Four of the clock it was, so I as I guesse

August 5, 2009 12:57 PM

How can I estimate the time that a photo was taken?

I'm working on a rephotography project (the link is my profile website, if you want an idea of what I'm dealing with). I've gotten the hang of many of the common problems I encounter when duplicating photos, but one thing still gets me -- time of day. I try to estimate based on the angle of a shadow, but that usually only gets me within a three hour window, and I hate sitting around waiting for the sun to move.

I'm a good enough geometer that I can usually figure out what the angle of a shadow is relative to true north using Photoshop, a street map, and a protractor. I have the date each photo was taken. I know NYC is about 15 minutes off of true noon for the time zone, and I know the angle varies seasonally. How can I do the math?

posted by zvs to media & arts (11 comments total)

Helios is a (rather pricey) iPhone app for this. If you don't have an iPhone, perhaps you can find a similar calculator for another platform.
posted by kidbritish at 1:09 PM on August 5

You don't happen to have the exact time in the (digital) photo's meta data?
posted by oxit at 1:25 PM on August 5

Keep in mind: the same time of day on different days of the year will result in a variety of sun positions. See "AnnalLemma".
posted by catkins at 1:32 PM on August 5

You don't happen to have the exact time in the (digital) photo's meta data?

If you click through, you'll see that his source material is a bit older than that.

(I don't have a good answer to the OP, but in case someone's interested, here's a tutorial that cover the basics.)
posted by effbot at 1:36 PM on August 5

(interested in this kind of photography, that is.)
posted by effbot at 1:55 PM on August 5

Google Sketchup can display sun angles based on the date and time. If you can roughly model the objects that produce large shadows you could probably spend a few minutes adjusting the date and time sliders while observing the scene to get a fairly good estimate.
posted by odinsdream at 2:00 PM on August 5

Echoing catkins, without an approximate time of year, the time of day still won't help too much. I think the images on your site look pretty good and in many cases it doesn't matter because there is no reference point between the old and new shots to compare with. Stick with trying to match the focal length and POV and make a good picture from that.
posted by JJ86 at 2:06 PM on August 5

To clarify, I know the time of year. Usually down to the day.
posted by zvs at 2:32 PM on August 5

Isn't this basically just a sundial problem? In which case, a sundial calculator like this one would help solve your problem. NYC is approximately 40.717 North Latitude, and plugging that in will give you time of day angles, with the (poorly formatted) table giving you adjustments for time of year. You'd also need to correct for longitudinal differences and Savings Time.

I'm sure there are other, better sundial calculators out there, that's just the first one I found that looked ok.
posted by fings at 3:03 PM on August 5

Helios looks great, but I don't have an iphone. That sundial table will totally work well enough for my purposes. Thanks!
posted by zvs at 4:30 PM on August 5

Can this be of any help?

sun from new york 4pm 8 august 2004
posted by hAndrew at 10:01 PM on August 9