniWB; you may want to do a search here for " Iliah Borg ".

Here you will find good material (crazy guy

http://www.aim-dtp.net/aim/content_main.htm

Kind regards

Raul

 Raul's gear list:Raul's gear list
Nikon Df
Alexandre Carneiro Junior Member • Posts: 46
Near UniWB for D700

Sorry for ressurecting this thread

I've been trying to upload the supplied D3 jpeg uniWB file to my D700, figuring with would work fine, since it's the same sensor.

I can't seem to get my D700 to recognize the D3 jpeg. I've tried renaming it, just as suggested before in this thread.

What am I missing here?

Thanks in advance,
Alexandre Carneiro

Alexandre Carneiro Junior Member • Posts: 46
Re: Near UniWB for D700

-bump-

Anyone?

WalterH Contributing Member • Posts: 927
Waiting for move

Raul wrote:

Great subject uniWB; you may want to do a search here for " Iliah
Borg ".

Here you will find good material (crazy guy

http://www.aim-dtp.net/aim/content_main.htm

Kind regards

Raul

-- hide signature --

It seems we have to wait for a move of the http://www pages.
In case this is outdated and a new site is known please post.
Cheers
walter

Gnomus New Member • Posts: 20
Re: Waiting for move

I'd also be interested in getting a file to set UniWB on the D700.

Can anyone help?

panos_m Senior Member • Posts: 1,504
Re: Near UniWB for D700

I do not have a D700 or a D700 nearUniWB file. But starting from the AdobeRGB rgb values of the OP of this thread I was able to make a nearUniWB file for my D3 in about 15 minutes. The values I get are 1.012, 1.000, 1.008. The AdobeRGB values I used are 113,67,97. For quick checking the cooefficients I used UFRaw. If someone can make a D700 nearUniWB jpeg I can host it on my server for future use.
--
Panagiotis

WalterH Contributing Member • Posts: 927
Workflow with Adobe ACR dedicated profiles

Hello,
thanks for posting.

I wonder if there is any specific workflow (and or correction) necessary in case one is using home-made profiles using the Adobe Labs DNG profile editor during ACR import and PP in CS4.

I produced such profiles for some typical shooting conditions for my D3 and at least for me these profiles reduce the number and extend of PP manipulations.

I did use near uni-WB in the past but stopped after I got my own profiles - not to complicate things too much^^.
Now i only use a D3 and this thread came up reminding me to go back to uni-WB.

I wonder if I can just shoot using uni-WB and then in post processing apply my own profiles after correcting WB to correct for the uni-WB settings.

I would try this out before asking but at present I only have a 2300K and "regular" fluorescent light available which would not be a good testing environment.
Perhaps you can safe me some time in testing.
Thanks for any hints.

Alexandre Carneiro Junior Member • Posts: 46
Re: Waiting for move

Why doesn't the D700 recognize the uniWB jpeg file for the D3. I made sure it's in the CF card, but it doesn't show in the preview of the D700.
I tested with a D300 jpeg and it was displayed.

Thanks,
Alexandre Carneiro

Keyboard shortcuts:
FForum MMy threads