Originally Posted by jodusc
Maybe I've missed some details, as this thread is very long, but it seems that this is a confirmed issue with all sets once they reach 500 hours. My set is way below that, so I'd LOVE for Panasonic to develop a fix before my set's voltage bumps up to compensate for panel degradation. I think you call that being proactive.
First time to this formum, and what an eye openner. I got my first plasma, a TC-P42G10, over the weekend. New out of the box, I went online to Panasonic to register, saw there was a firmware upgrade to 1.28, tried to figure out how to read the software version code on my set and ended up here.
Three hours of reading later, and I'm reading the same conclusion as jodusc, that perhaps if the software hasn't started the voltage ramp-up yet, then perhaps there should be a way to fix this with a software patch. My set was manufactued in Sept. 2009, and came with software version 1.24.
I'm guessing the complexity with patching the existing sets is that there are probably individual "life remaining" memory values for each pixel, which are being plugged into the voltage formula so that "older" pixels end up with stepped up voltages as needed.
The voltage step-up function is probably good for 99% of all mulit-color pixel aging situations, so a global fix to the function would screw up the way the main picture appears. The only way to patch a fix that would work on black only would be to adjust the weighted increments to the counter variables for pixels that are in "black" mode. However, just looking at the counter variables would not allow one to tell which pixels had been aged in "black mode" and which had aged in "normal use" mode.
Someone whose use counters are all very low could apply a patch to the weighting function and not have negative effects. Even though in both cases, one could not identify which counters to set back and which not to, setting back counters that are all relatively nil would not cause unintended consequences...
If I am thinking about this possibly implementation correctly, it would also explain why this type of aging is noticable mainly to people viewing letterbox stuff. In the normal state of the world with pixels doing other stuff than just black, the application of the voltage function would not be as noticable, even if there was a problem with the way it was working for "active black" time use.
I am more of a math / programming guy... I have no background in image stuff, but here's an example of how it might work:
Un = active use time for pixel n
Vn = F(Un) = activation voltages for pixel n
For each hour that pixel n is active, the software increments up Un in a weighted fashion. So if the pixel is showing 100% contrast, white, then, for example, Un increments by 100 units per hour. For time which the pixel is at 50% contrast, showing some mix of R/G/B, a weighted value would be computed and added to Un.
What has probably happened is that the weighting function for black (which, something I remember suggested that "true black" is implemented by panasonic by turning the pixel completely off), is adding too much to the Un value. This may be because of an assumption that most users will be viewing with picture cranked up to 100 for a glowing screen, or it could be something as dumb as plugging in a Pioneer algorithm for screen phosphors that age differently due to different manufacturing quality tollerances, etc.
But once you have Un computed, its impossibly to tell whether the value should be reduced or not, because you don't know whether it is the result of overstated "black", or normal color combinations.
It migh be possible to patch by pixel address in a way that would work to fix the black bars on top and bottom of the screen only, but it would again be difficult to apply across the board because the person who has watched 100% letterbox films and the person who was only watching 50% letterbox films would need those values reduced by different amounts in order to stay within the pixel/plasma life paramaters that the voltage increase function is here to account for in the first place.
So, again, even though this is mostly guess work, it would seem that any solution rolled out for existing sets will only work on sets that have very very low hours on them, and/or would need to be rolled out on sets with a higher number of hours on them accounting for situation specific variables in each case, meaning prohibitive cost.
Which leads me to the real point: can this problem be reduced by the so-called "break in" process? It would seem that it should at least reduce the amount of "black" contribution to the use variables as the voltage adjustments are made. I wonder how many people who are experiencing this problem actually went through the 100 hour break-in DVD prodedure, etc...