Originally Posted by markmon1
I'm not convinced. The majority of the processing power is already being done to measure the frames to enact dynamic laser dimming on a frame by frame basis. They also manipulate gamma during laser dimming already. I think they have the processing to do DTM to some degree if they can do laser dimming with gamma manipulation. Its very similar from the processor perspective. What they may lack is the needed memory to implement the feature. But I suspect they could can a few things no one really cares about similar to what they did on the NX line with keystone.
Purely speculative, I'm not familiar with type of software they use in those, and if they source them from a standard type board that is in heavy circulation, or if they just get them from generic Asia factories and program the firmware in. There are so many differences depending on HW/SW implementation, you never know.
Every time you modify an image in memory, you often have to keep the original image, and then often re-apply all the modifications in a row. The reason is because certain types of image enhancements are destructive.
For instance, when I write code to change the brightness of an image, what happens if you change the RGB all the way to 255,255,255,255. The image is essentially gone, it is blown out, and when it loses highlight, it loses information. So the solution is you often have to work with multiple matrices of data in order to maintain performance, but now-a-days this is less common, so now-a-days we can often avoid complex matrix shifting to alter the values and just use the tables themselves or the RGB values, or however you wish to do it. These image manipulation algorithms can be expensive, it just depends on what is available in the specific GPU implementation, and how they are shifting the values.
However, I would have no idea how that firmware worked or what kind of performance modding the data might cause, and I doubt anyone else would either, unless they worked for JVC...
Just my 2.5 cents...