I'm not sure if GMK would try to deviate far from the official colors that they are referencing. Many of the cases of "color matching" disasters I've read about involve the designer literally telling GMK to change the original color codes they provided them because they weren't satisfied on how they perceived the individual samples from them (GMK Necro, GMK Bread, GMK Peaches n Cream(?)), but personally I believe that their colors should be accurate even the first time around especially if they are also using color measuring instruments.
The post on reddit that gained some traction is pretty unfortunate as it seems very desaturated, probably not how the set actually looks in real life (I hope, and based on other pictures). I'm basing these assumptions off of my own RAL deck that I compared colors with (280 20 10
, which is nowhere close to that level of (un)saturation.
I'm curious to see how the set looks outside in direct sunlight, since that can help brighten the dark modifiers and make the color more identifiable and may bring the contrast between the alphas and mods to a more acceptable level compared to the original renders. But that won't really make the situation much better since nobody is using their keyboard like that
Okay as I'm making this rant thing I literally just stumbled upon some kind of mystery that I need to solve (
First, let's refer to the colors from the IC:
More
As far as I know, it was never announced which specific colors were in the set. The modifier color would obviously be RAL 280 20 10, but the alphas color is a bit harder to determine. I believed this was a very important piece of information, so I reached out to the person who did the renders, and they told me the alphas in the renders were RAL 270 30 10. So, there's our baseline! Let's see how Gtrx matched these two colors specifically (spoiler: he didn't?)
More
August 25, 2021: Grtx makes his first update post in his discord server with a timestamp and what looks to be two GMK Nord samples from 3/31/2021 (speaking of which, the Date above that reads 14.12.2020... does it really take GMK that long to send samples to the designer or is the 03-31-2021 date something else? idk, haven't gone through their process in a couple years so this is new to me
More
Unfortunately the yellow sample's color code is obscured, but you can clearly see that the gray is RAL 240 60 10. So based on the renders, that should be the color of the
alpha legends.
This isn't too relevant, but if you've read to the end you would realize that the final alpha color used is closer in lightness to the renders' alpha legends than the renders' alphas lolMore
These are the samples he posted on 10/30/2021 in his discord server, I'm assuming the original batch of samples. Can't say for sure, but the alpha and mod samples here seem like RAL 280 20 10 and
RAL 270 30 10 keep reading...
So the relevant timeline goes like this:
11/04/2021: Gtrx receives/evaluates the first samples and approves the yellow, green, and pink. He states that the alphas and modifiers needs to be less saturated. Quote "New colors will be picked today/tomorrow"
03/04/2022: New colors are picked and sent to GMK for another batch of samples. Here's the result.
More
05/28/2022: pictures of the new batch of samples in his discord server. The bottom right two are the ones of concern. Unfortunately it's not taken in the same condition as the first batch of samples, so I guess this photo is just here for reference. I would say that the modifier color looks less saturated in this image though.
More
Now, this is the official render from GMK I got in Gtrx's discord server from their Candykeys rep. I don't know when they originally got this, it was posted 9/20/2022 but this would probably be the final colors before production started. some people are seeing this and bashing GMK for not having eyes and how does this look anything like the set and stuff but I can assure you GMK's renders are probably the last thing you want to base color opinions off of especially for dark colors
now wait
wait a second
look at the colors
alpha keys: RAL 250 50 10modifier keys: RAL 260 30 05Well looks like color codes were changed after all. Maybe this is actually just how it's usually done if you're not satisfied with R1 colors? Makes me think back to the whole thing with Omnitype and their super objective color science lab machines they use, in those cases I don't think they actually changed the color code but for smaller designers who don't have access to color machines they probably need to tell GMK to actually change to a different color code or else they would reject them because they don't have proof that the color isn't accurate.I had to update this stuff several times... Once I found out what colors were used in the renders, this starts making a bit more sense, and at the same time, a bit less sense. Here's some basic RAL color information. The first number is the hue of the color in multiples of 10 (degrees). The second number is the lightness of the color in multiples of 10. The third number is the chroma of the color (basically saturation) in multiples of 5. So let's do some comparison...
Refreshing our minds: The colors used in the renders, what everyone bought the set for. I'm still not talking about the other colors, since I believe those problems if any pale in comparison to the main issue of alphas and mods here.
Alphas color: RAL 270 30 10 Mod color: RAL 280 20 10Now, here's the catch. RAL 270 30 10 doesn't exist! Which means that it doesn't appear in the physical RAL color book. For some reason, that and other mysterious missing RAL colors appear in online/virtual (even official) libraries. So what I'm assuming is the colors were submitted to GMK, and GMK followed up with "this color doesn't exist." This is not particularly the fault of the person doing the renders. If a color a designer requests you render with exists in the virtual library you're using to render, wouldn't you also not bat an eye (if you didn't have a physical RAL deck to reference)?
What would be a logical course here? One would assume to pick the closest matching colors in terms of matching hue, lightness, and saturation. And what a perfect way to do so using RAL's intuitive color numbering system! Which is apparently not what Gtrx decided to do - so now using that same system, we can find out how off the final colors he chose actually are.
Final Alphas color: RAL 250 50 10 Final Mod color: RAL 260 30 05If you understand how different these colors are, you would understand how the set turned out the way it did. So let me try to explain. From my observation, the main complaints are that
"the alphas are too light," the set is
"too gray," and
"the modifiers are too dark." Let's dissect these one by one.
"the alphas are too light,"Yeah, this one isn't too hard to dissect. Just look at the difference between the advertised render colors and the final alphas color (remember multiples of 10): RAL 270 30 10 vs RAL 250 50 10. That's 2 steps up in lightness, pretty damn big change in lightness to be making. As I said before, RAL 270 30 10 doesn't actually exist, so he would have been forced to pick a new color for the alphas regardless. But why make the new color so light? There's many other options that would have gotten close to RAL 270 30 10. Like maybe RAL 270 30 15. But I'm not going to look into what specific substitute would have been best because that's not my job here.
"too gray"First off, don't trust pictures online to evaluate colors, especially that one famous one floating on reddit. Anyways, why so gray? Chroma is based on the last value of the RAL color. For the advertised render colors, the value is both 10. For the final colors used, alphas are 10 and mod is 05. Even though 05 is a step down in chroma, that shouldn't explain why the alphas look gray. To be honest, I don't have a good theoretical explanation for this. Remember that the original alpha color was supposed to be 2 steps darker than what actually was decided in the end? I think the sheer lightness difference of the final alphas (and the less saturated mod color doesn't help that case either) contributed to that "light gray alphas to dark gray mods" effect instead of "a not-too-contrasty two tone complement to the mod color," if you are getting my thinking.
"the modifiers are too dark."I'm gonna pull something I said above down here:
"I think that in a lot of cases the darker colors when rendered in a virtual environment can cause them to look lighter than they actually look in real life under most lighting conditions."Even though this section is crossed out above, I still think this holds truth. It's good that most renders are done in a perfect condition with perfectly placed lighting to illuminate the colors of the keyset well, which does help with portraying accurate colors. However, if not corrected manually, I believe that pretty dark colors like the one used in this set are often portrayed as lighter than they actually are in most real life lighting conditions. Looking back to the final modifier, let's pay attention to the lightness (middle value): Mod color: RAL 280 20 10, Final Mod color: RAL 260 30 05. You can see that the modifier color used the render is actually a step darker than the final color used.
And some people say the final modifier color is still too dark? Well, what I just said plays a factor into this, but also perception. If you compare the original alpha and mod color (remember multiples of 10), the lightness between them is only 1 step apart (30 vs 20). Now comparing the new colors, they're 2 steps apart (50 vs 30). That's pretty much my reasoning here.
I haven't even mentioned the hue changes (first number), my guess is that it contributed a tiny bit to everything, maybe that extra sprinkle of "something doesn't look right" to the final colors.
I've attached some relevant images from my RAL deck. As always, these are pixels on a screen, so grain of salt here.
More
RAL 250 50 10 (final alpha color) vs RAL 270 30 15 (possible close match to advertised render alpha color)
More
RAL 260 30 05 (final mod color) vs RAL 280 20 05 (advertised render mod color)
So what's the conclusion here? Honestly, I don't know why Gtrx decided to go with the colors he did, especially cranking that lightness slider on the alphas up. My personal guess is that after learning that RAL 270 30 10 did not exist, he tried to pick a new color. However, he might have tried to pick the new color based on the renders through his monitor instead of relying on a physical deck to choose a close match. Maybe held up RAL 250 50 10 from his deck up to his monitor, not paying attention to how objectively different it actually was, and said, "250 50 10 is quite accurate"
Not saying it's actually what happened, just my sort-of best guess of the mindset used.
In the end, this post is just my two cents having gone through the process myself, don't take my words as absolute as I'm not super familiar with color theory and what exactly happened to this set. I made this so hopefully people could try to understand better what may have happened and how it ended up like this. I remember actually liking the look of this set back then when I was relatively fresh to this scene and hoping it would turn out well, and I'm sad that it had to turn out like this after all this time waiting. I don't know the actual metrics of this, but it makes me think that sometimes when designers say that they need "another round of colors," it's not GMK's errors but a blunder of the designer, and sometimes they need to request a completely different color for GMK to make new samples. You can't tell GMK (or any other manufacturer probably in that regard) to "just make this color a little more desaturated" or "make this color a little lighter/darker." They work on data, so unless you have data to back up them not making the color you told them to accurately, they won't budge. That might be why for some sets you've seen that had to get multiple rounds, they have exhibited an effect of being interestingly or depressingly different.