Hacker News new | past | comments | ask | show | jobs | submit login
Hockney–Falco thesis (wikipedia.org)
70 points by tosh on Dec 26, 2017 | hide | past | favorite | 20 comments

The film Tim’s Vermeer has a nice exposition of this theory, including a few good scenes with David Hockney himself. Great Look the intersections of art, technology, and personal obsession.


I wish people would be a lot more focused on Tim's experiment than Hockney's theory. Great documentary, and Tim is a great hacker and a participant in the community.

Tim's experiment was very convincing for me, a layman. Especially the fact that it reproduced details that the brain can't really process well enough to draw (e.g. the subtle changes of color of the wall) and the fact that the paintings exhibit optical defects such as lens blur, aberration, vignetting, etc, is pretty stong evidence to this theory.

Yeah, the problem with it is that while it's proof that you can do that with Vermeer's technology, there's no proof that Vermeer did it that way. And there are no artifacts like those lenses and mirrors found in illustrations of painting workshops or artifact collections, and no books or other documents that talk about that technique.

It would be awesome an art historian would look at a bunch of paintings for those defects, or if someone could work out how many portraits are 1:1 scale, like Tim's second "Claire with a Pearl Earring" experiment.

Yeah, very true. However, isn't the color aberrations in the paintings a dead giveaway that a lens was used? You can't really get aberrations any other way!

Yeah, I like that one. I also like the argument about the brightness gradient in the white wall being super accurate. But these sorts of arguments aren't considered strong by critics.

Now imagine that we found a half-burned-down painter's workshop in the countryside, with a bunch of artifacts. Or a woodcut or drawing of a painter's workshop with these devices in use. Those would be very convincing to a critic. But we don't have those.

Yeah, fair enough, I guess I just disagree with critics there because chromatic aberration is a very specific defect, and not just "random colors gone wrong". Art critics will know more than me on this, though, so they must have their reasons.

Hopefully we will find a workshop somewhere with a bunch of photographic equipment and crack this puzzle.

What's with the spam link? You have a really good comment history, so I'm thinking either your account has been penetrated or somehow your "yeah" was autocorrected to a bogus hyperlink, or you meant to link something else.

Thank you for the compliment about my comment history. There is an alternative explanation. A surfeit of eggnog. Luckily I didn't post this link: http://flaminglips.com/history/songs/the-yeah-yeah-yeah-song...

But we do have such images:



And their existence is part of critic's arguments against the Hockney thesis: many tools have been used in various ways at various times, and there's been little or no shame in acknowledging and documenting it.

To accept Hockney-Falco, you'd have to accept the opposite: that such devices were widespread amongst artists but kept secret or never mentioned. I think Occam's razor suggests that if such devices were commonly used, it wouldn't be so hard to find evidence of it.

Those aren't images of the tools that Tim used. And I totally agree that the tools used by Tim aren't found in middens or descriptions or drawings. Occam's Razor does apply, but, it leaves me wondering why The Music Lesson is so great. Or why the Girl with a Pearl Earring is 1:1.

I was first introduced to Hockney's theory in a New Yorker article from 2000. The full article isn't available online unless you have a subscription but you can find a summary of sorts here:


It looks like the article preceded publication of Hockney's book and originated at a time when he was still rehearsing and refining his argument.

His main antagonists in the article are the art historians who come off a bit like stuffy reactionaries. I was firmly in Hockney's camp by the time I had finished the article. It was one of those ideas that changed the way I look at the world slightly.

Never read the book. Tim's Vermeer was the coup de grace for me and I was gratified to see Hockney make a cameo in it.

It's a cupser-cool project, but it's sad that it only half-engages with its subject. Tim builds himself a replica of Vermeer's studio and then painstakingly commits a visual record of it to canvas using optical tools.

There's virtually no discussion of the design of Vermeer's paintings, which was elaborate to say the very least. I was so hopeful that time would apply his technique to some other subject, but it stays almost myopically focused on the rendering process.

This is an interesting idea, but it's impossible to know for sure if it's true or not. It's unfalsifiable, basically.

The reason -and it's a technical, not a philosophical reason- is that theories like this are really trying to determine a process from examples of its output, where the output in this case is some behaviour of a living organism (a human artist creating a painting). Unfortunately, for a sufficiently complex behaviour there may be any number of processes capable of reproducing it to the same degree of fidelity- "any number" as in an infinite many [1].

The best you can hope to do then, is to approximate the target process to some arbitrary measure of error. You can easily measure the error by comparing examples of the target process with examples of an approximator. However, you cannot prove that an approximator _is_ the target process- not even given 0 error in the reproduction, because finding one approximator is more of a hint that there are more approximators possible than any sort of proof that the approximator you found is the only process capable of reproducing those examples with that error.

So the problem with the Hockney-Falco idea is that there is no way to tell whether a work of art was created using an optical device, rather than without. Only that it could have. And that's really the strongest claim that one can legitimately make: "the Old Masters could have used optical devices to create their masterpieces". From there to "... and they actually did" is a big leap- of faith, not reason.


[1] There's maths on that- learnability results, particularly related to automata induction and so on. I'll dig out the refs if anyone is curious. But oh god, please don't :)

Well, you could prove it by digging up texts and commentaries from the time that mention these devices. Diary entries, etc.

It's unlikely that so many masters would use this technique without leaving some evidence of it behind. So, in the absence of that kind of evidence, I think it'd be okay to dismiss it as a primary force of innovation during the renaissance, especially if other techniques exist that were more readily discussed.

I love David Hockney, but Secret Knowledge is an extraordinary work that I’d recommend to everyone regardless of their interest in art criticism. Most people frame the work as using science to comment about art criticism but I think that’s backwards. Hockney, instead, uses the program of art criticism to evaluate scientific discovery (and progress). By the end I felt that science became slightly more humane.

I once attended a lecture by Stork that very convincingly disproved this thesis. But Hockey was not there to defend it.

At least it allowed a lot of non trivial math (computer vision).

> David Stork, an imaging scientist and former Stanford professor with a side career in computer-aided art analysis, was the major scientific critic of Hockney and Steadman a decade ago. One of his main counterarguments was that, using only a camera obscura, Vermeer would have had to paint upside down and the projected image would be too dim to be useful. Jenison figured out that using a second mirror solves both problems. So in his apparatus, the image is projected through the 4-inch lens onto a 7-inch concave mirror on the opposite wall, and then onto the 2-inch-by-4-inch mirror he’d have right in front of his face as he painted.


Stork discusses some of his criticisms of Hockney in this Youtube video: https://www.youtube.com/watch?v=rbL_Y-hZh20

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact