[Bizarr]

Can you post a link to a raw?

https://drive.google.com/open?id=1_TJK90bxtUodfm9RQ9EV9m0dwxc87Y6X

[UncleVanya]

https://drive.google.com/open?id=1_TJK90bxtUodfm9RQ9EV9m0dwxc87Y6X

I'll look at it in the morning. Not sure what I expect. Clutching at straws.

[photoptimist]

Now left side. Sometimes it happens, sometimes not, but most often it happens when camera is in vertical position. IBIS failure ? Just guessing.

---------- Post added 03-18-19 at 12:11 PM ----------


Nothing. Everything looks fine.

That image proves it's not sensor damage. But it definitely looks like a reflection of the image of the sun off the walls of the mirrorbox.

[UncleVanya]

We appear to be playing STUMP the CHUMP. (Cartalk reference)

I zoomed in and can see that weird patterning on some but not all of the pink. I really wonder what caused that pattern.

[photoptimist]

We appear to be playing STUMP the CHUMP. (Cartalk reference)

I zoomed in and can see that weird patterning on some but not all of the pink. I really wonder what caused that pattern.

Here's my guess:

The black plastic of the mirror box is not perfectly black in the near-IR (which can happen if the "black" is made from tiny particles which would tend to absorb shorter-wavelength light better than longer wavelengths). Thus, the reflection of the sun is blindingly bright in the near-IR.

The IR filter of the camera isn't perfect so a tiny % of the near-IR gets to the sensor. (Remember that the sun is over 100,000X as bright as the scene so if even 0.001% gets through, it's going to be very bright.)

The different colors of the Bayer CFA also let in some near-IR which creates a very strange signal.

The demosaicer is seeing some impossible numbers in the different color channels which it ends up interpreting as bright magenta but the odd-even row-column patterns of the RGB filter pattern end up creating odd-even fluctuations in demosaiced values.

[UncleVanya]

The demosaicer is seeing some impossible numbers in the different color channels which it ends up interpreting as bright magenta but the odd-even row-column patterns of the RGB filter pattern end up creating odd-even fluctuations in demosaiced values.

Shouldn't this mean different raw processors would give different results?

---------- Post added 03-21-19 at 03:22 PM ----------

Edit: I'm just asking. I'm lost. I follow the idea but have no background in this problem.

[photoptimist]

Shouldn't this mean different raw processors would give different results?

---------- Post added 03-21-19 at 03:22 PM ----------

Edit: I'm just asking. I'm lost. I follow the idea but have no background in this problem.

This might help explain the basics: Understanding Digital Camera Sensors

If you look at the layout of the Bayer color filter, all the odd-numbered rows consist of red and green pixels and all the even numbered rows consist of blue and green pixels. A similar odd vs. even effect occurs for columns. If there are large anomalies in the color data, the demosaicer can create a strange alternating row or alternating column effects.

[UncleVanya]

This might help explain the basics: Understanding Digital Camera Sensors

If you look at the layout of the Bayer color filter, all the odd-numbered rows consist of red and green pixels and all the even numbered rows consist of blue and green pixels. A similar odd vs. even effect occurs for columns. If there are large anomalies in the color data, the demosaicer can create a strange alternating row or alternating column effects.

Thanks.

[Bizarr]

This might help explain the basics: Understanding Digital Camera Sensors

If you look at the layout of the Bayer color filter, all the odd-numbered rows consist of red and green pixels and all the even numbered rows consist of blue and green pixels. A similar odd vs. even effect occurs for columns. If there are large anomalies in the color data, the demosaicer can create a strange alternating row or alternating column effects.

Ok, so... is it broken ? It's still under warranty. I dont think its normal.

[photoptimist]

Ok, so... is it broken ? It's still under warranty. I dont think its normal.

The only abnormality might be heat or burn damage to the inside the mirror box which will not be covered by the warranty. If the camera spent some time on a tripod pointed toward the sun, the plastic inside the mirror box may have melted, become shiny, and now reflects a lot more stray light.

But it's more likely that this is an unavoidable consequence of the sun being too close to the frame and the lens not having a deep enough hood to keep direct sunlight out of the camera.

[Bizarr]

The only abnormality might be heat or burn damage to the inside the mirror box which will not be covered by the warranty. If the camera spent some time on a tripod pointed toward the sun, the plastic inside the mirror box may have melted, become shiny, and now reflects a lot more stray light.

But it's more likely that this is an unavoidable consequence of the sun being too close to the frame and the lens not having a deep enough hood to keep direct sunlight out of the camera.

It has never been on a tripod If it was exposed for sun, its just for few seconds. Hood was always on. Normal use, nothing out of ordinary. I may say that I observe this phenomenon since day 1 i bought this camera.

[UncleVanya]

The only abnormality might be heat or burn damage to the inside the mirror box which will not be covered by the warranty. If the camera spent some time on a tripod pointed toward the sun, the plastic inside the mirror box may have melted, become shiny, and now reflects a lot more stray light.

But it's more likely that this is an unavoidable consequence of the sun being too close to the frame and the lens not having a deep enough hood to keep direct sunlight out of the camera.

It has never been on a tripod If it was exposed for sun, its just for few seconds. Hood was always on. Normal use, nothing out of ordinary. I may say that I observe this phenomenon since day 1 i bought this camera.

I routinely shoot the sun in frame using a DA 15 (and some other lenses) without seeing this result. The pink color is really strange.

I wonder if a warm colored flood light would do it also. It might be worth some test shots.

[photoptimist]

I routinely shoot the sun in frame using a DA 15 (and some other lenses) without seeing this result. The pink color is really strange.

I wonder if a warm colored flood light would do it also. It might be worth some test shots.

This problem seems more likely linked to images where the sun (or extremely bright light source) is just outside of the frame rather than in the frame. Extremely bright sunlight is entering the lens, hitting the wall of the mirror box and reflecting back on to the sensor. And it might be more an issue with some specific models of cameras depending of the specific materials and coating used inside the mirror box as well as the exact spectral transmission properties of the IR-cut filter and sensor pixel color filters.

You are right that the effect should also show up with a bare tungsten light bulb that are just out of the frame.

P.S. If you search the internet, you'll find that other cameras have this same problem. In the case of Fuji X-Trans, the problem even includes a strange grid-pattern in the flare ( Why and How Fuji Cameras Produce a Strange Purple Flare/Grid Artifact ). The grid pattern with Fuji cameras is different from the one in the KP because the X-Trans color filter is different from the Bayer design used in the KP.

[reh321]

Hello, as title suggest - i`ve got problem with pinkish sun on my photos when directly pointing my KP towards sun Any suggestions ?

At what ISO was this photo taken?

[UncleVanya]

This problem seems more likely linked to images where the sun (or extremely bright light source) is just outside of the frame rather than in the frame. Extremely bright sunlight is entering the lens, hitting the wall of the mirror box and reflecting back on to the sensor. And it might be more an issue with some specific models of cameras depending of the specific materials and coating used inside the mirror box as well as the exact spectral transmission properties of the IR-cut filter and sensor pixel color filters.

You are right that the effect should also show up with a bare tungsten light bulb that are just out of the frame.

P.S. If you search the internet, you'll find that other cameras have this same problem. In the case of Fuji X-Trans, the problem even includes a strange grid-pattern in the flare ( Why and How Fuji Cameras Produce a Strange Purple Flare/Grid Artifact ). The grid pattern with Fuji cameras is different from the one in the KP because the X-Trans color filter is different from the Bayer design used in the KP.

OK this is finally making sense to me. The KP like the Fuji mentioned has on sensor PDAF (not sure if these are used in any mode but they are there) this creates the mask that causes the shadow that can result in this purple/pink flare. The K-3 I have does not have this mask (nor did the K-50 or K100D Super before it. So I literally have had no opportunity to see this in the past.

@bizarr this is apparently a consequence of on sensor PDAF sensors. The KP sensor from all indications has these even if it isn't clear when they are used (maybe during AF-C in liveview/video?). The presence of these appears to cause this (see the petapixel article about fuji purple flare for an explanation. This does not appear to be anything wrong with your KP - but it does appear to be something that is specific to that camera and the K-70 which the K-3/K-3ii would not exhibit due to the lack of PDAF embedded sensors in those K-3 models.

---------- Post added 03-22-19 at 05:42 PM ----------

Thank you @photoptimist for getting me there.