Which White Balance

[Shadowman]

Outdoor shot lit by streetlamps.



Tutorials, tips say choose the one that looks right to you. I like the warmer version and it looks like the scene as I remember it. But the shot was taken five months ago.

[Glenn NK]

I always use the WB that I think enhances the image.

I would suggest that there are very few images where there is only one WB (light temperature). Shadows in a landscape image have a different temperature than objects in full light. Shadows in an artificially lit image can have several different temperatures depending on the mix of different coloured lights combined with shadows.

Perfection exists only in one's aspirations.

In your image, I might have picked something between the two images, but then, that's my personal opinion and it's only worth what someone else thinks of it.

[Shadowman]

I always use the WB that I think enhances the image.

I would suggest that there are very few images where there is only one WB (light temperature). Shadows in a landscape image have a different temperature than objects in full light. Shadows in an artificially lit image can have several different temperatures depending on the mix of different coloured lights combined with shadows.

Perfection exists only in one's aspirations.

In your image, I might have picked something between the two images, but then, that's my personal opinion and it's only worth what someone else thinks of it.

Glenn,

Thanks for the comments. I'm also sure that since this particular scene was lit by multiple light sources, WB is made so much more difficult to precisely measure.

[pnodrog]

I think the first one is probably reasonably close to a correct WB (maybe a bit green in skin tones) but I would be inclined to add a very slight warming photo filter to it.

[ashcroft]

I like the warmer version and it looks like the scene as I remember it. But the shot was taken five months ago.

But that's the point, isn't it? You are not the only person looking at the image, and we were not there. Had you just presented the one on the right some people might have said it was too warm. I think somewhere in-between which looks 'natural' whatever that means

[Shadowman]

I think the first one is probably reasonably close to a correct WB (maybe a bit green in skin tones) but I would be inclined to add a very slight warming photo filter to it.

L. Paul,

Thanks for the suggestions, it does look more natural; but the golden tones look oh so cool!

[Shadowman]

But that's the point, isn't it? You are not the only person looking at the image, and we were not there. Had you just presented the one on the right some people might have said it was too warm. I think somewhere in-between which looks 'natural' whatever that means

Agreed, perhaps only looking for preferences amongst viewers.

[rpcrowe]

Skin tones are always my primary interest regarding color balance. I selected the skin of the man and woman then reduced the green and the blue. This is the result:

[Shadowman]

Skin tones are always my primary interest regarding color balance. I selected the skin of the man and woman then reduced the green and the blue. This is the result:

Richard,

Nicely done, I often apply a skintone adjustment to most of my images. So far, I've only applied a highlight/shadow, white balance, and noise reduction to the two examples. Much easier to edit indoor images than outdoor images, when I apply flash I expect to run into more challenges.

[Manfred M]

John - You might find the following reference handy.

I will go back to it when my skin tones look wrong. Nothing looks worse than an image where the people don't look quite right (unless of course you are doing so on purpose).

http://www.graphicconnectionkc.com/s...orrection.html

[IzzieK]

If you are only asking about which of the two I prefer, I would have said #1 because it looks more natural because the other one on the right looks like it was shot during the day, but the edit of Richard made #1 looks better because of the added subtle warmth.

[Mark von Kanel]

If you are only asking about which of the two I prefer, I would have said #1 because it looks more natural because the other one on the right looks like it was shot during the day, but the edit of Richard made #1 looks better because of the added subtle warmth.

+1

[Shadowman]

John - You might find the following reference handy.

I will go back to it when my skin tones look wrong. Nothing looks worse than an image where the people don't look quite right (unless of course you are doing so on purpose).

http://www.graphicconnectionkc.com/s...orrection.html

Manfred,

Thanks for the link, very useful.

[Shadowman]

If you are only asking about which of the two I prefer, I would have said #1 because it looks more natural because the other one on the right looks like it was shot during the day, but the edit of Richard made #1 looks better because of the added subtle warmth.

Izzie,

Thanks for the comments and selection. The second image was captured around 8:44PM in August, the golden hour was fading away and the light source was a combination of street lamps and exhibition flood lights.

[chauncey]

This is my Guru...https://www.youtube.com/watch?v=dWaFDKrNrwc

[Shadowman]

This is my Guru...https://www.youtube.com/watch?v=dWaFDKrNrwc

Interesting link, thanks for posting. Something that seems to be left out of the tutorial was time of day that image was captured. Skintones should be affected by time of day somewhat and if I'm not mistaken L*a*b* can be adjusted (D illuminant) to emulate lighting conditions. Does the L*a*b* model change illuminant based on RAW data, is there enough information in a jpeg file to do the same?

[mknittle]

Perfection exists only in one's aspirations.

This often gets lost in peoples pursuit of technical perfection. Not so much here but I have seen images that tell a story better "too dark , too light" etc. Than if they were "better exposed/processed" etc..

[Shadowman]

This is my Guru...https://www.youtube.com/watch?v=dWaFDKrNrwc

Interesting link, thanks for posting. Something that seems to be left out of the tutorial was time of day that image was captured. Skintones should be affected by time of day somewhat and if I'm not mistaken L*a*b* can be adjusted (D illuminant) to emulate lighting conditions. Does the L*a*b* model change illuminant based on RAW data, is there enough information in a jpeg file to do the same?

bump

[Mike Buckley]

Skintones should be affected by time of day somewhat

Everything is affected by the varying light that occurs during the day. For whatever reason, we tend to find small "errors" in skin tones more bothersome than larger "errors" in other items.

I tend to solve those issues by first attending to the white balance, which by definition means making the white tones the way we want them even if doing so doesn't make them white. I then selectively attend to the other areas of the image, especially the skin tones, by changing the color balance. Part of that workflow has to do with the fact that my software adjusts white balance only to the entire image whereas the color balance can be adjusted in selected parts. If your software adjusts white balance selectively, I would argue that there is no meaningful, practical difference between white balance and color balance.

[Shadowman]

Everything is affected by the varying light that occurs during the day. For whatever reason, we tend to find small "errors" in skin tones more bothersome than larger "errors" in other items.

I tend to solve those issues by first attending to the white balance, which by definition means making the white tones the way we want them even if doing so doesn't make them white. I then selectively attend to the other areas of the image, especially the skin tones, by changing the color balance. Part of that workflow has to do with the fact that my software adjusts white balance only to the entire image whereas the color balance can be adjusted in selected parts. If your software adjusts white balance selectively, I would argue that there is no meaningful, practical difference between white balance and color balance.

Mike,

In the link provided by Chauncey, the presenter was expounding on how L*a*b* color model was preferable to cmyk for adjusting skintones, however if skin can be adjusted by white balance alone then perhaps there's no need for selective adjustments. I apply white balance adjustments when the image is converted from RAW to tiff and if I feel I need more selective adjustments; I'll use the color/skin filter in Elements. I've tried adjusting skintones using cmyk method, however just as the presenter in the video stated, skintones can be subjective and vary within an ethnic group, you wouldn't a group photo where everyone's skintone looked identical.