Strange streak discussion: 2004 Dec 7 APOD

[sctdixon@earthlink.net]

My name is Samantha. I live in Las Vegas and I too have seen these streaks in the sky and was too stupid to have not brought my camera. Funny thing is that it only happens when it's cloudy. I've seen up to 3 streaks at one time, all going in the same direction. The only thing I can think of is shadows, but the streaks seam awfully dark. Could it have something to do with water? The streaks that I've seen seem to be in the direction of Lake Mead. This photo is also over water. Go figure.

[Wombat]

Thanx to Victorengel for providing the (de)convolutions. As I see it, the most significant point that they show is that the more processing of the image, the sharper the "object" and the blurrier (is that the right word?) the background. Does anyone dispute that?

Surely this would very strongly support the theory that the object (bug, human hair, whatever) really is very close up? Because if the object was indeed distant (on the dock / jetty / whatever), the processing would not have made them sharper. Is there a flaw in that argument anywhere?

[victorengel]

First of all, I’d like to refute the “Bug Hypotheses”.

1. The shadow is straight line – generating a line with any photo editing program will eliminate the illusion of an arcing path.

It's actually not straight, as I've demonstrated earlier. It curves first to the left, then to the right, then to the left again. It's slight, but it's there.

2. It is just too coincidental that the line intersects the light flash.

The light flash is a reflection off the insect. When you flash a strobe at something shiny, it's not a coincidence that it makes a bright spot.

3. If the insect were to have crossed the field of view, it would have been pretty darn close to the lens. You can calculate velocities at increasing distances from the lens if you want – remember, shutter speed is 1/20.

These calculations have all been done and show values well within the range of insect flying speeds.

Anyway, here’s what I think. The camera simply caught the the bright flash of a light bulb in the act of burning out. The shadow and related aura around the flash are simply photographic anomalies produced by the camera during exposure.

An anomaly on a digital camera that goes for some distance in a slight undulating pattern at about 33.6 degrees is very much of a stretch to be a camera anomaly.

[redxeth]

Just found this website-- more fuel for the bug theory.

http://www.cloudbait.com/science/darwin.html



From his difference image, the line caused by the bug does appear to get wider as you move to upper left. And the line isn't perfectly linear.

This implies that the bug started out closer to camera and is flying slightly away. That the bug has translucent wings explains why only the body of the bug darkens the exposure.

************
One other thing, putting the time/data/location into Starry Night Backyard (an astronomy program) shows the sun to be 58 degrees above the horizon putting it outside of the image-- above the brightly lit clouds.

Regards,

Dan

[victorengel]

Hey, all you BUG people---why isn't the trail visible below the rear of the boat, on water, if it's a bug?? What, it flew out of the flash? If it's supposed to be so close to the lens, then where is the continuation on the bottom? Looks more and more like the flash is something produced from the boat in the picture, maybe a burp from an exhaust pipe, diesels do that, and the trail could be unrelated to water scene and explained by something from landmass in rear of scene---ya think?

I don't even understand what you're suggesting. The bug went from upper left (notice the trail starts some distance away from the edge of the picture) to lower right. You can see the image of the bug. The bug trail goes all the way to the bug, including that short distance in front of the water in the distance. The flash of the camera goes off either at the very start of exposure or very end. In either case, we would not expect the trail to go beyond the image of the bug.

If I misunderstood what you were saying, please reword.

[Ruben]

When I was coming home today, I saw a black streak in the sky just like the one in the photo in question (It immediately reminded me of this putative mysterious photo). It was obviously a jet contrail--another jet was flying the same course. Looking closer at the photo, one can see that the flash does not quite match up with the streak--for a closer correlation, the flash should appear further out in the water. The streak is a jet contrail disappearing over or coming from the distant horizon. The flash is simply the light burning out at that moment. The only explanation lacking is why the streak only appears in one photo, but that is hardly something to get all exercised over.

[Guest]

All this and what we all know about bugs: They tend to fly erratically. It keeps them alive longer that way. If this bug is so close that its path appears straight, the flash should have caught it. If its so far away that the flash didn't get it, its very unlikely this bug would have flown an arrow straight trajectory!

In my view, the logic stacks up to easily against it being a bug/insect.

Again, I didn't mean to be rude, my apologies if it came across that way.

THe thing is, it didn't fly in an arrow straight trajectory! There was a post by someone with my name starting with "v" but I have forgot who it was sorry! but it clearly showed the black streak was not arrow straight.

skyglow1

[victorengel]

Go here to see some work I did on sorting out a sense of scale for this debate. I think the 'bug' theory is up in smoke, if you'll pardon the jab.

Please explain. How does your "analysis" disprove the bug theory?

[victorengel]

OK, after looking at the image in photoshop the linear artifact is way too linear to be caused by the flight of the "bug". This doesn't eliminate the bug theory.

What are you talking about? Keep your TV tuned to Animal Planet on Sunday (I think). They have a program where I believe they strap a camera to a flying insect. I saw a preview today where they specifically commented on how straight bees can fly. This should be interesting stuff.

Anyway, the path is NOT straight.

[Guest]

All this and what we all know about bugs: They tend to fly erratically. It keeps them alive longer that way. If this bug is so close that its path appears straight, the flash should have caught it. If its so far away that the flash didn't get it, its very unlikely this bug would have flown an arrow straight trajectory!

In my view, the logic stacks up to easily against it being a bug/insect.

Again, I didn't mean to be rude, my apologies if it came across that way.

THe thing is, it didn't fly in an arrow straight trajectory! There was a post by someone with my name starting with "v" but I have forgot who it was sorry! but it clearly showed the black streak was not arrow straight.

skyglow1

Oh I remember now, it was thanks to Victorengel who posted the image that showing the flight path was more of a curve.

skyglow1

[Ed in Oregon]

Lets talk about the odds here:
If it is an exploding street lamp, what are the odds that it exploded just as the camera shutter was open. The life of a street lamp bulb varies (with type and manufacturer) from 6000 hours to 24000 hours or thereabouts. They burn on average 12 hours a day, making the life from 500 to 2000 days. Lets assume that they only burn out on startup, and always burn out with a flash (which is probably very wrong, but is OK for this calculation). This means that on a given day, the odds of the bulb burning out are from 1 in 500 to 1 in 2000. The man is taking a picture every 15 seconds, and the exposure is 1/20th of a second. This means the camera is "looking" 1/300th of the time. Let's assume that we know the camera is in operation during the time that the light should normally come on. (Which I don't believe is the case, because the pictures are taken within a minute or two of true sunset, and in my experience street lamps don't come on until 20 or 30 minutes after sunset, unless it is full overcast, which is not the case here). SO, the odds of catching the "burnout" flash is 300 times 500 or 1 on 150,000 to 1 in 600,000. Note that I have made several assumptions that reduce the odds by a lot. The real odds are probably 1 in several million.

Now, what are the odds of a bug flying in front of the camera within a few feet just as the flash goes off. Remember, this is Darwin, Australia, at dusk, in the local late Spring, next to the harbor, with lots of bushes and other vegatation around. Go stand there without bug repellent and tell me what it's like. Also remember that he took about 50 pictures. (I don't remember the exact number, and I'm not going back through this thread again to find out.) I'd put the odds at about even that one of those pictures will have a bug at just about the right range. If he duplicates the experiment, he'll catch another bug. It won't look precisely the same because the odds are that the orientation, position, and the species of the bug will be different, but he will catch one.

Why isn't there a shadow of the wings? They are moving very fast. They are semi-transparent. They obscure the background by about 1/1000th of the amount that the insects body does. Not enough to show. Remember that most of the images that show the shadow distinctly are processed to bring out the shadow to the limit that noise will allow.

Odds of a meteor? In a given year, lets say that the odds of a meteor hitting the ground within range of an operating camera are greater than 1 in a billion. (Only a few strike the ground anywhere in the world in a given year. Divide the area of the earth in terms of camera range by that number). There are 31 million seconds in a year. Multiply that by 20. Divide by 50. Multiply that by a Billion. That's 1 in 12,000,000,000,000,000.

This has been fun!

[Guest]

I like the idea related to the bulb flashing out and leavning a shadow, it strikes me as odd though that the shadow should have such a length. That seems a reasonable explanation though.

The day is fairly cloudy, though, and the puff of white vapor/smoke strikes me as odd as well. Is it possible that a large ice particle (hail or whatever)could have hit the light with enough force to jar the bulb into a flash out while at the same time the ice sublimates into the vapor from the heat of the lamp and impact speed... with the light in front of the vapor, the cloud would appear more substantial, but perhaps even with the light effectively behind it, this could explain it. A rapid temperature change could also cause the bulb to go out which could help the idea that a piece of hail or some other ice particle hit it.

Also, is it possible for dirty ice to carry a charge?

[victorengel]

I didn't mean to pick up my flamethrower; I just meant to point out that, based on more easily observable, rather mundane phenomena -- like arcing light-poles

You see an arc somewhere? Please show us. Also explain the flow of electricity. Recall that there was no damage to the light -- only that it was not working.

or workmen leaning ladders against live-wires in industrial areas, etc. --

I don't see anything like that at all. Do you see a ladder or workmen somewhere? Are there any wires in the vicinity of the feature you allege is an arc?

that the probability that this is a bug requires more Ifs and Ands and Thens than the more straightforward and, frankly, more likely explanation.

So list them and show what's wrong with them. The odds that an insect would be flying around is pretty high. There are insects everywhere. That it flew in front of the camera at normal speed is a coincidence, but not very odd.

I don't believe bugs don't sometimes get in the way of images, causing strange side-effects. I just don't believe this is a case of that happenstance.

Why? So far I haven't given a reason.

Even at this relatively low resolution in the larger original image, to me, a puff of smoke and a bright lightsource in line with a utility pole spells out the fairly obvious.

They are NOT in line. Close, but not in line. They are offset by what works out to be a couple of meters I think. Someone else did the calculations. They're in the thread somewhere.

Also, being an avid photographer tells me from experience that the line observed in the image is likely a failure of the device.

Lines resulting from failures in a camera tend to be straight (this one isn't) and orthogonal to some structure in the camera (33.6 degrees is not orthogonal to any feature of the camera so far as I know).

Film and digital sensors only have a 4 stop range (avg).

I don't think so. I've done testing on the Canon 10D showing it has about 10 stops of dynamic range (http://forums.dpreview.com/forums/read. ... ge=9651949). Please feel free to take issue with my method here or there. I'm quite satisfied it's fairly sound. While I recognize the G3 probably has a lesser dynamic range, I don't know by how much, but I doubt the difference is that much.

The brightest part of the cloud and the bright-spot are already well beyond that range, so they show as "pure-white only". Nothing surprising. The 'smoke' is lit by that same light source and renders fairly accurately as a bluish-white puff we typically observe as, well, smoke.

OK. But if it's smoke, then it's near the lamp. At that distance, the size of the smoke cloud is fairly large (several meters across). A personal email from the photographer states that the winds were calm. The relatively small waves confirm this. So what happened to the smoke? Smoke over this large a volume lingers for more than 15 seconds. But there's absolutely no evidence of it in the subsequent shot.

There's a simpler explantion for the blue color of that part of the image. You're seeing the wings or other parts of the insect illuminated by the flash. The flash is much cooler in color temperature than the setting sun. Hence, it looks blue. The bright gold spot evidently on the insects abdomen could be pollen, like in the photos of several bees previously posted. Or it could be the reflection of an iridescent insect. I don't think there's any way to discern the species of insect, but it's clear that there is enough variety in the insect world that something would appear like what we see.

If this was a bug, relatively close to the camera, anywhere within say 10-15ft but perhaps as far as 20ft, the flash which everybody is going on about would have caught it "within useful range" for a consumer camera such as this Canon and your bug would be a cleanly described, if not terribly in-focus, nicely frozen image, in front of the rest of the shot, since in ~1000/sec it would not move very far at all!

I believe it's much closer than 10-15 feet, but I haven't run the numbers. Something like a gnat could be just inches away, but would likely be blown out by the flash. I don't think the EXIF data say anything about the flash exposure, and I don't know enough about the G3 to know how much power would be used in a situation like this. I hope someone else who does know will weigh in on this, because we can probably get an upper and lower bound for the distance of the insect, and hence an idea of the size (assuming the insect theory is valid).

All this and what we all know about bugs: They tend to fly erratically.

They also tend to fly in straight lines. Megaloprepus, for example, flies in straight lines. Obviously that's not what we have here since it's Mexican. Bees fly in straight lines, too, when they're flying long distance. They fix an angle to the sun and fly to that angle. Many insects navigate this way. That's how they spiral around light bulbs. They're essentially dead reckoning against the light source, which their hard-wired system assumes is located at infinity. Their wiring breaks down when the light source is nearby, causing them to spiral in to the light source.

If this bug is so close that its path appears straight, the flash should have caught it.

It did. Look at any number of difference pictures. You can see what appears clearly to be an insect, frozen in flight.

[Ed in Oregon]

Forget the bulb flash. Any shadow from the bulb (assuming there is enough mist in the air to support a shadow, which there is obviously not!!) will be very conical in shape. Remember, the bulb is part of the lamp fixture and pole. It's very close to it. Any shadow from the pole/fixture would be very wide, and get much wider farther from the pole.

[wombat]

I'm confused by the arguments over the "Flash" info shown in EXIF. In theory, we have the photographer's word that he DID use flash, so that is the end of speculations.

But as someone said, let's play the game. I see this as important, because the existence of the flash seems to me totally essential to the bug theory.

At least one post here has the EXIF flash info as "Yes" or "On". Someone else has posted "Undefined". Mine shows "Flashmode Off"; the other values seem to agree with those mentioned by others.

So I'm confused, a frequent occurrence. Can someone unconfuse me here?

[HomeAl0ne]

I converted all three images to tiff format. Then, using pixel math, I
1- Took the average of the before and after images.
2- Subtracted the result of 1, above, from the image showing the hit.
3- I "stretched" the result of 2, above, wth the black level at 0 and the white level at 3000. I then converted this back to jpg format with 70% compression.

You can see the result at:
http://tjpstewart.org/images/Hit-minus- ... ess-70.jpg


The result of 3 appears to show that the upper left terminus of the dark streak is much more definite than it appears in the unprocessed "hit" image.

Terry Stewart

Any idea what the apparent arc is to the left of the bright blob? Go back up the streak about two bug wingspans, then follow the arc out to the left and down. The left hand end appears to be the left most lamp post alongside the road on the land in front of the wharf. The center of the circle the arc sits on would appear to be near the top of the other lightpost in the middle ground.

[Cloudbait]

I'm confused by the arguments over the "Flash" info shown in EXIF. In theory, we have the photographer's word that he DID use flash, so that is the end of speculations.

But as someone said, let's play the game. I see this as important, because the existence of the flash seems to me totally essential to the bug theory.

At least one post here has the EXIF flash info as "Yes" or "On". Someone else has posted "Undefined". Mine shows "Flashmode Off"; the other values seem to agree with those mentioned by others.

So I'm confused, a frequent occurrence. Can someone unconfuse me here?

I guess it is just how some programs interpret the data. The FLASH parameter in the header actually has a numeric value, which I suspect is the flash power. If an interpreter is expecting a boolean value, it might consider this undefined, or report ON/OFF based on the value being even or odd.

I've viewed the header in Photoshop CS, ACDSee 3, ACDSee 6, and the Canon Digital Camera File Viewer Utility version 1.3.1.9. They all report the flash was used.

[Guest]

The EXIF information (one shown below) ALL show FLASH USED as YES.

EXIF
Camera
Manufacturer = Canon
Model = Canon PowerShot G3
Dimensions
Width = 2272
Height = 1704
Advanced
Focal Length = 9.1mm
Focal Length (34mm equivalent) = 1169mm
…..

Black and white = no
Flash Used = Yes
Encoding = Baseline

[victorengel]

From his difference image, the line caused by the bug does appear to get wider as you move to upper left. And the line isn't perfectly linear.

I did my best to try to show this clearly. I rotated the image 33.6 degrees counterclockwise and then did manipulations to objectively measure the width of the trail. This involved things like contrast adjustment, gaussian blur, resizing, and finally smart blur with appropriate settings to show the outline of the flight path. The first image showing the path clearly is this one.

This is just after applying the smart blur to the image after being significantly squeezed width-wise. This resizing operation effectively integrates much of the variability over the length of the path and also makes it easier to use the smart blur tool to show the outline consistently. I then reversed the outline, and resized it to a bit wider, so you can get a meaningful view of the path.

At this point, it's pasted onto a duplicate image from the difference picture, so you can see the outline relative to the context. I resized back to the original dimensions, stretching the outline.
http://the-light.com/Photography/outlinewide.jpg
Then I rotated this back to the original orientation (clockwise 33.6 degrees) and pasted the outline layer to the original difference image. Here you see the outline showing the path width in its correct original context.
http://the-light.com/Photography/outlinecontext.jpg

This was some tricky Photoshop work, but I wanted a procedure that applied equally to the whole image I was processing in order to make an objective measurement.

The smartblur tool works on a threshold basis. So the width change could be a red herring if the path crossed a gradient in the background. The constant threshold would tend to drift because of the bias imposed by the gradient. The image the smarblur tool was applied to, however, had no noticeable gradient, because it was it was the difference image, which has the gradient subtracted. My conclusion is that the width indicated in these pictures is about as objective a measure as I can come up with.

P.S. Edit to post involved removing IMG tags on the two wide images, in order that the text in other messages are easier to read.

[caseyoconnell]

It seems this topic is another example of the desire of many to express what they are capable of, which doesn't include careful analysis.

The movie of the bug is quite a pursuasive visual.

No arguments address all of the data except the bug argument.

Very few people are going to read through this topic and actually understand what is being said.

When looking for an explanation, it is not enough to make a guess and then start supporting your guess.
That is how people succeed at fooling themselves.

Once a hypothesis is made, it is the responsibility of one seeking to understand to try to find difficulties with the hypothesis.

Difficulties have been found in all other explanations except the bug explanation.

[HomeAl0ne]

Flies and bees have wingbeat frequencies around 200 Hz.

http://hypertextbook.com/facts/1999/Mic ... egan.shtml

In 1/20th of a second the purported insect would have beaten its wings approx 10 times. (Insects the size of flies and bees and smaller rarely glide, the apparent viscosity of the air makes this energetically uneconomical for them).

Is someone smart/unemployed enough to attempt some sort of analysis on the streak to see if they can discern a regularity to it? Either a regular up and down displacement of the thorax/abdomen, or a preferential shadowing either side of the 'flight path'.

Absence would indicate nothing, presence would support the bug theory slightly.

[caseyoconnell]

Flies and bees have wingbeat frequencies around 200 Hz.

http://hypertextbook.com/facts/1999/Mic ... egan.shtml

In 1/20th of a second the purported insect would have beaten its wings approx 10 times. (Insects the size of flies and bees and smaller rarely glide, the apparent viscosity of the air makes this energetically uneconomical for them).

Is someone smart/unemployed enough to attempt some sort of analysis on the streak to see if they can discern a regularity to it? Either a regular up and down displacement of the thorax/abdomen, or a preferential shadowing either side of the 'flight path'.

Absence would indicate nothing, presence would support the bug theory slightly.

It is indeed amazing that someone would post intelligently.
This author is applying the Newtonian method of making a conclusion based on the hypothesis that can be tested for.

One difficulty is the limitation of pixels, another difficulty is that it is an attempt to prove without allowing the possibility of disproving.

Any test of a hypothesis needs to allow the possibility of disproving the hypothesis.

Any hypothesis that cannot be disproven is useless.

[victorengel]

Flies and bees have wingbeat frequencies around 200 Hz.

"The number of wing beats varies greatly from 4 - 20 in butterflies to 190 beats/second in bees and up to 1000 beats/second in a small fly."

We can't yet rule out a small fly. That would be 50 wingbeats during the exposure.

Perhaps we could consult som Roswell Rods pictures to see what the normal spacing between wing beats is as a function of wingspread. This might get us a rough estimate of total wing beats and thus a pointer at the size insect we're looking at.

[Harry Law]

Could it possibly be a drop of rain? Or some water that were collected on a near by tree that just happened to falls off at that time and caught by the camera? If that is so, it explain the straigthness of the shadow and the shadow itself, as well as the flares like effect at the end. With the reflection of water, something like this is very possible

I can't think of any other possiblity that get close..

[smith-at-canada-com]

Big congrats to victorengel for the horizontal compression technique. I tried that on my own diff image, only I ran the horizontal compression to 10:1 -- there is NO WAY that line is straight.

In an earlier post, I commented that only the exceeding straight line made me nervous about the bug conclusion. But the line is not straight.

It's a bug.

I have made up a composite image of my diff, with an inset blowup of the bug outline, and another inset of the 10:1 compression on the streak. It looks like a nice presentation of some well placed image analysis.

The bug wins.