Hacker News Reader

Comment by qayxc | Hacker News Reader

You are correct. NASA employs artists to generate published press images.

The individual pictures sent by the spacecraft are indeed monochrome images taken with different filters. These filters don't usually correspond with RGB and calibration is required to approximate human perception.

Most of the time there's no equivalent at all (this applies to pretty much all deep space imagery) and the colours are basically made up (the intensities aren't - they correspond to frequency responses, just not necessarily within the human vision spectrum).

Here's a discussion on the topic: http://ivc.lib.rochester.edu/pretty-pictures-the-use-of-fals...

View on HN

I'm still conflicted. The article you linked goes deep into philosophy of knowledge, but still arguably misses the most important point - false color can be a tool for analysis, pretty pictures are entertaining (and/or good marketing), but neither of them let us - individuals - get closer to the phenomena.

When I look at photos of distant places on Earth, I do so because I'm trying to imagine how would they look like if I was actually there. This is the main reason people care about photos in general[0] - they're means to capture and share an experience. As a non-astronomer, when I'm looking for photos of things in space, I also want to know - first and foremost - how would these things look to my own eyes if I was close enough to see them.

The article mentions NASA defending their image manipulation as popular astronomy's equivalent of red-eye removal. But it's not that. Red-eye removal exists to correct for the difference between a flash-equipped camera system and human eyes. It's meant to manipulate the image strictly to make it closer to what a real human would've seen on the scene. Where is astronomy's actual equivalent of that? Pictures manipulated in such a way to make them maximally close to what an astronaut in a spacesuit would see, if they were hanging around the astronomical object in question? That's what I'd like to see.

Such pictures will not be as exciting as the propped up marketing shoots, but they'll at least allow the viewers to calibrate their understanding of space with reality. I would think this would be seen as more important in the truth-seeking endeavor of science.

[0] - Marketing notwithstanding - the reason imagery is so useful in advertising is because of this desire; carefully retouched and recomposed pictures are superstimuli, the way sugar is for our sense of taste.

> [...] what an astronaut in a spacesuit would see, if they were hanging around the astronomical object in question? That's what I'd like to see.

Well, sorry to disappoint you there but that's technically impossible. For most deep space images human eyes would likely see nothing at all (because we cannot perceive the frequency bands depicted). In other cases it's hard to tell because light conditions are vastly different from our daily perception (lack of atmosphere, harsh contrasts, very little sunlight) and the sensors collected photons for hours - something that human eyes just can't do.

One problem is that the equipment used is simply incapable of recording images as human eyes would see them - try to snap a picture of the evening sky with a smartphone camera to see this first hand (unless your newfangled device sports AI image enhancement, in which case it's just as fake [0]).

> I would think this would be seen as more important in the truth-seeking endeavor of science.

Most scientists see this "truth" as something objective though, and that's strictly not what human perception is in the first place. [0] has a RAW image of a smartphone next to an AI's interpretation and that's a good analogy to what happens with scientific data as well. Neither of the two versions matches what a person would see, yet we accept the AI version as being "good", even though it neither is depicting what the camera sensor picked up nor showing what a person would have seen.

So what is "the truth" in this case? Is it the grainy, desaturated and dark raw data from the CCD sensors or the artificial construct that tries to mimic what a human observer might have perceived?

[0] https://www.washingtonpost.com/technology/2018/11/14/your-sm...

>For most deep space images human eyes would likely see nothing at all

Depending on the definition of "deep".

Presumably in close orbit around a planet, a person would see something. In fact, I've read that even on Pluto, sunlight at noon is still much brighter than moonlight on Earth.

And considering how far away Orion is, what if someone was ten times closer? Is it unthinkable that it might be comparable to the Milky Way?

> Where is astronomy's actual equivalent of that?

Removing StarLink satellites? ducks

An aspect of this I find fascinating is the different ways photos from Mars are coloured.

The cameras they have sent are quite good and can take 'true colour' images.

From the raw images they can be coloured as if a human was standing on Mars, but more often are coloured as if that part of Mars was on Earth. The reason for this is so that scientists can use their Earth-adjusted eyes to study the images - you want minerals in the photo to look familiar to the Earth scientist.

Reply on news.ycombinator.com