Space Shuttle Over Kansas: A Discussion

[DaveC426913]

A) The object could be in the distance, greater than S1. I mentioned this possibility (that it is further away from the tree) only in order to disprove and rule out that possibility. If it were true, the con-trails in the far distant background show us the asymptotic limit of the radius of confusion on the far end, which is lower than the radius of confusion of the object.

The object and the contrails are effectively at infinity. They will have the same level of focus.

You have not demonstrated how out-of-focus the contrails are, you are just taking it as a given.

C) The object could be in the foreground. This coincides with all visual evidence.

It coincides with one piece of evidence if you interpret it that way. Your weapon is pointing at your foot and the safety is off...



One of most likely sources of the blur is camera-movement. That would explain why everything in the picture has (more or less) the same amount of blur and why the blur has quite sharp edges (look at the branches of the trees).

 

[RonL]

I'm not quite following the points you raised, however, the fact that it is missing a vertical tail fin seems to definitively rule out the perspective I suggested.

After taking another look at the original image, I realized that it's absolutely impossible for this to be a plane of any kind at all. Based on the focus of the tree in comparison to the object it's clear that the object is significantly nearer to the camera than the tree, and that it's size is therefore quite small.

Then taking another look at the image, I'm almost positive that this is a picture of a bird. There is a downward curve in the front characteristic of a beak, and a flat tail in the back with no vertical fin indicative of a bird's tail feathers, and the odd structure over the torso is easily explained by motion blur between the flapping wings.

The bright spots aren't lights, they are reflections. Bird feathers can be quite reflective, especially when wet.

http://img5.imageshack.us/img5/3016/crowp.jpg

[/URL]

I had to scan the entire thread to see if anyone else would mention a bird, this is a sight I see quite often in my back yard, as birds glide through the trees making an approach to land near the water pans, and my pool, which is now a fish pond.

 

[junglebeast]

The object and the contrails are effectively at infinity. They will have the same level of focus.

The con-trails are effectively at infinity (this was part of my argument)...the object is not.

You have not demonstrated how out-of-focus the contrails are, you are just taking it as a given.

I did not explicitly point that out because I thought it was obvious. If you zoom in on the two regions, it is evident that the object has a blurring radius that is larger than the maximum resolution of the contrail (see figure below). This shows that the object cannot possibly be at infinity with the con-trail as you suggest.

http://img10.imageshack.us/img10/1279/focus.png
(and now with altered levels just to show that the foreground object does indeed blur out to that amount)
http://img16.imageshack.us/img16/6537/focusb.png

Indeed, this is the basis of my argument...the con-trail being effectively at infinity shows the maximum radius of blurring that is possible to occur for an object due to being farther away than S1. Therefore, since the foreground object has significantly more blurring than the contrail, it must be closer than S1 and closer than the tree.

 

[DaveC426913]

You are putting far too much faith in your analysis of blurring.

 

[junglebeast]

You are putting far too much faith in your analysis of blurring.

Blurring can be quite a reliable depth cue. It may not have high precision, but it is still reliable. It is well known that blur analysis is used as a depth cue in humans (on several levels). We even have computer software that can make 3D reconstructions based on blur analysis.

If you took a photograph of 2 balls that you knew were the same size, and you measured one of them to have a radius of 5 pixels and the other to have a radius of 100 pixels, you would take it as proof that the larger one is closer to the camera. That is the same level of confidence that can be drawn from the blurring in this image, the only difference is that we don't need to know how large the objects are to do the analysis.

My measurements may be crude, but I am only using them to show relative depth changes. In this case, all I needed to show to prove my point was that one circle is not the same size as the other by a statistically significant amount. In this case it's about 4 or 5 times larger, and the number of pixels under the area are effectively the number of measurements that corroborate that determination. So why are you so reluctant to believe it?

 

[DaveC426913]

My measurements may be crude, but I am only using them to show relative depth changes. In this case, all I needed to show to prove my point was that one circle is not the same size as the other by a statistically significant amount. In this case it's about 4 or 5 times larger, and the number of pixels under the area are effectively the number of measurements that corroborate that determination. So why are you so reluctant to believe it?

The trouble is, they're not the same type of object. They couldn't be more different. One is black-against-light-sky, solid, opaque and sharp-edged, the other is white-against-light-sky, ephemeral, transparent and diffuse.


Dark objects and light objects do not blur the same way (light and lack-of-light are not two sides of the same coin).

 

[junglebeast]

The trouble is, they're not the same type of object. They couldn't be more different. One is black-against-light-sky, solid, opaque and sharp-edged, the other is white-against-light-sky, ephemeral, transparent and diffuse.

Dark objects and light objects do not blur the same way (light and lack-of-light are not two sides of the same coin).

Blurring due to being out of focus is a purely geometrical effect caused by the aperture not being a perfect pinhole, and white objects do blur exactly the same way as black objects.

If anything, the con-trail having a softer edge in real life only strengthens my argument that the con-trail is more in-focus than the blob, because despite having a softer edge in real life, it still appears crisper in the image than the blob.

As for being transparent...it may be translucent up close but from afar it appears opaque. The only relevance that this would have on the analysis is that if the contrast from the background is too subtle, it would not be possible to determine the extent of the object's blur radius. In this case, the edge is distinctly visible, so it's not relevant.