News The dress color controversy.

[Pablo49]

Only ever seen blue and black. When first exposed to the image I heard it was white and gold before I viewed it, and didn't know anything about the controversy, and I still saw blue and black. It is interesting how our brains account for the low amount of color information in either direction. Perception is neato.

 

[Pipcard]

For me, I can only see white and gold. I can see both now.

edit: I took the image and inverted the colors.

The opposite of white is black, and the opposite of yellow is blue (light, not pigment).

 

[Unstung]

|

 

[Hielor]

For me, I can only see white and gold.

edit: I took the image and inverted the colors.

The opposite of white is black, and the opposite of yellow is blue (light, not pigment).

Now you're falling for the opposite of the effect--those lighter areas in that inverted picture are most definitely a light shade of blue, not white. Proof that it's gold, not black!

 

[RGClark]

Veteran Mars imaging aficionados will note similar color controversies occurred with respect to the "real" colors of Mars:

http://www.nbcnews.com/science/space/color-controversies-started-mars-not-thedress-n314601

I wonder what a high resolution film camera would see both in the case of the dress and on Mars.

Bob Clark

---------- Post added at 03:15 AM ---------- Previous post was at 03:12 AM ----------

The xkcd pic is on the right track but misses the mark by a hair.
I'd say the confusion is more related to the situation of these cubes.
Visual color continuity relies heavily on context. Neutral colors have the unique property of being warmer than cool colors and cooler than warm colors.
The problem with xkcd's example is that people swap the color without changing the color of the background.
In my example, the illusion originates in the neutral colors recorded by the camera lack information about whether its a backlit white or a dark blue being washed out by a bright, warm light. If the background wasn't yellowish, it might eliminate the illusion.
There is nothing new or novel about this discussion, the topic is covered academically in the field of artist's color theory.

How would it look if the bright sunlit background was replaced with a blue background?

Bob Clark

 

[SolarLiner]

|

This post is soo cool because I can finally see this "black and blue" thing everyone is talking about.
I see blue and gold, so yeah.

I first assumed it was a white balance problem because of the blue, if you set a colder white balance it will be finally completely white. However it appears that there's no gold in it whatsoever, the factual answer is black and blue http://www.romanoriginals.co.uk/invt/70931?colour=Royal-Blue

 

[PhantomCruiser]

Like I said if I look long enough I can see blue/black. But I can never see it as the "right" shade of blue, it's kind of pale. Here's the dress...

 

[Unstung]

The dress keeps changing colors to me periodically. First it looked white and gold, then black and blue, it recently reverted to white and gold, now it's back to black and blue. Ugh. My brain.

 

[PhantomCruiser]

Only one way to fix that.
Let me introduce you to one of my best friends Al.
Al Cohol
We've got Jack, George, Granddad, Jim, Jonnie and that hippie girl Moon Shine. They should hook you up just fine. :thumbup:

 

[statickid]

Veteran Mars imaging aficionados will note similar color controversies occurred with respect to the "real" colors of Mars:

http://www.nbcnews.com/science/space/color-controversies-started-mars-not-thedress-n314601

I wonder what a high resolution film camera would see both in the case of the dress and on Mars.

Bob Clark

This reminded me of the Mars photos as well, although that is a white-point balance discussion. I thought the dress photo was kind of interesting because it shows how difficult it is to pin down "actual color" ...as in when they post a photo and say "similar to what it would look like if you were there."

I did a quick photo experiment just for fun.

Column: A=AWB, B=Manual adjustment of A1 to "true-color", C=Fluorescent camera preset

I used my phone to take a picture of a white cup. I wanted to see if I could get the cup to appear blue in the photo. The best I could do without moving from my desk was photo A1. I took the photo using "auto white-point balance" setting, or AWB. If you super saturate the photo you can see the cup has a healthy blue-dominant color channel.

I tried to use manual post-editing of THE SAME photo, A1, to make it look more like "what I saw with my own eyes." That's B1. It looks pretty clear that I thought the photo was too dark, and the cup didn't look blue to me in person like on the camera. Interestingly, if you blow out the colors on B1, the light/shadow interface becomes kind of a swapped-up rainbow... there is blueish/purpleish light, turning into a reddish shadow where sunlight is tangent, and a greenish back glow from the room/computer monitor. The healthy mix of hues shows the neutral nature of the "white" that I adjusted it to in the B column.

For kicks I also took a second photo with the fluorescent light white-point balance camera preset, photo C1. The photo looks alot warmer than A1. The funny thing about it is if you blow out the colors of C1, It's pretty much bichromatic... having only red and yellow with no blue at all.

All this reminded me of the mars photos because even if you have the absolute best best calibration, it still might not look EXACTLY like what a person standing on Mars would see, not that any photo really does, but I bet even the camera experts would say... "ahhhhh I see now" when they got there and stepped out the door, and thinking about what they got right and what they didn't expect.

---------- Post added at 07:46 AM ---------- Previous post was at 07:28 AM ----------

Okay,my take on this.

...

(title of this thread) is stupid.

:lol:

 

[Andy44]

This happens all the time when I take photos with my phone, particularly sunsets and the like. No matter how I fiddle with the settings the colors of the photo never look like the colors in the sky in front of me.

Of course, I usually shoot black and white film, thus bypassing all of this...

 

[Ghostrider]

I see it as blue and black.
And now that I see that it seems to be A B Normal, I'm reporting to Area 51 for dissection.

 

[Quick_Nick]

 

[Linguofreak]

http://youtube.com/watch?v=jexnhNfOzHg

http://youtube.com/watch?v=Tw0PlGpaqy4

I can see the dress as either white or blue, but I've not really been able to see the trim as anything but gold until I saw the printout in the second video, where it registers as black or something fairly close. I was able to get the same effect by printing out the photo instead of looking at it on screen, and then looking at it in very dim, indirect, incandescent lighting.

On screen, I still can't see anything but gold.

 

[RGClark]

...
All this reminded me of the mars photos because even if you have the absolute best best calibration, it still might not look EXACTLY like what a person standing on Mars would see, not that any photo really does, but I bet even the camera experts would say... "ahhhhh I see now" when they got there and stepped out the door, and thinking about what they got right and what they didn't expect.

:lol:

Since you are experimentally inclined, perhaps you can emulate the dress scenario by imaging a blue colored sheet with black stripes with bright sunlight streaming in behind it.

I was one of those dissatisfied with the "true color" photos released by NASA of Mars. They were formed by using different color filters to image the scene then combining the filtered images. But NASA adjusted how the filter images were combined. I thought, why not use a "color camera" that does all the colors "at once"?

It's not that trivial a matter though as I'm sure someone familiar with digital cameras will explain.

Bob Clark

 

[PhantomCruiser]

Not only just a digital camera, there is a world of difference between a projected image and a reflected image. The image shown on the computer monitor rarely equals the printout held in the hand.

 

[C3PO]

I wonder if this had gone viral if the picture showed the same effect but the object was a two-tone car. :lol:

 

[Urwumpe]

I wonder if this had gone viral if the picture showed the same effect but the object was a two-tone car. :lol:

Likely they would have just said "cool pearl effect paint!" and go on.

 

[statickid]

I was one of those dissatisfied with the "true color" photos released by NASA of Mars. They were formed by using different color filters to image the scene then combining the filtered images. But NASA adjusted how the filter images were combined. I thought, why not use a "color camera" that does all the colors "at once"?

It's not that trivial a matter though as I'm sure someone familiar with digital cameras will explain.

Bob Clark

There are a few reasons... First and foremost, a "color camera" takes color photos the same way. The difference is that instead of taking three separate photos with big filters (like lens covers), they have millions of tiny filters, one for each pixel; its a Bayer filter.


There are some advantages to using big, interchanging filters.

1. Take only one camera! Space and weight are premium, custom built cameras are expensive.

2. More light is available to a camera with NO filter, so having the option is nice.

3. Higher resolution with available pixels. With a Bayer filter it takes 3-4 pixels to make one "color pixel". Using a whole aperture filter uses 3 separate photos, each at full resolution.

Obvious disadvange:

Time lag. Arguably something could change within the timeframe needed to take three separately filtered photos. Luckily Mars landscapes dont change much from second to second. However this is probably the reason we dont have any color dust devil photos.

 

[Gerdih]

I thought this forum was bettet than this... am I the only one who doesn't care at all about that darn dress?