Finding the angle that a photo was taken from

[William W]

. . . The street (Mill St.) in question doesn't have an appreciable incline . . . GDHS0249 shows the buildings on the right. Etc . . .

Thanks for the extra reference photos, very useful. Taking all the images into account, and regarding image one that you posted here, it looks like the footpaths either side of the road are not at the same ELEVATION, maybe 12 ~24 inches difference.

Assuming the ladder theory is correct, and taking into account all the reference images, it looks to me that there is a slight DECLINE in the ROADWAY between the doorway area of FOREGROUND building LH side of image and where the ladder would have been placed. This concurs with my counting bricks earlier.

My guess is that ground level where the ladder was placed would have been about 3’ lower than the ground level of the doorway of the Foreground LH Building – ergo the “Camera Elevation” by ‘counting bricks method’ would have been about 17’.

Working backwards – to get to an elevation of around seventeen feet, a bloke 5’9” could use a 12, 14 or 16 ft double step ladder… and I bet there were a few of those around that street at the time and it doesn’t look as though it was an heavy car traffic area or that the local Bobby would have any objection.

Of course the roadway may have been resurfaced and today's roadway might not have the same undulations as what appears in the images made decades earlier.

***

I think that if the image was lifted from a ‘post card’, then it would be even more likely that it was made with a 6x9 Bellows 620 Roll Film Camera.

Interesting website, thanks

WW

Lots of people like trains – thanks Ian

[loosecanon]

It is worth remembering that it is only since 1970 that we have very wide angle lenses. Most archive images would have a 35mm equivalent of 35mm to 135mm. Often images are from upper floors of buildings long gone. Sometimes from buses.

To take images from higher up now I use a heavy duty light stand, can give a 5metre height, and a very different perspective.

[pnodrog]

Years ago (16?) at a photographic conference I attended a lecture by a Professor of Forensic Photography that happened to be in NZ to give some training to our police photographers. It was a fascinating lecture and he had some interesting stories related to crime scenes and contrived photographs used to try and defraud insurance companies.

He mentioned he had a program that helped in the determination of viewpoint, angle of lens and measurements of both distance and size of objects within the photograph. From memory he said it was most accurate he could visit the site and take some reference measurements. I understand most police photographers will know about or have a similar programs. I suggest you contact contact your local police photographer and get his/her help.

[george013]

Years ago (16?) at a photographic conference I attended a lecture by a Professor of Forensic Photography that happened to be in NZ to give some training to our police photographers. It was a fascinating lecture and he had some interesting stories related to crime scenes and contrived photographs used to try and defraud insurance companies.

He mentioned he had a program that helped in the determination of viewpoint, angle of lens and measurements of both distance and size of objects within the photograph. From memory he said it was most accurate he could visit the site and take some reference measurements. I understand most police photographers will know about or have a similar programs. I suggest you contact contact your local police photographer and get his/her help.

That's what I meant. It's all gonio and in the picture. If you know one real distance or size the others can be calculated.
Not by me.
George

[Saorsa]

Here is a quick way to estimate.

I took your photograph and added lines projecting things that I would expect to be level into the photograph. Here is the result.



There are a lot of ways that we could improve the precision of the answer but I believe that this is accurate enough to show a position just above the level of the window shown near the convergance. Since the street seems to curve and a corresponding window on another building might be a bit lower it looks like it was taken from a first floor window or from a ladder on the pavement. Perhaps by a window cleaner who was also a photo bug? Or a documentary photographer who used a ladder to get a more desirable perspective.

[george013]

Here is a quick way to estimate.

I took your photograph and added lines projecting things that I would expect to be level into the photograph. Here is the result.



There are a lot of ways that we could improve the precision of the answer but I believe that this is accurate enough to show a position just above the level of the window shown near the convergance. Since the street seems to curve and a corresponding window on another building might be a bit lower it looks like it was taken from a first floor window or from a ladder on the pavement. Perhaps by a window cleaner who was also a photo bug? Or a documentary photographer who used a ladder to get a more desirable perspective.

What do you mean with level????

George

[DavidAllen]

We did wonder if a bus had been involved

[DavidAllen]

Years ago (16?) at a photographic conference I attended a lecture by a Professor of Forensic Photography that happened to be in NZ to give some training to our police photographers. It was a fascinating lecture and he had some interesting stories related to crime scenes and contrived photographs used to try and defraud insurance companies.

He mentioned he had a program that helped in the determination of viewpoint, angle of lens and measurements of both distance and size of objects within the photograph. From memory he said it was most accurate he could visit the site and take some reference measurements. I understand most police photographers will know about or have a similar programs. I suggest you contact contact your local police photographer and get his/her help.

Interesting, I may explore that.

[DavidAllen]

Here is a quick way to estimate.

I took your photograph and added lines projecting things that I would expect to be level into the photograph. Here is the result.



There are a lot of ways that we could improve the precision of the answer but I believe that this is accurate enough to show a position just above the level of the window shown near the convergance. Since the street seems to curve and a corresponding window on another building might be a bit lower it looks like it was taken from a first floor window or from a ladder on the pavement. Perhaps by a window cleaner who was also a photo bug? Or a documentary photographer who used a ladder to get a more desirable perspective.

Wow! That's incredible. Thanks for your efforts.

[Saorsa]

What do you mean with level????

George

In my first use above the picture I meant perpendicular to a plumb line.

In the second use I meant the height of the top of the window on the far building on the left hand side of the street.

Here is a crop of the area of convergence.




These don't converge tightly for several reasons, lens distortion could be one reason but the greatest contributor is probably that I arbitrarily selected things that I expected would be perpendicular to a plumb line. Since I'm working from a contrasty photograph of buildings which may have settled with lens induced distortion it comes pretty close but not perfectly. Since I was using a relatively crude drawing tool with wide lines it's also likely that I didn't follow the line I selected exactly.

Oh, one more thing to introduce, these may well have been taken with a bellows camera which allowed some latitude in terms of swings and tilts to the lensboard. This capability was often present even in modestly priced cameras.

[Saorsa]

The first wasn't too difficult but the second was more problematic. I had to expand the canvas size to provide a background for the perspective lines because they are running out of the picture. Here is the result.



The line of the guttering on the longest wall is about as close as I could guess. The line of the roof ridge is slightly downward and the line of the bottom of the windows is slightly upward.

[dem]

This is what Mill Street looks like on Google maps:



The parallel lines from building on the left and those from the building on the right will converge at the same point only if the two buildings are parallel to each other. Which they are. If the camera was levelled perfectly (which it was), this convergence point will be at the same height as the camera.



According to Google Maps, the length of the house on the left is about 16.7 meters. The height of the house in the image appears smaller further away from us but it is obviously the same in real life. Since linear magnification is inversely proportional to camera-subject distance, we can work out the position of the camera.



x/(x+16.7) = 137/381

from which x=9.4 meters.



* the ladder is not shown

[Tronhard]

Your project reminded me of a set of images take in London during WWII and merged with ones taken recently. Here is the link, I think it's very impressive and they seem to have got the location and perspective bang on...

http://www.theatlantic.com/photo/201...graphs/481851/

[Cantab]

Your project reminded me of a set of images take in London during WWII and merged with ones taken recently. Here is the link, I think it's very impressive and they seem to have got the location and perspective bang on...

http://www.theatlantic.com/photo/201...graphs/481851/

Trev, that's a fascinating link. Thanks for posting it.

[Saorsa]

I just noticed this was Gamlingay!!! I've been through there visiting friends in Great Gransden. I was stationed in Bedforshire when I was in the USAF in the 60s.

[george013]

This is what Mill Street looks like on Google maps:



The parallel lines from building on the left and those from the building on the right will converge at the same point only if the two buildings are parallel to each other. Which they are. If the camera was levelled perfectly (which it was), this convergence point will be at the same height as the camera.



According to Google Maps, the length of the house on the left is about 16.7 meters. The height of the house in the image appears smaller further away from us but it is obviously the same in real life. Since linear magnification is inversely proportional to camera-subject distance, we can work out the position of the camera.



x/(x+16.7) = 137/381

from which x=9.4 meters.



* the ladder is not shown

And from those figures it must be possible to calculate the angle of view.
And then there's what I call the perspective in the picture: the suggestion of the third dimension in a 2 dimensional plane. The wider the aov, the more change in height, the narrower the aov, the less change in heigth or flattened.

You know the first uncensured pictures of Amsterdam where from a russian sattelite. There where some military objects in the city, marine, and on streetmaps they where represented as water. Now we've Google.

George

[dem]

And from those figures it must be possible to calculate the angle of view.

Not quite. You need to know a dimension of something perpendicular to the line of view, a height or a length but not the depth, at a known distance form the camera. Say the height of the house on the left, the line of 381 pixels, corresponds to about 5 meters and we think that it is about 9.4 meters away from the camera. The width of the whole image is 700 pixels which corresponds to 5 meters * 700/381 = 9.2 meters across the width of the image at 9.4 meters away from the camera.

So the width of the image is about the same as the camera-subject distance.

This is a field of view typical for a 35 mm full frame equivalent (any X100 shooters here?), about 53 degrees horizontal.

[William W]

. . . You need to know a dimension of something perpendicular to the line of view, a height or a length but not the depth, at a known distance form the camera. Say the height of the house on the left, the line of 381 pixels, corresponds to about 5 meters and we think that it is about 9.4 meters away from the camera. . .

Hi Dem,

"Counting Bricks" gets me 70 bricks for the 381 pixel line of the LH Building.

Bricks and Mortar in those days - that's about 6.2~6.5mtrs for your blue line.

What do you reckon?

WW

[george013]

Hi Dem,

"Counting Bricks" gets me 70 bricks for the 381 pixel line of the LH Building.

Bricks and Mortar in those days - that's about 6.2~6.5mtrs for your blue line.

What do you reckon?

WW

I think too 5m is on the low site for 2 floors and a separation, and the ground floor being a shop. And the length of the house not being there anymore. I don't know the sizes of English bricks used that time, but the Dutch are smaller, especial the older bricks.
More important is the way of thinking and calculating.

George

[DavidAllen]

I'm getting lost now. Are we going for leaning out of a window or using a ladder in the road?