Wednesday, January 24, 2007

Silvers and Gold

Over the last few days, I've been having a very enjoyable e-mail exchange with Rob Silvers. If you don't know Rob, he's the guy that invented photomosaics. Well, I had a more-or-less brilliant idea of having next year's Skyline background be a photomosaic of book covers. I told Rob I could provide more than 12,000 cover images. If it's possible to tell that someone is drooling over email... I could tell. Apparently he usually only gets a few hundred to work with. Anyway, he seems to be a very likeable guy — the enthusiasm is palpable. He even offered me a very generous discount on my project because we're a smaller company.

How much would it cost to do what I had in mind? Well, I'll keep the exact numbers confidential, but I'll let slip that he very generously knocked it down from five figures to four. We'll see what the bean-counters have to say, but I have my doubts, to say the least!


Money, money, money, money.
(Click for a larger view. It's worthwhile.)

3 comments:

Jonadab said...

Other than a lack of experience with the technique, or a lack of discretionary time, what prevents you from doing one on your own?

Granted, it would be much more elegant to write a program to automate the process, and that means familiarity with graphics-manipulation programming techniques, but I'm thinking it *could* also be done mostly by hand, in a pinch.

Andy said...

Well, there's the fact that I'm not a masochist... and the fact that I don't want to be sued for copyright/patent infringement...

But just for giggles, let's consider: Last year's design was 144 x 60" and ran into the two-gigabyte file size limit several times. At the intended scale, that's 5,760 cover images. We'll round up to 6,000. Let's say that I could find and place a book cover per minute, choosing the correct shade each time to make a perfect match. That's 100 hours. Hmm. Then again, I do have about two years...

I could do it the easy way, which is to just do a scatter of cover images, and then some sort of type treatment over the top, but my initial experiments there indicate that legibility suffers considerably.

Jonadab said...

Regarding copyright infringement, I am absolutely certain that the mere idea of making a big image out of a bunch of small ones is not copyrightable. Copyright protects a specific expression of an idea, i.e., particular words, or a particular picture. Merely redrawing an image does not get you out of copyright trouble if the image you're redrawing is copyright, so creating another George Washington that looks like theirs would be trouble, maybe even if you used book covers instead of money, but you'd presumably be creating a whole new image, *not* a copy of any of their images, and that's altogether another thing. Copyright just doesn't apply there.

As for patent, that protects the method or means by which something is accomplished. You could potentially run into patent trouble if you write software to do it, and your software happens to do it approximately the same way as their software, but again, patent cannot protect the mere concept of making a big image out of little ones.

However, I don't have a good answer for the "I'm not a massochist, and it would take too long" argument. You're probably right. I don't think you could place a book cover per minute, and do it well, not by hand, and after thirty hours or so, you'd probably be extremely tired of the project.

I could suggest ways in which a computer program might select and place them (my initial thoughts are in terms of matching greatly scaled-down versions of the covers against small sections of the source image; once you've decided where to put each cover, the full-res vesion can be substituted), but even assuming writing a custom program was something you wouldn't mind undertaking, it could potentially get you into patent trouble, if it works in the same way as theirs. Not knowing how theirs works, I wouldn't suggest undertaking that without checking their patent and making sure to avoid its technique.

Ah, well. It was worth thinking about.