15 year old ghost mystery possibly solved

[richardm]

I stand corrected about the "cap curls". Comparing the features side-by-side like that, and sized to match, they do appear to be the same image after all. Good work, Orthoptera!

Seconding this. I was a bit dubious about the quality of the match, but this spells it out perfectly. Nice work

 

[Stray Cat]

Sorry, I still disagree. Though the images of the girl are remarkably similar, they are not an exact match. This is as close as I can get and some of the details are still jumping around a bit. I have mostly concentrated on lining up the creases in the bonnet and making sure the belt is as lined up as possible.

 

[Pythra]

Stray cat, what struck me about your animation was the way the belt and collar of the dress line up absolutely spot-on.

I do think there are tiny differences in the face, but maybe those can be explained by the way the source image was enlarged and possibly slightly distorted when it was added to the photograph. I wouldn't say I'm 100% certain it's a match, but that higher quality image Matty found is very compelling evidence. I think it's the same image.

 

[JcR]

Is the sidewalk a ghost to? Sorry I mean the wall to.

 

[P.J. Denyer]

Stray Cat- great animation, funny thing is it had the opposite effect on me, I thought they were similar but not the same image until I saw your comparison (and Orthoptera's too), now I'm 100% convinced it is the same.

 

[Skeptical Greg]

You can see more of the girl's face in the ghost picture. Not a match, unless her face was retouched in that picture.

Yeah, no way that happened .. Must be a real ghost ...

 

[Stray Cat]

Stray Cat- great animation, funny thing is it had the opposite effect on me, I thought they were similar but not the same image until I saw your comparison (and Orthoptera's too), now I'm 100% convinced it is the same.

Well I did say it's "remarkably similar". But it's certainly not 100%
If I get some time, I'll do an enlargement to highlight the differences.

 

[Pythra]

Ignoring the face for a moment, just look at the belt.



The shape matches perfectly, from where it does up in the middle to the outline on the left, where it's concealed by the girl's sleeve. Above the belt on the left is a wedge-shaped shadow, which also matches. Even if two girls were dressed identically, or even the same girl was pictured twice, you'd have a hard time replicating such uncanny similarities.

In the 'ghost' picture, though it's hard to say for sure, it also appears that the belt overlaps the railing in front of the girl. Not that anyone here needs convincing that the photo is fake, but that is only possible if the image of the girl was added on top of the original image.

 

[richardm]

There's an excellent link here that lets you fade and zoom between the two pictures. It makes it pretty clear that it's the same picture.

Edit: The fading is useful, I think, because it makes the changes in contrast easier to work past versus a flick from one to the other (Still, nice work lining them up, Stray Cat - but even before trying the link above I thought that your animation made it look like a good match!).

 

[Stray Cat]

Hi Richard
Originally I did the animation with a slow fade, but decided it allowed the eyes to adjust to the slight variations and disguised them to an extent.

 

[Pythra]

Yeah, I don't really like fades (as compelling as it is) because the details in the transitional images make you see matching details where there are none.

Although at this point I am completely convinced it's the same image of a girl.

 

[Careyp74]

When you take a photo of another photo, you are going to lose a lot of detail. When you take a photo of an old post card, you are going to get even less of the original detail.

The newly discovered copy of that postcard shows how losing details in contrast makes it appear that the angles are different.

There are so many of the same details found in Stray Cat's comparison, that it is proof that the post card image was used in the fire picture.

Any details that are not the same, are contributions from the fire picture. The rail for example. ParrotPirate, were you joking? There is a rail on the building that is on fire. That is why you see it in the picture. There is no fire in the street scene, either. Point?

 

[Lucky]

Stray Cat, you've done a pretty good job of proving that it is the same image!

What details are you claiming don't line up? Using your animation I checked the base of her nose, the edges of the eyes, the three central frills on her cap, the bottom of the tie (if that's what it is), the folds of her dress round her waist and the shadow caused by these folds, and (following Pythra's observation) the full outline of the wedge-shaped shadow next to her right sleeve. The positions all match almost perfectly, except for the tip of the cap frill to her left (which is very blurred in the copy).

What is the probability, do you think, of all these matches happening independently - head and body angles, frills, folds and shadows (including details such as the waist shadow and the wedge-shaped shadow above being cut off in the same place by the right sleeve)?

I'd say the correspondence is so striking that it has to be the same image (or another one taken within a few seconds, which seems far less likely).

Pythra, it's not a belt - it's a shadow caused by her blouse overhanging her skirt where it's tucked in at the waist. It's obvious that the image of the child can't be on top of the railings.

 

[sophia8]

I thought this one had been solved ages ago.
A couple of years after the photo came out, there was a TV show which demonstrated that the image of the girl must have been photographed from a TV screen - the image showed telltale horizontal lines when blown up. The show mocked up a near-identical photo, using photography and imaging equipment the photographer was known to have.

 

[Pythra]

Pythra, it's not a belt - it's a shadow caused by her blouse overhanging her skirt where it's tucked in at the waist. It's obvious that the image of the child can't be on top of the railings.

Oh, sorry, you're right about it being a shadow! Optical illusion.

If you're saying the image of the child isn't double-exposed on top of the railing, how do you think the forgery was made? What I'm referring to, though it's impossible to say conclusively, is where the shadow crosses the railing in the ghost image. There's a darker stripe on the railing at that point, which appears to line up with the shadow.

 

[LTC8K6]

It's the same image. Can't understand why anyone would doubt it.

 

[richardm]

Hi Richard
Originally I did the animation with a slow fade, but decided it allowed the eyes to adjust to the slight variations and disguised them to an extent.

Yeah, I don't really like fades (as compelling as it is) because the details in the transitional images make you see matching details where there are none.

The point is a good one. I liked it anyway, not least for that very reason - because it disguised the minor variations and drew the eye to the considerable similiarities.

 

[Hannibal]

Looking at the more in-depth image comparison it has swayed me - probably* the same image.



*I rarely admit to 100% volte-faces!

 

[Stray Cat]

It's the same image. Can't understand why anyone would doubt it.

And I don't understand how any sceptic could be so certain at this point.
Certainty is a step too far.

To clarify, I'm not saying it isn't, just that there are differences. When time allows I'll have a closer look at both pictures and compare them pixel by pixel.

 

[TheSkepticCanuck]

Well, I am certainly convinced. Definitely the same image in both pictures.