Old bicycling photo - "real" or staged?

Factual Questions
181

You’re contradicting yourself here. If they’re using fixed lights, then the exposure time required for this photograph is long. Film with an ISO of 25 or so, aperture of f11 or so at widest, and light that is at best a couple stops lower than full sunlight. That gives us a shutter speed in the vicinity of 1/10s at the fastest. “On the order of seconds” is indeed more likely. If we grant that they’re moving 3km/h (which imo is much, much too slow to maintain their angle/position on the track, but whatever), that’s .83m/s, or 8cm over the course of a 1/10s exposure. That would make for unmistakable motion blur. A more reasonable supposition of 15km/h and 1/2s exposure would have them traveling 2 metres during the shot!

If it’s live action the lighting has to be brighter than the sun, given the lack of motion blur, the film speeds available at the time, and the observed depth of field. Turn-of-the-century stage lighting isn’t brighter than the sun. The only possible way it’s live action is if someone did one of the elaborate flash powder setups that’s been hypothesized.

182

I’ve combined all 3 photos here for better comparison (I’ve also increased the contrast in the last two). Viewed this way it doesn’t appear that the guy on the floor has moved, the last photo was just taken from the end. It appears that all three were taken at the same time if it was an action shot. I was originally in the real camp but am slowly edging toward the staged camp.

Another thing… the shadow of the cage on the right side seems to change on pics 2 and 3.

Also, doesn’t it seem that the transition from floor to cage is a bit abrupt?

183

It would be if it was a ramp that you approached head on, but if you’re riding in a circle and thus approach the edge of the ramp at a very slight grazing angle, the effective angle you’re riding up onto is also slight.

184

To me, it looked like the nearest rider in picture 3 was moved relative to the other two pictures, but it was a little hard to be sure. I decided to see if it could be shown objectively. I read the three images into Matlab, and used some code I have to select points in the image and get their coordinates. I selected the heads of the four bike riders, and other points I could identify in all three pictures, including some in front of the rider and some behind them. For picture 3, it was hard to identify points in the scenery. See Picture 1, Picture 2, and Picture 3. The green circles are the selected points, and the lines connect them in the same order in all pictures.

I then took the third picture as my “standard” picture, because it was a mostly horizontal direction of view. I chose the point midway between the heads of the riders on the left and right as my origin. Using all three pictures, I estimated the position of the selected points in the third dimension for picture 3. Having the circular track helped a lot here.

I then rotated the points from picture 3 to match up as well as I could with picture 1 and with picture 2. I had four parameters: three angles and an overall scale factor. I didn’t apply a shift, since my origin is near the scene center in all three pictures. I also didn’t take into account the distance from scene center to the camera.

Comparing the points in pictures 2 and 3, the alignment is very good, except for the point for the near rider. The red points are the rotated points from picture three, and the black points are from picture 2, so the near rider in picture 3 is to the left of his position in picture 2. I marked the near rider and scenery points to make it a little easier to figure out what you’re seeing.

Comparing the points in pictures 1 and 3, the alignment is also good, except for the point for the near rider and for the three points on the front of the track. I suspect the misalignment for the three points in front of the track is mostly due to the camera being much closer to the scene. I’m essentially assuming the cameras are all “very far” from the scene. Again the red points are the rotated points from picture three, and the black points are from picture 2, and the near rider in picture 3 is to the left of his position in picture 2.

I’d like to make a comparison between pictures 1 and 2, but that’s a little complicated, since I have to use the out-of-page coordinates from picture 3. Also, this approach could be improved by taking into account camera distance, but that’s even more complicated. It isn’t really necessary, though. For both comparisons, the heads of the other three riders all align well, but the near rider doesn’t. The near rider in the third picture is in a different location than in the other two pictures.

185

Well, the near rider is the one that seems the most realistic. He’s not doing anything strange.

186

Well, if the whole thing - stage, track, cameras, lighting, backdrop - were on a turntable centred on the middle of the track, so you could spin everything and the bikes would stay in place…

Or you could build a mocked-up stage outside and photograph on a bright but overcast day, with big mirrors to create the multiple shadows…

Or you could set up pots of flash powder with big fulminate detonators for lighting, and set them off simultaneously with an electric blasting machine (invented 1878). Confined flash with fulminate detonation may be fast enough that you can use low light and an open shutter…

187

Thanks for the enhanced comparisons.

The middle-of-the-track rider has moved between 1-2 and 3; he’s gone from the 6-o’clock position in 1-2 to the 9-o’clock position in 3, much close to the rider in the cammo jacket. You can see his bike position through the track in 3. He’s really close to the left rider, which he obviously is not in the first two. Enlarging the pics, in 3 it seems you can see the cross-brace and frame of floor-guy’s bike and he’s actually standing to one side of it.

Meanwhile, the light has gone out in pic3 on the lower right, so the stage under the riders is darker. Also, the back rider puts an obvious shadow on the backdrop in 1-2, but the lighting is changed in 3, no shadow from 12-oclock rider and the fringe on the second top curtain is not as prominent. yet in all this change, the 3 riders are perfectly in the same place.

You can gauge the angles by where you see the braces for the track. In pic1 You must be in an opera box on the side (see also side view of opposite side curtains), pic 2 in the balcony (by the elevation and more head on) and 3 - lower down - in the seating ground level, toward the back.

My guess is pic 3 was taken, the lighting was updated a bit and the middle guy moved more into his quarrter of the track then pics 1,2 were taken.

Staging. Also note the perfect quarter positions; the very back rider’s bike is tied to the top track brace so he’s more visible from the audience level in the picture, the others at 3 oclock and 9oclock tied at the bottom brace rung. The ground-floor rider is low enough to not block the view of the back rider… nicely staged.

188

Here’s another nail in the “action shot” coffin. I was trying to determine where the camera that took picture 1 would be in the other two shots. Looking at picture 1, the top of the head of the rider on the left blocks the bottom of the front wheel of the far rider. Looking at picture 2, you can draw a line between those two locations, and (assuming the two pictures were taken at the same time) the camera had to be on that line, to the left of the left rider, when picture 1 is taken.

Next, in picture 1, the front of the neck line of the near rider’s shirt is in line with almost the bottom of the track, at some point a little in front of the rider on the right. I measured about 6 percent of the way up from the bottom. In picture 2, you can fit an oval to the base of the track (I used PowerPoint) and estimate where an oval 6 percent up the side of the track would be. The picture 1 camera position would have to be on a line between the neck line of the near rider’s shirt and some point on this oval. But these possible lines never cross the line that was found in the first paragraph. There’s no location in picture 2 (including beyond the edges of the picture) where the camera that took picture 1 can be. Thus picture 1 and picture 2 are also taken at different times.

This, plus the movement of the near rider in the third picture, means all three pictures were taken at different times, but with the three riders on the track in exactly the same place in the same pose. It’s simply not plausible that this would occur if these were true action shots.

189

Personally, my impression was that the back three riders (at 12, 3, and 9 if you consider the back wall to be the “top”) were wired up, with the one closest to the viewer (at 6) having his bike free-standing on the floor. Which is why he’s the only one that moves.

190

A bit off topic, but in reading this thread I’ve been wondering what the act was and who they were. A bit of googling led me to a few newspaper articles of the period that refer to a ‘Cycle Whirl’.

From the February 2, 1902 ‘Sunday Globe’ - Washington DC newspaper ad (PDF) promotes:

*Keith’s Sensational Cycle Whirl

Not three, but four of the swiftest riders in America whirling 'round with such rapidity they appear as but a single rider.*

Another from The Daily Ardmoreite, October 19, 1903

Free! Free!! Free!!! The sensational cycle whirl, the whirling wheelers. Sensational, nerve thrilling and dexterious (sic) feats of skill. Double cross and zig-zag with rapid whirl at break-neck speed. Fancy trick and rough riders. Single and double sweve (sic) and sway on one wheel, attempted by no other bicyclist. Smallest trick in the world, only 70 foot diameter at top. Embankment 80 degrees. The most sensational act of the age. Will be seen in Ardmore October 26 to 31.

Finally, here is an additional photo of a different track than the one in this thread:

http://museumofnyc.doetech.net/detlobjps.cfm?ParentListID=0&ObjectID=42595&rec_num=20&From=obj_key.cfm#42

191

I suspect you’re joking, but I don’t believe they would. If the track rotates and the riders are stationary WRT the ground, there will be no centrifugal force to keep them in place up the side of the track, leaning over.

Rotation is not purely relative like linear motion is.

192

He’s not saying the bikes are stationary with respect to the ground, but stationary with respect to the track/stage/cameras.

Moot, now, with ZenBeam’s excellent work, of course.

193

Actually, I think he was riffing on the treadmill thing.

194

Blinks…

You’re correct of course, but that’s not what I was visualising! The riders should be stationary with respect to the track but rotating with respect to the ground, along with the stage, camera, backdrop etc.

On a pyro forum I enquired as to whether confined flash powder could give a much shorter duration flash, to allow an open-shutter technique to be used. Apparently it can, in fact flash powder can detonate, so you could probably get sub-millisecond flashes. The downside being that you’d get a hell of a bang and a bit of a shockwave. The device would basically be a pyrotechnic maroon.

Someone on that forum introduced me to the Yablochkov candle as a potential source of illumination, which I thought was pretty cool for its name alone…

195

I see what you mean now - yes.

196

Another photo
http://museumofnyc.doetech.net/image.cfm?ImageKey=14626&ImageFile=/VoyagerImages/Z001/Z00108/Z0010863.jpg&TableKey=OBJECT:41125
http://museumofnyc.doetech.net/srchobj.cfm