value check

Become a member of the CGSociety

Connect, Share, and Learn with our Large Growing CG Art Community. It's Free!

 
Thread Tools Search this Thread Display Modes
Old 06 June 2014   #1
value check

Was looking around in the forums and google for an answer but I either missed it or don't understand what they mean. What exactly does it mean to do a value check? Am I looking at where the darks and lights are and how they interact depending on the mood I'm trying to achieve? For example in this sketch if I was to do a value check, what am I checking and where am I looking? Thanks for the help!

 
Old 06 June 2014   #2
My understanding is that a value check is to ensure that your image has good contrast. When working with colour, there's a chance that all the values end up being quite similar, decreasing the readability. This may not be immediately obvious until the image has been converted to black and white.
__________________
3D Reel, Photography and Blog
 
Old 06 June 2014   #3
Kevin is right--it's mainly to check to make sure your colored image has effective value contrast, since often our brain is often fooled into thinking that hue and saturation differences are creating value differences too, when they actually don't. By converting the image into B/W, you will see exactly what the values are. The easiest way to do this and still keep your colored version and all the layers, is to simply put an Hue/Saturation/Lightness adjustment layer on top of the layer stack and turn the saturation all the way down. You can then just turn that layer on and off to check the values whenever you feel the need, without messing with your colored layers underneath.

Beyond making sure there is adequate contrast for readability, you have to also make sure the overall tonal composition works. A lot of people make the mistake of having excessive micro-contrast everywhere in the image, as if every part of the image must be turned up to the max in contrast. That is a very amateurish mentality and all it does is create lots of visual noise everywhere. Instead, you have to look at the image from the macro level, and make sure the values are managed and arranged in a way that is well-balanced. Some areas are not supposed to have strong contrast, such as supporting areas, while some areas like the main focal points, can have stronger contrast to lead the eyes.
 
Old 07 July 2014   #4
Hello.
Note that desaturating an image with a Hue/Saturation adjustment will cause the image to be converted to a grayscale representation that has each pixel averaged in its channel values. While this results in a grayscale image, it does not take into account the luminance of each pixel.

For a grayscale representation that accurately matches the luminance of the original image, you need to convert the image from a colour mode to grayscale mode (you can do this in the GIMP, Photoshop etc.) and then "undo" to continue working.

Original image:


Desaturated (channel values are averaged, some values are lost):


Grayscale mode (channel values are summed in weights, values are preserved ):


You can bind the "convert to grayscale mode" action with a hotkey so the process is quicker.
 
Old 08 August 2014   #5
That's interesting - I've been desaturating my images to assess the values but greyscale looks like the way to go.

I just did a little test trying to come up with a non destructive method. I added a grey fill layer using 'color' blending mode to remove the colour. This does that but the result is much darker than when converting to grey scale. Perhaps gamma or something else funky is the cause here?

I tried an exposure adjustment layer on the result to try and make sense of it but I couldn't.

However, if you're using LAB colour then either method works. I don't use LAB, but would like to get a quick non-destructive way to do this...

Last edited by MisterS : 08 August 2014 at 09:45 AM.
 
Old 08 August 2014   #6
I was thinking some more about this -

I think that desaturating an RGB image via an adjustment layer would be best to asses values from an artistic perspective as the result is the perceived value. What do you guys think?
 
Old 08 August 2014   #7
Grayscale conversion is definitely more accurate in terms of actual visual perception, so if you want accuracy, I second Rafael's suggestion of using an action to do it quickly whenever you need to check the values.

You really wouldn't be checking your values constantly anyway--it should be something you do at certain intervals of your workflow when you really need the extra assurance. Most of the time, just squinting your eyes is plenty effective in helping you make judgments about the values. And it's only when you have colors that are very close in values should you be worried. If squinting merges them together too easily, then you know you need to differentiate them more. If not, then you're okay and don't need to worry about it.
 
Old 08 August 2014   #8
Hello again.
I did some more research.

Originally Posted by MisterS: I think that desaturating an RGB image via an adjustment layer would be best to asses values from an artistic perspective as the result is the perceived value. What do you guys think?
Humans perceive green more strongly than red, and red more strongly than blue. This is due to the fact that we evolved in a world where there are more shades of green than those of the other colours.
So grayscale conversions can take this into consideration and generate images that more accurately reproduce the brightness of the original colours that we see, but for this we need to use weighting. You can read some more here:

- http://entropymine.com/imageworsener/grayscale/
- http://www.poynton.com/notes/colour...AQ.html#RTFToC9

- - - - -

We are looking for the most accurate way to convert an image to a "value" only representation so that we can value-check it, but we also want the convenience of being able to turn this representation on and off so that it doesn't introduce any more friction to our workflow of digital painting (so that we don't have to undo, or convert to any other modes etc.).
There seems to be two methods for this:

Method 1
Since version 6.0 Photoshop has an adjustment called Channel Mixer, which can be used for selectively mixing each RGB channel based on a percentage (a weight).
  • Add this adjustment as a layer and move it to the top of the layer stack.
  • Open the options window for it.
  • Turn on the "monochrome" checkbox, and set the channel slider values to the following:
    Red: +21%, Green: +72%, Blue: +7%

Those percentage values are from the latest CIE specification, and can be seen in the second article linked above.
Now you can value-check your work by simply turning on the visibility for this adjustment.
The result is still slightly different than the Photoshop grayscale mode, but the difference is so subtle that this can safely be used for value-checking on production material.
I compared the result of this adjustment with the GIMP grayscale mode, and they're the same.

Method 2
If you own Photoshop CS3 and up, a new adjustment was introduced named Black and White.
There are several sliders to configure this adjustment, but simply applying it in its default settings will produce an accurate grayscale conversion.
Since this can be used as an adjustment layer, it can also be turned on and off.

Last edited by Kryzon : 08 August 2014 at 04:51 AM.
 
Old 08 August 2014   #9
Hehe now we're talking! Thanks guys.

On a similar note, anyone got a good method to Linear add/subtract, using just one adjustment layer and one blending mode!

If I want to add 10% to all values - I can stick a 10% grey at top of my stack set to linear add. Or subtract. Using a solid fill layer I can adjust the 'exposure'.

Or use a levels adjustment set to Linear Add, with the three sliders to the right, and white level at zero, you can increase this to linearly increase exposure. But not decrease which would be very handy.

I've been studying colours and values along side using the 'zone system' (photography). I'm now trying to figure out if there is a systematic method where I could pre calculate what values particular hues would yield under different lighting conditions.

If I'm wasting my time, feel free to tell me.
 
Old 08 August 2014   #10
You're wasting my time.

I'm joking. I believe you can use the Curves adjustment for that, and hand-draw the points that you wish to modify. It can affect different tonal points in different ways, and on separate channels.
 
reply share thread



Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
CGSociety
Society of Digital Artists
www.cgsociety.org

Powered by vBulletin
Copyright 2000 - 2006,
Jelsoft Enterprises Ltd.
Minimize Ads
Forum Jump
Miscellaneous

All times are GMT. The time now is 01:05 AM.


Powered by vBulletin
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.