Friday, April 12, 2019

Black Hole Photo

As we all know, Scientists Have Photographed a Big Black Hole. With an array of radio telescopes and some software, apparently.

Lewis Bush and I had a short discussion about how one ought to think about this "photo" and then I thought about it some quietly by myself. As always in these things, you should assume that anything intelligent in what follows originated with Lewis, and anything stupid, with me.

A JPEG file is a collection of data, and a print made from it is a visualization of that data. It is in some ways no different from a chart of recorded temperatures over time, except that the data it is helping you visualize is the pattern of light cast by a certain lens on a certain surface, not a temperature. Analog film does not materially change this. It too collects readings of light-pattern data, and allows the visualization of that data on paper in much the same way a piece of litmus paper enables you to visualize the acidity of a solution.

Still, there is, it feels obvious, something critically different between a print from a JPEG and a paper strip recording temperature over time. Both are "mechanical" reproductions of data, and thus trustworthy in a certain sense. But the picture feels quite different.

The difference, I think, is that the picture allows us to apprehend the data in a way that mimics the original experience of seeing the whatever-it-was. Imagine that instead of a paper strip with temperatures, we had a device that "played back" the temperatures, perhaps by heating or cooling our fingertip. Now, rather than simply seeing "it got hotter at 10am" we can re-experience that heat to a degree.

The critical difference is that the print of the JPEG offers us the idea that it looked like that in a limited way, in the same way that the temperature gadget offers us a more direct line to it felt like that.

Scientific imaging is frequently presented in a way that is both data visualization, and it looked like that but only sometimes is the second part true.

Electron microscopy of a bug's nose is, arguably, a proper photograph. Sure, we would have to be very very small, but that is what the bug's nose would look like. Reconstructed color imagery of Saturn? Again, if we were hanging in space in the right spot, and maybe had somewhat more sensitive eyes that are nonetheless much like the eyes we have, sure, Saturn would actually look like that.

Those pictures of individual atoms that pop up from time to time? We begin to stray into strange territory. We might imagine ourselves very very small indeed, but with what are we "seeing" these atoms? Photons are the size of basketballs. There is really no way to imagine that atoms look like that in any meaningful way. These things are data visualizations that, while I think useful, stray out of the bounds of a proper photograph.

In the same way, we might imagine ourselves very large and in space near the center of galaxy M87, looking at this supermassive black hole.

Firstly, this thing is, like a JPEG print, a data visualization. Secondly, it is apparently extrapolated quite a bit by software (apparently the process of emulating an Earth-sized radio telescope with a lot of human-scale radio-telescopes scattered around the surface of an Earth-sized object like, say, Earth, leaves some gaps that have to get filled in, in some sense.) So, it resembles computational photography, or perhaps Adobe's content-aware fill. It's a visualization of an intelligent and reasonable guess at a much larger, notional, data set.

Would it look like that if we were in the right place? What if our eyes were sensitive to radio waves? How much different from our human selves would we have to be to directly perceive something that looks like the Black Hole Photo, and is it even physically imaginable that we could?

This is an interesting question, and I do not know the answer.


  1. Thank you for the post. The question is very good.
    I am not sure about it too. Scientists do not either, I suppose. I would say they will have many versions of JPEG picture, this one looked nice, so they give it public. They need a public support, they need public money, so they release something attractive. So, they probably know, that public will take this picture as a picture of real.
    What I am sure, scientist do not use these synthetic "pictures" as a pictures, they use them to support imaginantion about physical properties of theoretical objects. Imagine they have two diferent takes of the same data. They can look at them and discusss differencies, they identify problems in theories, suggest new teories. I think these pictures serves them more like metaphores.
    I do not think, the scientific value of the images are in images itself. Or I should write I hope in it.

  2. "How much different from our human selves would we have to be to directly perceive something that looks like the Black Hole Photo, and is it even physically imaginable that we could?"

    Look at the movie "Interstellar".
    You will know.