It's times like these where doing these exercises in a class would be both harrowing and helpful. I'm relieved I'm not being graded on my ability to determine the values of colors. But I would really like some feedback both on my technique and how to double check myself. Following, please find my thoughts on how to check the assessment of value of a color using an image editing program.
In my image editor of choice, GIMP, if I select Curves from the Color menu and look at the Value graph it appears that I have one big spike in value right around the middle and a tiny curved bump in the top quarter, which I'll blame on the non-green or gray portions of the image. This seems to confirm my guess as to the Munsell value of this green.
I have a few different options for turning color images into greyscale images. One option is to desaturate the image. There are three ways to desaturate an image: lightness, luminosity, and average. Desaturating by lightness has the exact same effect as selecting the Hue-Saturation option on the Color Menu and sliding the Saturation toggle to the bottom.
Desaturation by luminosity appears to confirm that the chip and the background are the same value.
Desaturation by the average (I'm guessing the average of lightness and luminosity), as one might expect, results in an image about halfway between the relatively high contrast of the image desaturated by lightness and the very low contrast of the image desaturated by luminosity. Not sure it actually tells me much.
I started with the second example first because this first example is so clearly wrong. When I pull up the Value graph, I get at least two distinct bumps, implying I've got at least two distinct values in this image. If I were correct, there should only be one bump because the value of the blue background and the value of the gray chip should be the same. At least that's my layperson's understanding of the Curves/Value graph in GIMP.
I didn't even bother to run the average desaturation considering even the luminosity desaturation didn't make these two look like the same value.
The original sample for this image included a lot of other colors than the one with which I was trying to work, so I trimmed out the other colors to check it. Just like the previous example, when I pull up the Value graph, I get two distinct spikes, implying I've got at least two distinct values in this image. If I were correct, there should only be one spike because the value of the orange background and the value of the gray chip should be the same.
But, unlike the blue example, this desaturation by luminosity actually looks pretty good. How can my Curve have two spikes, but my desaturation by luminosity looks almost as good as the first example?
Granted, the average looks a lot like the desaturation by lightness, implying that it's more different than it is the same, but it still seems like a closer match that the blue example.
Click here to return to This is My Heart home page
It's times like these where doing these exercises in a class would be both harrowing and helpful. I'm relieved I'm not being graded on my ability to determine the values of colors. But I would really like some feedback both on my technique and how to double check myself. Following, please find my thoughts on how to check the assessment of value of a color using an image editing program.
In my image editor of choice, GIMP, if I select Curves from the Color menu and look at the Value graph it appears that I have one big spike in value right around the middle and a tiny curved bump in the top quarter, which I'll blame on the non-green or gray portions of the image. This seems to confirm my guess as to the Munsell value of this green. 
I have a few different options for turning color images into greyscale images. One option is to desaturate the image. There are three ways to desaturate an image: lightness, luminosity, and average. Desaturating by lightness has the exact same effect as selecting the Hue-Saturation option on the Color Menu and sliding the Saturation toggle to the bottom.
Desaturation by luminosity appears to confirm that the chip and the background are the same value.
Desaturation by the average (I'm guessing the average of lightness and luminosity), as one might expect, results in an image about halfway between the relatively high contrast of the image desaturated by lightness and the very low contrast of the image desaturated by luminosity. Not sure it actually tells me much. 
I started with the second example first because this first example is so clearly wrong. When I pull up the Value graph, I get at least two distinct bumps, implying I've got at least two distinct values in this image. If I were correct, there should only be one bump because the value of the blue background and the value of the gray chip should be the same. At least that's my layperson's understanding of the Curves/Value graph in GIMP.
I didn't even bother to run the average desaturation considering even the luminosity desaturation didn't make these two look like the same value.
The original sample for this image included a lot of other colors than the one with which I was trying to work, so I trimmed out the other colors to check it. Just like the previous example, when I pull up the Value graph, I get two distinct spikes, implying I've got at least two distinct values in this image. If I were correct, there should only be one spike because the value of the orange background and the value of the gray chip should be the same.
But, unlike the blue example, this desaturation by luminosity actually looks pretty good. How can my Curve have two spikes, but my desaturation by luminosity looks almost as good as the first example?
Granted, the average looks a lot like the desaturation by lightness, implying that it's more different than it is the same, but it still seems like a closer match that the blue example.
No comments:
Post a Comment