rscheffler wrote:
Dan, I think it's one extremely hot pixel that is somehow overloading the column and resulting in an artifact that appears as a vertical line, but I have no scientific way to explain it.
The sensor is read out in rows so something is causing it to 'glitch' at the same point along the horizontal run as the rows are read out. Maybe @snapsy@ or Jim Kasson could provide a correct explanation. But it would only explain why and be of no benefit to the OP with their dealings with Canon.
Yep, the readouts are oriented on rows, so column-based anomalies are uncommon. @Leah Hallett, which orientation was the photo taken in, landscape or portrait?
Here are three raw files, hopefully they will download for you. Two days before these shots the defect wasn't there. Copy and paste each into address bar:
Leah Hallett wrote:
I really don’t think a camera store (Fort Worth Camera) would look up a receipt by serial number or inventory number, if that’s what that is on their sticker on the box, and give a copy of it to someone whose name isn’t on the receipt? Right
Response to my PM asking about getting a copy of the receipt:
“Hi I’m sorry to hear that happened, but I don’t have the receipt for it. I buy and resell gear, this one didn’t come with a receipt.”
The seller's ad stated it came with 'paperwork.' Since there isn't a receipt, I assume the paperwork refers to at least the Canon warranty card? If so, is it a USA/Canada warranty card? If so, then at least the camera was 'officially' imported into the market by Canon USA (or Canon Canada). IOW, it wouldn't be a gray market product and would normally receive warranty service, which is a good thing.
You could try contacting Canon and say the camera was received as a gift and unfortunately the purchaser no longer has the receipt, is out of the country for X amount of time, moved out of the country (everything is in storage/moving boxes), recently passed away, etc., etc. But you should still contact the store. Explain to them the situation (you received the camera from someone who apparently bought it there but no longer has the receipt). They are not machines and should have some leeway in helping you out if the camera was legitimately purchased from them. If they don't want to reprint the original receipt with the purchaser's info on it, can they provide you with a generic one? I think all that matters from Canon's POV is that the receipt includes the serial number.
It's a bit of legwork, but the alternative is probably having to pay out of pocket for an expensive sensor replacement even though the camera should still be in warranty coverage (for the original purchaser).
The gray zone, IMO is the seller you bought it from. He's in the 'business of buying and reselling.' That suggests to me he maybe got the camera from someone else and not from the store. This could be totally legit, or it might not be...
It's also a good warning for all of us with respect to being a bit more skeptical about 'new in box, zero activations' types of B&S listings.
I’m thinking that the laser theory is extremely unlikely since the problem pixels all seem to be in the same physical column on the image. Aligning a laser that perfectly during a performance, especially while hand-holding the camera is not impossible, but it is about as close to impossible as you can get and not be there.
I think this also makes physical damage (for example by something striking the sensor surface) equally unlikely.
My bet is that this is a sensor issue, though that could include some kind of issue with the connections to the sensor in the camera. It is weird that it skips over pixels though — when I’ve seen stuff like that it is usually a whole row or half of a whole row.
Also, when I’ve seen issues caused by bad cards (or bad connections to cards) they have looked a whole lot different than this — often multiple horizontal lines (not vertical) and with odd patterns.
In any case, that camera is either going back to the seller (assuming that it arrived with this issue) or to Canon for warranty service. I would at least contact the seller and have a conversation about it — though if you have had the camera for a while without noticing it (even if because you did not use the camera much) it may be too late to say/do much about it.
Perhaps you should educate yourself on the subject?
Here's one of many reports of low wattage laser damaging sensors. Take special note of the damage pattern at 1:45
And another segment at 3:20 when he's talking to CPS about the damage.
And just for the record, I was simply tossing some ideas out there that might have explained how it could have happened, which seems to not be the case based on the OP's comments.
Here are three raw files, hopefully they will download for you. Two days before these shots the defect wasn't there. Copy and paste each into address bar:
Got 'em, thanks. The raw files were very revealing. It's a repeating grid of 3x2 bad-readout pixels, all on the same Green_2 bayer channel, with unaffected pixels on other channels between them. These two bad-readout rows are followed by 6 correctly-read rows, then the 3x2 bad-readout pixels repeats.
This is very unusual and indicates the issue is unlikely to be damage to the sensor surface area itself (like a laser) but instead a malfunction in an electrical readout component that's specific to that color channel. The affected pixels even extend beyond the active sensor area into the optically-masked pixels, which further supports the cause being a readout component on the sensor.
The center column of the 3x2 bad area is full-white (saturation) on one raw, then darker than normal on another.
Here is an animation between two of the raws, showing only the Green_2 pixels. You'll see two pixels stuck to saturation vertically on top of each other, surrounded by darker pixels also affected:
To demonstrate why this is so unusual, here's a diagram of the RGGB bayer sensor layout, with an "X" drawn on the 3x2 bad-readout pixels on the G2 channel:
If you notice on this diagram, the Blue pixels sandwiched between the affected "X" G2 pixels are fine. If something external damaged the sensor in that area then it would've damaged those blue pixels as well, along with the red pixels in the row between them.
Wow, thank you for all your effort figuring this out. Do you mind if I try to send your info to Canon repair to help them more quickly fix it? Not sure if there’s a way to add info to my repair submission now, but I’d like to try if it’s ok with you. I’m amazed at your ability!
lighthound wrote:
Perhaps you should educate yourself on the subject?
Here's one of many reports of low wattage laser damaging sensors. Take special note of the damage pattern at 1:45
And another segment at 3:20 when he's talking to CPS about the damage.
And just for the record, I was simply tossing some ideas out there that might have explained how it could have happened, which seems to not be the case based on the OP's comments.
&ab_channel=AlainMartinez
Thanks for this. The damaged pattern @ 1:45 you referenced definitely looks like the OP's damage, down to even the repeating blocks of pixels with unaffected regions between them. That might have me rethink my theory that this isn't laser damage for the OP's camera based on only one color plane affected. Perhaps the specific wavelength of the sensor light (green in this case) is why only that column was affected, with the other pixels protected by their non-green bayer color filter that prevented the laser energy for that green wavelength from reaching to the pixel wells underneath.
Leah Hallett wrote:
Wow, thank you for all your effort figuring this out. Do you mind if I try to send your info to Canon repair to help them more quickly fix it? Not sure if there’s a way to add info to my repair submission now, but I’d like to try if it’s ok with you. I’m amazed at your ability!
Leah
Sure. But just note that after I posted my analysis I watched the video shared by Lighthound and based on the damage I could make out after YT compression I revised my theory that it could actually be laser damage.
You mentioned in your OP that it's not on every image. Was there a time that it wasn't on any image, and then one day started showing up regularly? If that's the case then it would add weight to the component malfunction theory. However if it's been there on/off since you originally bought the camera then that would add weight to the laser theory.
It never showed up on any image until the day I shot a Great Blue Heron on the island in my pond. I was shooting a Pileated Woodpecker two days before and it wasn’t there - then it was on the first pic two days later. I was not around any lasers and the sun wasn’t near either of these birds. I didn’t notice it until shooting the Belted Kingfisher that was at the pond. It’s on every shot now.
Leah Hallett wrote:
It never showed up on any image until the day I shot a Great Blue Heron on the island in my pond. I was shooting a Pileated Woodpecker two days before and it wasn’t there - then it was on the first pic two days later. I was not around any lasers and the sun wasn’t near either of these birds. I didn’t notice it until shooting the Belted Kingfisher that was at the pond. It’s on every shot now.
Leah
Good. I asked that in consideration of the laser damage possibly occurring during the camera's previous ownership. If you've owned and used the camera for a good amount of time and never saw the issue before that day then that certainly adds weight for my first theory of it being a component failure.
lighthound wrote:
Perhaps you should educate yourself on the subject?
Perhaps you should try being civil.
I did not say that lasers cannot cause damage to sensors. I'm aware of that possibility and have been for some time.
Go back and read what I wrote again and why THIS particular example is almost certainly not caused by that. It is important to the OP to understand that.
The specific image issues on THIS camera are more consistent with some kind of sensor-related problem.
(I now see that the OP has confirmed that the camera was not exposed to lasers when the problem began.)
Leah Hallett wrote:
I really don’t think a camera store (Fort Worth Camera) would look up a receipt by serial number or inventory number, if that’s what that is on their sticker on the box, and give a copy of it to someone whose name isn’t on the receipt? Right
Response to my PM asking about getting a copy of the receipt:
“Hi I’m sorry to hear that happened, but I don’t have the receipt for it. I buy and resell gear, this one didn’t come with a receipt.”
Perhaps use the serial number to check if it’s been registered with Canon. As a caution to buyers looking for a good deal for a “new” camera, always insist on the original receipt as it could be stolen goods. Inquire if there’s any warranty and if they’re the original buyer. I’m hesitant to purchase gear here from re-sellers as opposed to BH, Adorama, etc. because of customer service.The seller’s description for your camera’s ad is a bit deceptive and seems to imply that it’s their own gear. Never understood the usefulness of printing someone’s name and date in the photos as some kind of firewall protection for security.
Leah Hallett wrote:
I bought it as “brand new in box” and the defect was not there before that day. Thanks again!
Good luck with the servicing! I would push for an in-warranty repair even though you don't have a receipt proving the original date of purchase is within the Canon USA 1-year warranty. The R5 II was announced on July 17th, 2024 and didn't start shipping for some time after that, which means all USA bodies in circulation should be within their 1-year period. I would argue that if Canon gives you any resistance.
I didn’t read the whole thread, but did you ask the seller where they purchased the camera? New cameras with zero shutter clicks should always have proof of purchase. If not, it’s probably stolen goods. Original receipts as with original shipping boxes have value in resale. Saving a hundred dollars or two is enticing, but buying from authorized Canon dealers is always better for the customer.
I registered the camera when I filled out the Canon Repair info. Would it let me register if it had already been registered? The camera box has a sticker from where it was purchased, see below. I'm pretty sure I checked the shutter count after I got it to make sure it had none - but I can't remember now. It has a USA/Canada warranty card in the box and everything was sealed. I don't believe he is selling stolen goods!
snapsy wrote:
Got 'em, thanks. The raw files were very revealing. It's a repeating grid of 3x2 bad-readout pixels, all on the same Green_2 bayer channel, with unaffected pixels on other channels between them. These two bad-readout rows are followed by 6 correctly-read rows, then the 3x2 bad-readout pixels repeats.
This is very unusual and indicates the issue is unlikely to be damage to the sensor surface area itself (like a laser) but instead a malfunction in an electrical readout component that's specific to that color channel. The affected pixels even extend beyond the active sensor area into the optically-masked pixels, which further supports the cause being a readout component on the sensor.
The center column of the 3x2 bad area is full-white (saturation) on one raw, then darker than normal on another.
Here is an animation between two of the raws, showing only the Green_2 pixels. You'll see two pixels stuck to saturation vertically on top of each other, surrounded by darker pixels also affected:
To demonstrate why this is so unusual, here's a diagram of the RGGB bayer sensor layout, with an "X" drawn on the 3x2 bad-readout pixels on the G2 channel:
If you notice on this diagram, the Blue pixels sandwiched between the affected "X" G2 pixels are fine. If something external damaged the sensor in that area then it would've damaged those blue pixels as well, along with the red pixels in the row between them....Show more →
It appears to be 16 rows until the defect repeats again. I'd guess 16 rows are read out with each 'sweep' of the sensor and there is a defect with a specific component in the readout cycle that causes the repeated pattern.
Any idea for the one very hot pixel (or small area of pixels)?
