Home · Register · Join Upload & Sell

Moderated by: Fred Miranda

New fredmiranda.com Mobile Site

New Feature: SMS Notification alert

New Feature: Buy & Sell Watchlist

FM Forums | Canon Forum | Join Upload & Sell

2 end

Sensor error, or valid data? (5Ds high ISO)

Jeff
Offline
• • • • •
Upload & Sell: On

p.1 #1 · p.1 #1 · Sensor error, or valid data? (5Ds high ISO)

I'm kinda stumped here, and really not a sensor/tech/data guy like @snapsy is.

The following is a 100% crop from one of 1477 sequential images captured for a time-lapse (this is pre-edit, i.e. out of the camera, raw file). That I have seen, it is the only single frame that has this sort of oddity. At first I assumed it to be some sort of weird noise artifact, but I have never seen such a thing in all the night images I've captured over the years, and the nature of it seems to be both chaotic and yet non-random at the same time.

Has anyone ever seen something like this, or might you have an idea what it is attributable to? Does the angle of the line make any sense in regard to how the sensor reads the data?

The reason I became interested in this is because I can think of one reason that this could be valid data, however large a stretch it may be. I won't divulge yet so as to not color the discourse, but I'd love to hear people's ideas on what this could be.

Thanks,

Jeff

13s, f/2.2, ISO 4000 (1:1 crop, bottom left corner)

full frame added for scale perspective (ignore the airplane in the center)

Edited on Aug 15, 2023 at 07:47 PM · View previous versions

Aug 15, 2023 at 07:44 PM

EB-1
Offline
• • • • • • •
Upload & Sell: Off

p.1 #2 · p.1 #2 · Sensor error, or valid data? (5Ds high ISO)

Meteors? Alien lasers?

EBH

Aug 15, 2023 at 07:46 PM

Jeff
Offline
• • • • •
Upload & Sell: On

p.1 #3 · p.1 #3 · Sensor error, or valid data? (5Ds high ISO)

EB-1 wrote:
Meteors? Alien lasers?

EBH

That's what I was thinking! lol

Aug 15, 2023 at 07:47 PM

bballfreak6
Offline
• •
Upload & Sell: Off

p.1 #4 · p.1 #4 · Sensor error, or valid data? (5Ds high ISO)

I've seen that once before back when I had a 5D4; I forgot the exact reason/name but it's some kind of phenomenon. I can try to find the old email chain I had with the Canon tech when I am home from work haha.

Aug 15, 2023 at 08:07 PM

Jeff
Offline
• • • • •
Upload & Sell: On

p.1 #5 · p.1 #5 · Sensor error, or valid data? (5Ds high ISO)

Would love to see what you've got if you can find it, along with a screenshot of yours if you still have it.

bballfreak6 wrote:
I've seen that once before back when I had a 5D4; I forgot the exact reason/name but it's some kind of phenomenon. I can try to find the old email chain I had with the Canon tech when I am home from work haha.

Aug 15, 2023 at 08:31 PM

EB-1
Offline
• • • • • • •
Upload & Sell: Off

p.1 #6 · p.1 #6 · Sensor error, or valid data? (5Ds high ISO)

Check the solar flare activity. Maybe something is activating the sensor.
https://www.spaceweatherlive.com/en/solar-activity/solar-flares.html

EBH

Aug 15, 2023 at 08:37 PM

Jeff
Offline
• • • • •
Upload & Sell: On

p.1 #7 · p.1 #7 · Sensor error, or valid data? (5Ds high ISO)

EB-1 wrote:
Check the solar flare activity. Maybe something is activating the sensor.
https://www.spaceweatherlive.com/en/solar-activity/solar-flares.html

EBH

Is that a thing?

Aug 15, 2023 at 11:23 PM

snapsy
Offline
• • • • • •
Upload & Sell: On

p.1 #8 · p.1 #8 · Sensor error, or valid data? (5Ds high ISO)

Line is too perfectly angled to be a naturally-occurring phenomena. Also its angle doesn't match the star motion trail angle, so it doesn't appear to be some derivative artifact from a star trail. However if you zoom in close on the 100% crop there is a green blob that appears to exhibit coma, which implies whatever it is was at least resolved through the optics rather than being exclusively a sensor or DIGIC artifact.

Having access to the raw would help.

Aug 16, 2023 at 03:10 AM

Jeff
Offline
• • • • •
Upload & Sell: On

p.1 #9 · p.1 #9 · Sensor error, or valid data? (5Ds high ISO)

snapsy wrote:
Line is too perfectly angled to be a naturally-occurring phenomena. Also its angle doesn't match the star motion trail angle, so it doesn't appear to be some derivative artifact from a star trail. However if you zoom in close on the 100% crop there is a green blob that appears to exhibit coma, which implies whatever it is was at least resolved through the optics rather than being exclusively a sensor or DIGIC artifact.

Having access to the raw would help.

Just sent you a PM with Dropbox link. Thanks for taking a look!

Aug 16, 2023 at 07:16 AM

snapsy
Offline
• • • • • •
Upload & Sell: On

p.1 #10 · p.1 #10 · Sensor error, or valid data? (5Ds high ISO)

Jeff wrote:
Just sent you a PM with Dropbox link. Thanks for taking a look!

Thanks, I looked at the two raws you sent, taken 13s apart, where the first doesn't show the line but the second does. It's approximately 1 pixel in width, spans all RGGB pixels across a 75.5 degree diagonal, and has a brightness value approximately equal to the stars but with a much higher concentration, which makes it stand out so much. It starts at the very bottom of the frame and extends to just before where the sky starts.

It's hard to imagine this being present in the actual scene. But it's even harder to imagine it being a sensor/DIGIC artifact since they rarely occur on a diagonal like this. The only time processing could even approach this kind of diagonal is a lens correction transform but Canon doesn't apply those to raw files anyway.

The line is also present in the embedded full-sized jpg, so it's not media data corruption.

It can't be a plane since it's so low in the frame. No natural emitter of light could be a perfect diagonal like this.

Can you describe what occupies that part of the frame? Is it just the base of a mountain?

Aug 16, 2023 at 08:06 AM

Jeff
Offline
• • • • •
Upload & Sell: On

p.1 #11 · p.1 #11 · Sensor error, or valid data? (5Ds high ISO)

snapsy wrote:
Thanks, I looked at the two raws you sent, taken 13s apart, where the first doesn't show the line but the second does. It's approximately 1 pixel in width, spans all RGGB pixels across a 75.5 degree diagonal, and has a brightness value approximately equal to the stars but with a much higher concentration, which makes it stand out so much. It starts at the very bottom of the frame and extends to just before where the sky starts.

It's hard to imagine this being present in the actual scene. But it's even harder to imagine it being a sensor/DIGIC artifact...Show more →

So, reading between the lines, you'd think this has to be an optical phenomenon, not electronic, despite however odd/unlikely it may be?

Aug 16, 2023 at 08:10 AM

melcat
Offline
• • • •
Upload & Sell: Off

p.1 #12 · p.1 #12 · Sensor error, or valid data? (5Ds high ISO)

Could it be something man-made, like a power line or fence, that glinted in an artificial light source like a car headlight or a camper’s headlamp for a moment when the angle was just right?

Aug 16, 2023 at 08:14 AM

Jeff
Offline
• • • • •
Upload & Sell: On

p.1 #13 · p.1 #13 · Sensor error, or valid data? (5Ds high ISO)

melcat wrote:
Could it be something man-made, like a power line or fence, that glinted in an artificial light source like a car headlight or a camper’s headlamp for a moment when the angle was just right?

Nope. Well, I actually am in the frame shooting from the base of the spire, but there are no wires, etc. If my headlamp had been on, it would be evident in the image, and the perfectly straight line obviates any type of flaring phenomenon. Add to that the fact that there was (literally) (probably) not a single human within 10 to 25 miles of my location. Believe it or not.

Aug 16, 2023 at 08:19 AM

RoamingScott
Offline
• • • • • •
Upload & Sell: Off

p.1 #14 · p.1 #14 · Sensor error, or valid data? (5Ds high ISO)

the color shift along the path of it makes me think it was some naturally occuring phenomenon. it almost appears to "pulse" in various wavelengths as it goes. as many have said, it just doesn't fit the normal type of sensor artifact hallmarks.

Aug 16, 2023 at 08:22 AM

snapsy
Offline
• • • • • •
Upload & Sell: On

p.1 #15 · p.1 #15 · Sensor error, or valid data? (5Ds high ISO)

Jeff wrote:
So, reading between the lines, you'd think this has to be an optical phenomenon, not electronic, despite however odd/unlikely it may be?

That wasn't my intention but I didn't address it nonetheless.

It looks like a shooting star trail but that doesn't seem possible because since it starts and ends below the horizon over a 13s period. It would be helpful to induce what direction the light emitter is moving, as it would provide some clues of possible sources. The concentration of pixels is higher near the bottom of the line but I'm not sure if that's a clue on the direction of travel.

Aug 16, 2023 at 08:26 AM

snapsy
Offline
• • • • • •
Upload & Sell: On

p.1 #16 · p.1 #16 · Sensor error, or valid data? (5Ds high ISO)

RoamingScott wrote:
the color shift along the path of it makes me think it was some naturally occuring phenomenon. it almost appears to "pulse" in various wavelengths as it goes. as many have said, it just doesn't fit the normal type of sensor artifact hallmarks.

The color shift is because it spans RGGB pixels and is only a pixel wide, thus the demosaicer renders it as different colors along the path.

Aug 16, 2023 at 08:26 AM

Jeff
Offline
• • • • •
Upload & Sell: On

p.1 #17 · p.1 #17 · Sensor error, or valid data? (5Ds high ISO)

snapsy wrote:
The color shift is because it spans RGGB pixels and is only a pixel wide, thus the demosaicer renders it as different colors along the path.

Do you have a representation (screenshot?) of it before the demosaicer gets hold of it that you could post? I'd be curious to see what that looks like.

Based on all of this, I'd guess that the single-pixel brightness level is fairly bright (compared to the black background obviously)? And, you did see that there was a pixel or two affected above the horizon along the exact same line?

Aug 16, 2023 at 08:32 AM

snapsy
Offline
• • • • • •
Upload & Sell: On

p.1 #18 · p.1 #18 · Sensor error, or valid data? (5Ds high ISO)

Jeff wrote:
Do you have a representation (screenshot?) of it before the demosaicer gets hold of it that you could post? I'd be curious to see what that looks like.

Based on all of this, I'd guess that the single-pixel brightness level is fairly bright (compared to the black background obviously)? And, you did see that there was a pixel or two affected above the horizon along the exact same line?

Here's a 300% crop showing the RGGB pixels without demosaicing, generated from RawTherapee using the RT's demosaicing method of "None (Shows sensor pattern)

Image

I thought I saw a trail above the horizon but can't find it now. Can you tell me the coordinates?

Aug 16, 2023 at 08:51 AM

Jeff
Offline
• • • • •
Upload & Sell: On

p.1 #19 · p.1 #19 · Sensor error, or valid data? (5Ds high ISO)

snapsy wrote:
Here's a 300% crop showing the RGGB pixels without demosaicing, generated from RawTherapee using the RT's demosaicing method of "None (Shows sensor pattern)

Image

I thought I saw a trail above the horizon but can't find it now. Can you tell me the coordinates?

Thanks for that. So, sure looks to me like something activated those sensor pixels?

The other spot is in the cloud above, and actually your demosaiced image shows there may be another (more faint) one in between the cloud and horizon:

Aug 16, 2023 at 08:59 AM

kirbic
Offline
•
Upload & Sell: Off

p.1 #20 · p.1 #20 · Sensor error, or valid data? (5Ds high ISO)

It's interesting that it is *exactly* one px wide... except where it isn't. There is that one spot where it is spread across four px, and perhaps a couple other spots where there is some smaller spread (though perhaps not after looking at the non-demosaiced image)
If this were an actual light source at distance, I can't imagine it being that sharp. Neither the horizon nor the stars are quite that well defined.
One thing to note, even though this was a 13s exposure, that doesn't necessarily mean that this feature took 13s to track across the corner of the frame. It could have been a momentary phenomenon.
It is exactly, precisely straight, at least at the limit of what can be judged in Ps, which does place some constraints on what it might be

Aug 16, 2023 at 09:07 AM

1 2 end
FM Forums \| Canon Forum \| Join Upload & Sell 1 2 end

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.