Strange streak discussion: 2004 Dec 7 APOD

[victorengel]

I think you missed my point. What I was suggesting was to get two data sets like this.

(click image for larger version)

One would be taken from data in the putative wing trail. The other would be taken just outside that range. The curve I want to see is a difference between the two.

[victorengel]

Hi I posted back on Page 61 (please see) , wanted to say thanks for the response but that pink spot is a sparkle of sunlight and there is a spot visible in the sky that looks like a white star in the big high def pics directly above in the sky, above Flash- this is not in my opinion a stuck pixel...

They are classic examples of stuck pixels. Anything else would require an explanation that trumpted the stuck pixel explanation. Quoting reports of UFO sightings does not do this, for several reasons, one of which is that the dates are wrong (how can reports of UFO sightings on December 4 explain pictures taken in November?).

[hazeii3]

I think you missed my point.

<snip>

One would be taken from data in the putative wing trail. The other would be taken just outside that range. The curve I want to see is a difference between the two.

Ok, what I put up earlier was the difference between a strip at y=0 (outside the trail) and the normalised sum of all other strips (which would include the trail). Still, if you look at the 3D version, that's effectively showing the differences between the FFT of the first strip and the FFT of the rest.

Here, then, is the difference between the FFT of a strip 5 pixels in from the edge of a 200-pixel cross-section across the trail (this strip would be outside the trail) and the FFT of a strip 96 pixels in (which would be about mid-span on the upper wing).


(click image for larger version)

The wing beat signal would, if present, be somewhere close to the left edge (on the full size image, each pixel represents about 10.75Hz, so a 200Hz flap would be about 18 pixels in from the left edge).

[RAJ]

I say its been over since early on, and congrats to "geon" on page one for first proposing the simplest (and correct) answer. Its a bug. I'll leave the conspiricy theorists and red herring chasers to continue the debate. I'm outta here!

[Guest]

I say its been over since early on, and congrats to "geon" on page one for first proposing the simplest (and correct) answer. Its a bug. I'll leave the conspiricy theorists and red herring chasers to continue the debate. I'm outta here!

And to Bob Harrington (top of page 6).

....

The images still have the EXIF info from the Canon Powershot G3 camera used. Included in that info is the fact that the flash was fired for each of the images (not sure why flash would be used in this situation, but perhaps part of the automatic functioning of the camera)

I suspect an insect flying by close to the camera - and thus out of focus - left the darker shadow during the 1/20 second exposure, and then was illuminated by the very brief (usually well under 1/1000 second) strobe of the camera's flash.

....
Bob

[Jwhandyman]

It doesn't seem to me that the streak and light exactly line up, and therefore may not be related. The dark streak appears to be a jet contrail, and the light appears to be on the water, not on the utility pole.
Light reflecting off of the windshield of a small boat (which would move fast enough to be out of the frames before and after this one) could explain this.

[victorengel]

Here, then, is the difference between the FFT of a strip 5 pixels in from the edge of a 200-pixel cross-section across the trail (this strip would be outside the trail) and the FFT of a strip 96 pixels in (which would be about mid-span on the upper wing).

OK. I think you have shown pretty conclusively that there is no pattern louder than the background noise. Thanks.

[meteorwright]

I support the theory this is a shadow of the lamp housing caused by the blow out of the lamp which in turn appears to be caused by a lightning strike to the near base of the pole. There appears to be a failrly clear lightning track which could easily cause the bright blow out of the bulb and subsequent shadow.

[Doug Huffman]

Victor- yes NOV 24 as opposed to DEC 4, I think I was suggesting that strange unexplained things had been happening in area- this picture certainly qualifies. The star is like those seen associated with crop circles and was there in the sky and photographed and is not just a pixel. See post page 61 for more pics to compare. I think crop circles are real at least some, and they are laid down over months clearly the circles at different locations and times are not unconnected events - what does one thing have to do with the other here- maybe nothing maybe a lot- depending what this is. I think this photo goes beyond the normal -easy, the shadow, the strike, the smoke behind pole and a possible star almost directly over the event (which is not even being acknowledged except by me and you). It appears as if a force in a that single moment appeared struck behind the pole (maybe the water) creating smoke behind the pole. The constant is that star sitting in a dark cloud in all three pics directly above the event location - that pink thing low our side is a sparkle on the water what I call ' a diamond dog' . This might be another form of communication (like the circles) or a once in a million natural event. The star is there though it deserves some attention. ANY light on the bridge you choose based on need can be called a stuck pixel (and you could explain) but its not and I think you are arguing from personal judgements (where we are equal) even if your technical expertise exceeds my own (I'm sure it does)- i.e. it can't be a star so I'll explain stuck pixels. Prove (don't just explain) 'the star' is a stuck pixel and not a real object caught by the camera in the sky if you can't there is another dimension to this that should be discussed- the audience can decide if this is so . Again you must use the high def big images to see it. In general when something can't be debunked its called fake or someone makes a judgement not based on the evidence that takes the heat off case closed reality is edited - that's happening here with more than just the star. Face it this could be the true unknown to be fair some may be already at this conclusion it says as much in the original posting.

[victorengel]

Doug,

It would be helpful if you divided your posts into paragraphs. It would make reading them easier. By the way, I was using the term "stuck pixel" as a generic term to refer to dead, hot, or stuck pixels. A dead pixel is one that is always black. A stuck pixel is one that is always maxed. A hot pixel is one that reads too high. It's always on, but not maxed.

Victor- yes NOV 24 as opposed to DEC 4, I think I was suggesting that strange unexplained things had been happening in area- this picture certainly qualifies.

I'm not so sure it does. Just because the picture is unexplained doesn't mean that the event that it captured is strange. It may actually be quite mundane. If the hair theory or the bug theory is correct, for example, the picture shows an event that is quite ordinary.

The star is like those seen associated with crop circles and was there in the sky and photographed and is not just a pixel.

How can you say it was there in the sky? Were you there? Did you see it? It couldn't be a star because a star would be blocked by the clouds. It's also daylight (barely) so only the brightest objects would be visible in the sky. Here again, it would help to have the other pictures. If this is present in all frames, including ones possibly pointing in other directions, the star hypothesis is out and the stuck pixel theory remains.

Or maybe you're suggesting it couldn't be a stuck pixel because more than one pixel is involved. My reply to that is that only one pixel of the RAW data is involved. The camera has to manipulate the data from the raw sensor data to an image data set that makes sense to image processors. This involves, among other things, a demosaicing process that integrates the values of several photosites to create a single pixel. The result is that the value of one photosite influences several pixels. The picture also shows signs of having been sharpened, which is normal for a digital camera. Such a process would also affect neighboring pixles. Additionally, the photo was saved as a jpeg. Jpeg artifacts are introducted.

that pink thing low our side is a sparkle on the water what I call ' a diamond dog'

It is a stuck pixel. Look at the waves in the picture surrounding the pink spot. They change. If it were really a sparkle, the sparkle would change with the waves. By the way, there are more stuck pixles than the two we just spoke of. Can you find them?

ANY light on the bridge you choose based on need can be called a stuck pixel

No. A stuck pixel would be consistent on all frames, including ones taken of other subjects or no subject at all (lens cap on).

Prove (don't just explain) 'the star' is a stuck pixel and not a real object caught by the camera in the sky if you can't there is another dimension to this that should be discussed

I'm confident the stuck pixels can be proven, but more shots from the same camera are required.

[Guest]

for the fourier-analysis crowd, mapping this image into fourier space (a.k.a. frequency space or k-space) will give you **spatial** frequency information, because this is a spatial image. I fail to see how this spacial frequency can be accurately interpreted as temporal frequency (a.k.a. time frequency, what the laymen normally think of when discussing frequency), and using the shutter speed as a time metric would not be very accurate or justifiable.

What metric or mapping was used to convert spatial frequency to temporal frequency? It seems to me that any fourier analysis could only be interpreted if this mapping was perfectly linear (i.e. everything in the picture was moving at a constant velocity for the duration of the shutter speed)... which could be an accurate assumption for slow-moving objects.

only for objects moving at a constant velocity in the place of the image...

The point is that the photo was taken during a finite amount of time, which, if the bug theory is correct, maps directly to a line in the picture. One end of the line corresponds to the start of exposure. The other end to the end of the exposure. So there is a direct correlation between time and space. The insect image is what makes the connection.

[Lambert]

Cosmic ray in the CCD?

[Ray Mann]

It seems that the position of the sun in the picture (somewhere beyond the upper right of the frame) would make it difficult for the dark streak to be a contrail shadow... it would be helpful to know the exact time of day and camera direction, then a likely position for the contrail could be determined. Also, I would love to know if the original photo is higher quality than the one on the web page - if so, the contrail question could possibly be answered by looking closely at the area of water between the "flash" and the horizon - if the "streak" crosses this area, then the flash and the streak are part of the same phenomenon. My guess would be that the streak is a shadow of part of the lightpost, except that one would expect a shadow of that kind to widen further from the light souce, and in the picture it's not clear whether that is the case. As for the origin of the flash itself... I would guess the bulb in the post flashed as it butned out, though that wouldn't adequately explain the white smoky smudges near the flash. It might also help to see the photo frames before and after this one, to see whether the flash corresponds exactly with the light bulb's position.

[Guest]

*****************************************************
TWO REASONS WHY IT CAN'T BE A FALLING OBJECT/METEOR...
*****************************************************

1) things falling that fast usually exceed the speed of sound; any smoke/compression cloud would appear behind the falling object, rather than ahead of it. See this pic from past apod for a similar effect:
http://antwrp.gsfc.nasa.gov/apod/ap030504.html
If the alleged meteor fell below the speed of sound, the compression front could shift ahead of the object, it would not still be burning. If a lamp strike occurred, I suppose it's possible that a flash could be caught by the camera, but extensive visible damage would be visible on the lamp. My neighbor had a 3" hole punched through the roof, passenger compartment, and floor of his parked car by a meteorite that was just 1" across! But people could still argue outta this reason, so I give you #2:

2) 1/20sec exposure could simply NOT capture ANY resolvable features of a falling object with great velocity; the bright spot could just *not* represent a burning meteorite... to be still burning, it would have to be traveling very very fast (>>speed of sound) and would just not be resolved on a camera with that exposure setting. Just imagine trying to take a picture of a tracer bullet (those ones that have magnesium or phosphorus to burn brightly as they travel); unless you're several miles away (clearly not the case here), looking at a profile of the bullet's trajectory in 1/20 sec will simply give you a streak; there will be no resolvable features on either end of the streak. Saying that this picture represented a falling object would be like taking a picture of said bullet in 1/20 as it happened to pass by the FOV of the camera, and managing to see a nicely resolved image of the bullet at the end of its trail. THIS DOES NOT HAPPEN... SIMPLE & WELL-KNOWN OPTICS/PHYSICS. If you don't believe my, go set up a camera and try to take a picture of a speeding bullet at 1/20. If that's out of your reach, then set up a camera and throw something glowing past it as hard as you can... then you'll understand.

must look elsewhere, at which point, relevance to apod becomes moot.

[victorengel]

This is just a reminder that there is a program on Animal Planet today, called Spy on the Wild that, among other things, promises an experiment to determine how straight bees can fly. Since this is directly related to this thread, I thought I'd post this.

I think I posted earlier that they hooked video cameras to the bees. It seems I saw the preview wrong. They hook a camera up to an eagle. They mount something on a bee, but I don't know what it is. Maybe it's some sort of transponder. In any case, it looks to be a very good show.

[DC]

Back a few pages someone wondered if the top end of the trail ended with a curve. This is what I get with an enhanced diff image, contrast enhanced and blurred. The places thought to be part of a curved trail in the earlier image show up as a slightly different color here, so I think the trail is straight all the way.

[Guest]

Aliens.

[Guest]

Back a few pages someone wondered if the top end of the trail ended with a curve. This is what I get with an enhanced diff image, contrast enhanced and blurred. The places thought to be part of a curved trail in the earlier image show up as a slightly different color here, so I think the trail is straight all the way.

The trail definitely isn't straight, which can be fairly easily seen if you rotate it to horizontal and compress it along the X axis. On the image below (posted earlier) the trail is shown compressed horizontally on the left.



I can't quite orient your image though - the bright diff on the left seems to match the bush on that side of the image, and the bright arcs look like the cloud edges above it in this image but the trail seems in the wrong place for that to be the correct assumption.

[Guest]

for the fourier-analysis crowd, mapping this image into fourier space (a.k.a. frequency space or k-space) will give you **spatial** frequency information, because this is a spatial image. I fail to see how this spacial frequency can be accurately interpreted as temporal frequency (a.k.a. time frequency, what the laymen normally think of when discussing frequency), and using the shutter speed as a time metric would not be very accurate or justifiable.

What metric or mapping was used to convert spatial frequency to temporal frequency? It seems to me that any fourier analysis could only be interpreted if this mapping was perfectly linear (i.e. everything in the picture was moving at a constant velocity for the duration of the shutter speed)... which could be an accurate assumption for slow-moving objects.

Simple; the exposure is well-documented as 1/20th second, amd since the trail starts and ends inside the image boundary we know how many pixels 1/20th of a second represents. From that it's simple to determine that a periodic signal like an insect wing flapping at say 200hz would leave 10 beats along the trail.

[Bob Peterson]

Streak=contrail with craft moving from left to right. Contrail is dissipating from left to right.

Flash=unrelated event. Probably the sun reflecting from the light pole or reflective surface, seemingly inline with the contrail.

Smoke/Haze around light pole area=unrelated event. Smoke from engine startup or other source nearby. Smoke drifting from right to left diffusing the "flash". I say right to left because the surface wind seems to be blowing in that direction, judging from waves in the harbor water.

I've only read a few posts. Did anybody request that the photographer post all 38 of the his photos, so this could go on for the next couple of years?

[Guest]

Excuse the awful graphics -I am an amateur at photo editing. Below are crops from the before-during-after sequence.



Maybe I'm wrong, but this seems to be a reflection of the flash in the foreground water. Did anyone else notice it? It doesn't apear in either the before or after photo. It seems too regular to be a random play of light on the water. It seems to stop right at the shore's edge, which is evidence that the light really is a reflection off the water and not a photography problem. It forms a curve which on average points towards the flash. I don't know why the reflection would be not a straight line - maybe because the waves have a certain structure close to the shore? Finally, this is the only "thin curve/line" type of pattern in the any of the water in any of the three photos. And it happens to be in the exact position one would expect if it were a reflection of the flash.

Anyway, I'm not convinced by this but I think it is a possibility.

My pic is good only for pointing out what I'm referring to, not for detailed viewing - please see the original photo instead.

http://antwrp.gsfc.nasa.gov/apod/image/ ... de_big.jpg

[victorengel]

Maybe I'm wrong, but this seems to be a reflection of the flash in the foreground water.

It's hard to tell, because the waves in the water make taking a difference difficult. In order to compensate for the waves, I took an average of the before and after pictures and then applied a gaussian blur with radius 4. I also applied a gaussian blur of the same amount to the flash picture and made a difference image using this and the average picture. The gaussian blur smooths out the small waves. However, the waves with long wavelength should still be visible.



I see that there is a bright spot in the water directly in front of the flash. However, it seems to be part of a larger wave pattern. I say this because on either side of the bright spot (mostly in front and behind) is a dark spot. So, while the water does seem to be brighter, I think that fact is a red herring. Also, the brightness extends farther to the left and right than one would expect if it were a reflection of the flash event.

[Straub]

Has anyone noticed the ofset of the horizon on both before and after photos? They would appear to have been altered. unless the land in the distance is shifting. I don't think photos like this belong on the astronomy pic of the day.

[Luis]

Has anyone noticed the ofset of the horizon on both before and after photos? They would appear to have been altered. unless the land in the distance is shifting. I don't think photos like this belong on the astronomy pic of the day.

I think that is a bit tough on APOD. The image was posted to generate this forum and find out or at least get an idea of what could have produced it. If it was a mosquito, a hair or a good photoshop work, it was not known at the time (and I don't think it is known yet). Retrospectively is easy to say, it should not be there, but when it appeared people thought it could have been a meteorite and it has already generated 84 pages of mostly interesting posts, a lot of burned neurons, and it has not yet been answered.

I say good work APOD for giving the chance to the web community to interact with it and let us put our bit of help to solve a potentially interesting scientific problem.

[HomeAl0ne]

Maybe I'm wrong, but this seems to be a reflection of the flash in the foreground water. Did anyone else notice it?...

Yeah, last Friday I toyed around with testing some predictions from the various competing hypothesis.

Any point source generating light in the middle ground/background (exploding lightbulb / space debris reentry / meteor explosion / lightning / rocket launch / firework etc) should result in relections from the water surface which chould be detectable in the image.

Flat water would produce a single reflection in a location determined by the distance between the 'flash' and the camera, and their respective heights above the reflective water surface. Waves will produce multiple images, preferentially aligned in the direction the camera is pointing but 'smeared', as various transitory wave faces reflect the light to the lens.

A flash in the middle of the lake is not supported by any obvious reflection from the water between the flash and the wharf.

A flash associated with the lamp is harder to rule out. You would predict the greatest relection of the flash to be coincident with the reflection of the top of the lamp post.

It's not clear from the photo exactly how far the lamp post is from the camera. I suspect it is mounted on the furtherst, lowest wharf. That would place the lamp roughly three times higher above the water than the road. The road casts a shadow very close to the base of the pylons, so the predicted position of the reflection would only be about level with the base of the pylons supporting the nearer road. (ie the other side of that headland, or in line with the rear of the white car).

Best case for producing a reflection has the lamp on the nearer road. That would have the half as high again off the water as the road surface, and the reflection would be half as far again as the raod shadow from the base of the nearer pylons. This puts the predicted reflection roughly in line with the high tide mark on the near headland. The trees and headland itself may block reflections close the land, but there may be some reflections from wave tops closer to the camera.

But for comparision, can you detect any reflection from known point sources of light at the same rough distance? Take a look at the illuminated lights on the wharf and buildings to the right? See any reflections from them? I can't.