Olympus C5050 pixel remapping failure

My Olympus C5050 seems to have developed a hot pixel - a white spot that
shows up fairly clearly. Unfortunately, when I try the pixel remapping
function of the camera, the spot remains.

It appears that this spot is a square of four pixels, appearing all white.
This is surrounded by a 1 pixel wide border that is darker than the area
around it. So, I have a 4 x 4 pixel bad spot, though I wonder if the dark
area may be a result of the bright 2 x 2 area in the center.

Looking back at old pictures, it apparently appeared while the camera was
still under warranty, had I discovered it then.

Unless someone can tell me of tricks to get the remapping to work better,
I guess it is off to pay Olympus a huge amount of money. Sigh.


(No, I am not ready to replace it. When I spent that much money, I expect
it to last a bit longer than this.)


Alan

"Alan" wrote

My Olympus C5050 seems to have developed a hot pixel - a white spot that
shows up fairly clearly. Unfortunately, when I try the pixel remapping
function of the camera, the spot remains.

It appears that this spot is a square of four pixels, appearing all white.
This is surrounded by a 1 pixel wide border that is darker than the area
around it. So, I have a 4 x 4 pixel bad spot, though I wonder if the dark
area may be a result of the bright 2 x 2 area in the center.

It's not 4 bad pixels, it's only 1 bad pixel. The way the Bayer sensor works
is to interpolate color values from neighboring pixels. 1 hot pixel will
always show up as at least 4 hot pixels by the time the image is convtered
to a RGB format.

Regardless, the whole point of pixel mapping is to get rid of those
problems. I'm surprised it doesn't work. I'd give Olympus hell for selling
you a broken feature.

In article ,
says...

My Olympus C5050 seems to have developed a hot pixel - a white spot that
shows up fairly clearly. Unfortunately, when I try the pixel remapping
function of the camera, the spot remains.

It appears that this spot is a square of four pixels, appearing all white.
This is surrounded by a 1 pixel wide border that is darker than the area
around it. So, I have a 4 x 4 pixel bad spot, though I wonder if the dark
area may be a result of the bright 2 x 2 area in the center.

It's not 4 bad pixels, it's only 1 bad pixel. The way the Bayer sensor works
is to interpolate color values from neighboring pixels. 1 hot pixel will
always show up as at least 4 hot pixels by the time the image is convtered
to a RGB format.

Obviously it's not only one pixel, otherwise pixel mapping would take
care of it, wouldn't you agree ? Pixel mapping maps out bad pixels by
using the neighbouring pixels.

Regardless, the whole point of pixel mapping is to get rid of those
problems. I'm surprised it doesn't work. I'd give Olympus hell for selling
you a broken feature.

It's more likely that behind those four damaged pixels there is a huge
block of broken pixels, so huge that even pixel mapping is overwhelmed.

In any case this doesn't help the original poster. The only thing he
might try is to leave the camera without batteries for 24 hours, so that
it fully resets to the factory default. Then try again the pixel mapping
and if that still doesn't help send it to Olympus.
--

Alfred Molon
------------------------------
http://groups.yahoo.com/group/Olympus_405080/
Olympus 5060 resource - http://myolympus.org/5060/
Olympus 8080 resource - http://myolympus.org/8080/

"Alfred Molon" wrote

Obviously it's not only one pixel, otherwise pixel mapping would take
care of it, wouldn't you agree ? Pixel mapping maps out bad pixels by
using the neighbouring pixels.

But the symptoms he describes are exactly what one would expect to see from
a single hot sensor. That creates a block of 4 hot pixels, bue to Bayer
processing of the image. Add some jpeg artifacts to it, and you've got a
ring around that block.

It sounds more like pixel mapping isn't working on his camera, or at least
isn't working on that specific hot pixel.

It's more likely that behind those four damaged pixels there is a huge
block of broken pixels, so huge that even pixel mapping is overwhelmed.

Possibly, although that's not what he describes.

In any case this doesn't help the original poster. The only thing he
might try is to leave the camera without batteries for 24 hours, so that
it fully resets to the factory default. Then try again the pixel mapping
and if that still doesn't help send it to Olympus.

Could work. Maybe the buffer that holds the hot pixel information is full.
I'm not sure if a reset would clear that, or not.

"Alfred Molon" wrote

Obviously it's not only one pixel, otherwise pixel mapping would take
care of it, wouldn't you agree ? Pixel mapping maps out bad pixels by
using the neighbouring pixels.

But the symptoms he describes are exactly what one would expect to see from
a single hot sensor. That creates a block of 4 hot pixels, bue to Bayer
processing of the image. Add some jpeg artifacts to it, and you've got a
ring around that block.

It sounds more like pixel mapping isn't working on his camera, or at least
isn't working on that specific hot pixel.

It's more likely that behind those four damaged pixels there is a huge
block of broken pixels, so huge that even pixel mapping is overwhelmed.

Possibly, although that's not what he describes.

In any case this doesn't help the original poster. The only thing he
might try is to leave the camera without batteries for 24 hours, so that
it fully resets to the factory default. Then try again the pixel mapping
and if that still doesn't help send it to Olympus.

Could work. Maybe the buffer that holds the hot pixel information is full.
I'm not sure if a reset would clear that, or not.

In article ,
says...
"Alfred Molon" wrote

Obviously it's not only one pixel, otherwise pixel mapping would take
care of it, wouldn't you agree ? Pixel mapping maps out bad pixels by
using the neighbouring pixels.

But the symptoms he describes are exactly what one would expect to see from
a single hot sensor. That creates a block of 4 hot pixels, bue to Bayer
processing of the image. Add some jpeg artifacts to it, and you've got a
ring around that block.

Actually Bayer colour interpolation is a bit more complex than that. It
uses more than just four adiacent pixels. See for instance this

R1 G1 R2 G2 R3 G3
G1 B1 G2 B2 G3 B3
R4 G4 R5 G5 R6 G6
G4 B4 G5 B5 G6 B6
R7 G7 R8 G8 R9 G9
G7 B7 G8 B8 G9 B9

To calculate the RGB values of the pixel at R5 you need to use the R
values at R2, R6, R8 and R4 - and it also wouldn't hurt to use the
remaining R values (R1, R3, R7 and R9).

In addition, without knowing how exactly the colour interpolation
algorithm is implemented, I suspect that also the neighbouring G and B
values are used to calculate the RGB value of the pixel at R5.

In any case there are more than just four pixels involved.
--

Alfred Molon
------------------------------
http://groups.yahoo.com/group/Olympus_405080/
Olympus 5060 resource - http://myolympus.org/5060/
Olympus 8080 resource - http://myolympus.org/8080/

In article ,
says...
"Alfred Molon" wrote

Obviously it's not only one pixel, otherwise pixel mapping would take
care of it, wouldn't you agree ? Pixel mapping maps out bad pixels by
using the neighbouring pixels.

But the symptoms he describes are exactly what one would expect to see from
a single hot sensor. That creates a block of 4 hot pixels, bue to Bayer
processing of the image. Add some jpeg artifacts to it, and you've got a
ring around that block.

Actually Bayer colour interpolation is a bit more complex than that. It
uses more than just four adiacent pixels. See for instance this

R1 G1 R2 G2 R3 G3
G1 B1 G2 B2 G3 B3
R4 G4 R5 G5 R6 G6
G4 B4 G5 B5 G6 B6
R7 G7 R8 G8 R9 G9
G7 B7 G8 B8 G9 B9

To calculate the RGB values of the pixel at R5 you need to use the R
values at R2, R6, R8 and R4 - and it also wouldn't hurt to use the
remaining R values (R1, R3, R7 and R9).

In addition, without knowing how exactly the colour interpolation
algorithm is implemented, I suspect that also the neighbouring G and B
values are used to calculate the RGB value of the pixel at R5.

In any case there are more than just four pixels involved.
--

Alfred Molon
------------------------------
http://groups.yahoo.com/group/Olympus_405080/
Olympus 5060 resource - http://myolympus.org/5060/
Olympus 8080 resource - http://myolympus.org/8080/