July 23, 2014, 02:16:45 AM

Author Topic: Which one is accurate?  (Read 5710 times)

duydaniel

  • 7D
  • *****
  • Posts: 341
    • View Profile
Which one is accurate?
« on: September 06, 2013, 07:14:20 PM »
I want to achieve accuracy
The first photo is how my eye saw it and the camera captured it.
But the second one is supposed "white balance corrected".

Which one is more accurate then?
Thank you

canon rumors FORUM

Which one is accurate?
« on: September 06, 2013, 07:14:20 PM »

Jim Saunders

  • 5D Mark III
  • ******
  • Posts: 656
    • View Profile
Re: Which one is accurate?
« Reply #1 on: September 06, 2013, 07:53:09 PM »
It depends on how you corrected the balance.  I tried the apparently white parts with the tool in LR5 and achieved similarly cool results.  However I've seen the colours get really cool like that when the white source you pick is blown.  Here's what I got from the auto white balance, but whatever looks good to you is the right setting:

I'd probably do better to invest more time and less money.

Valvebounce

  • 5D Mark III
  • ******
  • Posts: 732
  • Still can't use most of it to it's full potential!
    • View Profile
Re: Which one is accurate?
« Reply #2 on: September 06, 2013, 08:08:59 PM »
Hi, do we assume evening sun low and red, if so then the first would be nearer how I would expect it represented with all the whites having a red tinge. If midday with clear blue sky then pure whites would be the correct representation.
For my tuppence the first pic is more pleasing to me, and if that is what it truly looked like then that is how you want to remember the scene, not how some software with white balance settings thinks you should see it. Of course I have been wrong before.  ::)

Cheers Graham.
7D + Grip, 40D + Grip, 20D, EF-S 17-85 Kit lens, EF 70-200 f2.8 L IS II USM, EF 2x III, Sigma 150-500, Sigma 17-70 f2.8-4 C, 50mm f1.8, 550EX some Filters Remotes Macro tubes Tripod heads etc!
20D, BG-E2N, 17-85mm, 50mm are pre loved. :)
(300D Saved a holiday, E-FS 18-55 Cosine 100-300 retired)

neuroanatomist

  • CR GEEK
  • ********
  • Posts: 13513
    • View Profile
Re: Which one is accurate?
« Reply #3 on: September 06, 2013, 08:12:13 PM »
If you're doing copy reproduction of art, the color must be accurate.  If you're taking a picture, set the WB as it pleases you.
EOS 1D X, EOS M, and lots of lenses
______________________________
Flickr | TDP Profile/Gear List

duydaniel

  • 7D
  • *****
  • Posts: 341
    • View Profile
Re: Which one is accurate?
« Reply #4 on: September 06, 2013, 08:14:26 PM »
The first pic is how it looked.
The second pic is more colored accurate (after subtracted the sun) via WB correction.

color accuracy or staying true to the original scene?


Jim Saunders

  • 5D Mark III
  • ******
  • Posts: 656
    • View Profile
Re: Which one is accurate?
« Reply #5 on: September 06, 2013, 08:20:38 PM »
The first pic is how it looked.
The second pic is more colored accurate (after subtracted the sun) via WB correction.

color accuracy or staying true to the original scene?

Like Neuro said, if you need the colours to be faithful to the original scene then a grey card or something is necessary; If the photo is your artwork then whatever looks right, is right.

Jim
I'd probably do better to invest more time and less money.

duydaniel

  • 7D
  • *****
  • Posts: 341
    • View Profile
Re: Which one is accurate?
« Reply #6 on: September 06, 2013, 08:49:27 PM »
The reason I asked was because National Geographic says
"National Geographic Photography want to see the world through your eyes, not the tools of Photoshop or setup photography."

So I would assume the first photo is the one.
I don't plan to submit anything but I just want to go with their guideline

Thank you very much


The first pic is how it looked.
The second pic is more colored accurate (after subtracted the sun) via WB correction.

color accuracy or staying true to the original scene?

Like Neuro said, if you need the colours to be faithful to the original scene then a grey card or something is necessary; If the photo is your artwork then whatever looks right, is right.

Jim

canon rumors FORUM

Re: Which one is accurate?
« Reply #6 on: September 06, 2013, 08:49:27 PM »

paul13walnut5

  • Guest
Re: Which one is accurate?
« Reply #7 on: September 06, 2013, 09:00:02 PM »
I want to achieve accuracy
The first photo is how my eye saw it and the camera captured it.
But the second one is supposed "white balance corrected".

Which one is more accurate then?
Thank you

The joys of white balance, especially away from the presets.

I am interested in your term "white balance corrected"

By what means?

I'm being pedantic, but what did you reference off of in the scene? 

When it comes to sunsets and sunrises you would really need to white balance manually every 10s to be technically correct.

But at sunset and sunrise that would lose the impact of that magic hour light.  And besides, if the suns dipped, you aren't colour metering off of the light source, but reflected light source bouncing off of clouds and the sea, or your card is facing the wrong way etc etc.

So whats the answer to your question.

At magic hour there isn't really one.

A preset of daylight up until streetlights kick on will probably be accurate.. to your eyes recollection.
A preset of flouro will probably serve you well enough when lights kick on, until the sky is all indigo.
It's therafter a question of setting for the right source, 2000k for sodium, 3200k for incandescent..

Aesthtically I prefer the colours of the first one.  When theres high contrast betraying the position of the sun (i.e. low) I would just set to daylight and roll with the punches.  Unless you are shooting scenes of crime.

duydaniel

  • 7D
  • *****
  • Posts: 341
    • View Profile
Re: Which one is accurate?
« Reply #8 on: September 06, 2013, 09:04:19 PM »
Thank you!!!

in LR, I used the WB tool and click on one of the roofs to get the 2nd picture.
So it turns out the first one is better

I want to achieve accuracy
The first photo is how my eye saw it and the camera captured it.
But the second one is supposed "white balance corrected".

Which one is more accurate then?
Thank you

The joys of white balance, especially away from the presets.

I am interested in your term "white balance corrected"

By what means?

I'm being pedantic, but what did you reference off of in the scene? 

When it comes to sunsets and sunrises you would really need to white balance manually every 10s to be technically correct.

But at sunset and sunrise that would lose the impact of that magic hour light.  And besides, if the suns dipped, you aren't colour metering off of the light source, but reflected light source bouncing off of clouds and the sea, or your card is facing the wrong way etc etc.

So whats the answer to your question.

At magic hour there isn't really one.

A preset of daylight up until streetlights kick on will probably be accurate.. to your eyes recollection.
A preset of flouro will probably serve you well enough when lights kick on, until the sky is all indigo.
It's therafter a question of setting for the right source, 2000k for sodium, 3200k for incandescent..

Aesthtically I prefer the colours of the first one.  When theres high contrast betraying the position of the sun (i.e. low) I would just set to daylight and roll with the punches.  Unless you are shooting scenes of crime.

Pi

  • 1D Mark IV
  • ******
  • Posts: 937
    • View Profile
    • Math and Photography
Re: Which one is accurate?
« Reply #9 on: September 06, 2013, 09:20:32 PM »
Like Neuro said, if you need the colours to be faithful to the original scene then a grey card or something is necessary; If the photo is your artwork then whatever looks right, is right.

Actually, a gray card in yellowish light is yellowish. If you use it for WB, you make it look gray, which it is not (in that light). This is based on some absolute notion what gray is.

To take it to an extreme, if you have a monochrome light (laser), and do WB on a gray card, the algorithm will try to divide by zero somewhere.

duydaniel

  • 7D
  • *****
  • Posts: 341
    • View Profile
Re: Which one is accurate?
« Reply #10 on: September 06, 2013, 09:45:54 PM »
Like Neuro said, if you need the colours to be faithful to the original scene then a grey card or something is necessary; If the photo is your artwork then whatever looks right, is right.

Actually, a gray card in yellowish light is yellowish. If you use it for WB, you make it look gray, which it is not (in that light). This is based on some absolute notion what gray is.

To take it to an extreme, if you have a monochrome light (laser), and do WB on a gray card, the algorithm will try to divide by zero somewhere.

this is interesting!!!

Pi

  • 1D Mark IV
  • ******
  • Posts: 937
    • View Profile
    • Math and Photography
Re: Which one is accurate?
« Reply #11 on: September 06, 2013, 09:56:41 PM »
To take it to an extreme, if you have a monochrome light (laser), and do WB on a gray card, the algorithm will try to divide by zero somewhere.

this is interesting!!!

To be more precise, you need no monochrome light (since the CFA filters have broad spectrum) but light with spectrum which is seen as (almost) black by two of the filters. In reality, you divide by small numbers, not zero, and you get strong noise.

neuroanatomist

  • CR GEEK
  • ********
  • Posts: 13513
    • View Profile
Re: Which one is accurate?
« Reply #12 on: September 06, 2013, 10:01:14 PM »
To take it to an extreme, if you have a monochrome light (laser), and do WB on a gray card, the algorithm will try to divide by zero somewhere.

What if I used my IR laser on the gray card?  (Well, actually it's a 3 W laser, so it would set the card on fire, would give an even cooler WB than the second shot...)
EOS 1D X, EOS M, and lots of lenses
______________________________
Flickr | TDP Profile/Gear List

canon rumors FORUM

Re: Which one is accurate?
« Reply #12 on: September 06, 2013, 10:01:14 PM »

Jim Saunders

  • 5D Mark III
  • ******
  • Posts: 656
    • View Profile
Re: Which one is accurate?
« Reply #13 on: September 06, 2013, 10:04:32 PM »
To take it to an extreme, if you have a monochrome light (laser), and do WB on a gray card, the algorithm will try to divide by zero somewhere.

What if I used my IR laser on the gray card?  (Well, actually it's a 3 W laser, so it would set the card on fire, would give an even cooler WB than the second shot...)

Do it!  Do it!   ;D

Jim
I'd probably do better to invest more time and less money.

privatebydesign

  • Canon EF 300mm f/2.8L IS II
  • *******
  • Posts: 2048
  • Ermintrude says "moo"
    • View Profile
Re: Which one is accurate?
« Reply #14 on: September 06, 2013, 10:29:19 PM »
The only way you could get a "true" recreation of the scene is if you made a custom camera profile in the ambient light AND, measured the temperature and spectral characteristics of that ambient light. This is what auto white balance does and is why that is what you consider to be "what you saw", it doesn't simply make white toneless, it attempts to measure the ambient light and adjust toneless white to that temperature. Using the WB tool in post just makes the patch you sample toneless (all three colour channels the same number), which is why it works on white, grey and even black. On the occasions that you sample blown whites the software can't make any corrections as all channels are of equal value.

Profile cards like the X-Rite ColorPassport have a row of several "whites" only one of which is toneless, it is rarely the "correct" WB to actually use.

My first 1D actually has a small window on the front that was part of the ambient light measuring system.
The best time to plant a tree is twenty-five years ago. The second best time is today.

canon rumors FORUM

Re: Which one is accurate?
« Reply #14 on: September 06, 2013, 10:29:19 PM »