Jump to content
Bill Cotter

Some 1939-40 New York World's Fair images

Recommended Posts

Was Daylight Saving Time in effect?

It's a pleasant afternoon, 2:50 PM to be exact, at the Court of States.

court-of-states-pa.jpg

Share this post


Link to post
Share on other sites

Couldn't say - these weren't dated in any way, and there are some from both years of the Fair mixed in. If you can do your magic and tell us a date I will be duly impressed.

Share this post


Link to post
Share on other sites

Couldn't say - these weren't dated in any way, and there are some from both years of the Fair mixed in. If you can do your magic and tell us a date I will be duly impressed.

Well, the "bait" is duly noted, but I was hoping someone else would give it a try first... If you can send me a bigger scan, I can try to sort it out. Not knowing the year means more accuracy is needed.

[Edit: for you engineers reading this, more precision is needed, and accuracy too.]

Share this post


Link to post
Share on other sites

Got it

Share this post


Link to post
Share on other sites

OK, here's the first try at a date.

First of all, I noticed that these photos must have been taken with an adjustable-back camera, as the perspective "lean-in" of the buildings was approximately corrected. 008 missed slightly one way, and 004 the other way.

Lookig at what was available to measure, I decided to try 008.

7512112752_8b06c84f69_c.jpg

008 - Court of States towards Pennsylvania by old_tv_nut, on Flickr

First tempting objects to determine vertical were the flag poles - nice and big - but it turns out they are not necessarily vertical. Note the uneven spacing.

7512113086_b7b572bf20_o.jpg

008 flagpoles irregular by old_tv_nut, on Flickr

In this picture, there is a man in uniform at the left with a clear shadow. Since he is standing on the ground, it is not necessary to go through the procedure I used on the 1964 TIC-TOC picture ot determine how far the bottom of an object is from ground.

First moves are to correct the perspective so building edges are vertical, and then draw black perspective lines along various railings, which will meet in the vanishing point for the camera position. Once the vanishing point is determined, red perspective lines can be drawn to the person of interest and his shadow. Horizontal and vertical green lines can then be drawn to give measurement dimensions for height and shadow x-direction length.

7512113304_35d1137533_c.jpg

008 - xy by old_tv_nut, on Flickr

Y = 800 pixels; X = 746 pixels.

The shadow is not directly to the right, however, and is seen in perspective. To correct this, a "perspective crop" is done in Photoshop. The perpective crop procedure starts with adjusting a grid to match the red perspective lines.

7512112138_b008300e42_b.jpg

008 - perspective grid by old_tv_nut, on Flickr

When the grid matches the perspective, pushing Enter corrects the perspective so that the red lines are now vertical. This means that pespective of shadows on the ground is corrected to a direct overhead view. Now, the shadow lies on a circle, and the X and Z dimensons can be measured. This picture shows the upper half of the circle and the X-prime and Z-prime dimensions. If the perspective transform changed X prime to be different from X in the other picture, the Z-prime measurement could be corrected to Z. However, the transform conveniently changed only Z, so this correction was not needed.

7512116714_4baa0926dc_c.jpg

008 xprime zprime by old_tv_nut, on Flickr

Xprime = X = 746 pixels; Zprime = Z = 52 pixels.

Shadow length = the hypotenuse = 747.8 pixels - a small correction compared to X, but enough to throw the date off.

Knowing the height (746) and shadow length (747.8), the altitude of the sun can be calculated from the arctangent as 46.93 degrees.

Share this post


Link to post
Share on other sites

Part 2 - looking for the date.

This takes some trial and error with a sun-postion calculating spread sheet.

The question arises whether the clock showed daylight saving time. One source on the internet said that New York, Chicago, and a few other places maintained daylight time following WWI, running 7 months starting March 31.

It turns out sun altitude is too close to call for several dates, especially when trying to determine the year, which causes an equivalent 1/4 day offset (actually, there is an additional day offset due to leap year).

Candidates, assuming daylight saving:

Date -------------------Sun altitude at 2:50 pm, degrees

Sep 7 1939 ----------46.78

Sep 6, 1940----------46.87

Sep 6, 1939----------47.14

Sep 5, 1940----------47.23

May 3, 1940----------46.82

May 4, 1939----------46.89

May 4, 1940----------47.11

May 5, 1939----------47.18

If you assume the clock was showing standard time, the candidate dates are

May 20 - May 21, 1939 or May 19 - May 20, 1940; or

Aug 4 - Aug 5, 1939 or Aug 3 - Aug 4, 1940

An error of one pixel in the X or Y dimensions translates to 0.03 degrees in the sun altitude, so you can see the potential errors are too large to pick the year and probably represent a couple of days within in each year.

Share this post


Link to post
Share on other sites

We know it's 1939 due to the USSR Pavilion still being there.

From the clothing, I would guess later in the year rather than earlier. Of course, May could have been chilly. Now you need to pull some weather stats!

Share this post


Link to post
Share on other sites

Hi Bill,

Indeed. And, there is something just as scarce in that photo enlargement and it is not the camera guys! :)

Best wishes,

Eric

Share this post


Link to post
Share on other sites

Hi Randy :D

That is just one of the many light fixtures in the area - the scarce thing is the entrance sign to the Perisphere! Only thing scarcer are the two different exit signs for the Trylon/Perisphere; one interior and one exterior.

Best wishes,

Eric

Share this post


Link to post
Share on other sites

We know it's 1939 due to the USSR Pavilion still being there.

From the clothing, I would guess later in the year rather than earlier. Of course, May could have been chilly. Now you need to pull some weather stats!

Looking at the supposedly direct overhead view, I think I may have screwed up - the shadow should be about the same aspect ratio as a person, and it is much too thin. This means the Z dimension is not correct. Have to think about this some more.

Share this post


Link to post
Share on other sites

After a lot more thought, I can't see how to get accurate depth dimensions from the photo without knowing the dimensions of something in the picture, preferably a horizontal object. When you think about it, if every point in the scene was made proportionally further from the camera (or proportionally close), the picture would still be exactly the same. Unfortunately, there are no square tiles (or circular insets) on the walkway.

I think if someone knows the dimensions of the reflecting pool, that would help.

7512116636_e795cdd72b.jpg

008 - perspective crop by old_tv_nut, on Flickr

Share this post


Link to post
Share on other sites

Supposedly the original sphere (the one here is a copy) is 410 cm in diameter. Being built by the same man they well might be the same size.

Share this post


Link to post
Share on other sites

Reviewing my calculations, I found two errors not related to the problem

of measuring depth:

1) I got the arctangent upside down. The angle in the previous

calculation should have been ATN(746/747.8) = 44.93 degrees.

Second BIG mistake - I read the wrong spring month on the spreadsheet.

So, lets start again, and throw in an arbitrary correction prompted by

the shadow being too skinny in the original perspective transformation.

Suppose my mistaken value of Zprime is off by a factor of three.

Xprime = X = 746 pixels; Zprime = Zx3 = 156 pixels.

Shadow length = the hypotenuse = 762.1 pixels.

Knowing the height (746) and shadow length (747.8), the altitude of the

sun can be calculated as ATN (746/762.1) = 44.387 degrees.

Assuming Daylight Saving Time, the closest Spring date is March 27

(44.52 degrees). In fact, if the clock is one that moves a whole minute

at a time, ti could be closer to 51 minutes past the hour, at which time

the sun would be at 44.4 degrees. BUT - the fair didn't open until April

30, so the Spring date is not possible (this is where I made the mistake

in reading the spreadsheet before).

The closest Autumn date is September 13 (44.21 degrees altitude) [or if the clock is a minute slow, 44.45 degrees]. However, the temperatures are questionable for the clothing shown, with highs ranging from 69 to 85 degrees within +/- two days.

----

Suppose Daylight Saving was not in effect (at least not on that clock). Then, the closest Autumn date is August 15. This definitely does not match the clothing, as the high temps that week were between 84 and 90 degrees. The closest spring date would be May 5 (44.51 or 44.33 degrees altitude). On May 5, the high temperature was 79 degrees, but two days earlier it had been only 56 degrees, and three days earlier, 61 degrees.

So, if the clock was showing standard time, given the possible errors in estimating the shadow length, the picture might have been taken on May 3, 1939.

---

This is where the calculations stand now; if someone can find the dimensions of the pool, it may be possible to refine them.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


×