Originally Posted by SailorKane
I have to admit that a difference in image processing between the 732 and 831, particularly in deinterlacing 1080i to 1080p, would be significant to me. This is the first indication that there might be some actual difference in image processing between the sets (except the lamp). Mits has not admitted to any, but the factory information is so sketchy and misleading that I am not surprised by that.
The thing about 1080i processing is that we have no real 1080p source at this point. The HD sources so far come in as 720p or 1080i and must be converted to 1080p. (Bluray and HD-DVD may change that as this technology becomes mature and widely available). So, we are stuck with some kind of converting. There is no way around this right now. The STB that is available to me outputs in 720p or 1080i. I have to pick one. Knowing that the 831 is state of the art in this process is a really major thing to me. It means that native 1080i source, which is transmitted very well by the Verizon FIOS system, would be converted well by the Mits to its 1080p. And would change my opinion about the selection of the TV. The 1394 front input, the thin bezel, the negligible value of the 180w lamp, are NOT significant.
Personally, I've found the 1080i setting for the STB to be better than 720p for the 831, and I have posted this before. I also found my yamaha 2700, which purports to have a truly excellent chip set for upscaling and deinterlacing, to be identical to the Mits in performing that function, so I have continued to let the MITS do all the converting. It went counter to UMR's testing of the 732, so I felt a bit uncomfortable with my selection. It was my uncalibrated eyes vs UMR's equipment and skill. Now that UMR may have found a difference in the image processing of the 831, I feel a little better in my selection.
So, lets see what develops in the future. And what comments UMR can add to what he has already said. Hopefully he feels confident about additional posts. His is the voice of reason and empirical evidence that we really need here.
And I have sent another email to Mits to ask them about the image processing in their sets. Is there an engineering reason for improved processing. Hopefully they aren't tired of my questions and will answer this one too.
I have already decided not to return the 831. But this information will certainly make me feel better. It would turn my sense of the difference in value between 732 and 831 from about $100 to the retail price difference and justify the purchase. It might also help others to make the purchase decision between the two sets. Again, I think BOTH sets are excellent.
I think that it is very hard to come to clear generalized conclusions about what sets do what better or worse in terms of handling interlace, scaling, and pulldown. think about the worst case situation in your own system. If you are using the Ymaha, which is pretty good in its class, but not a very high end processor, you could have a lot of layers of conversion going on. For instance, ifyou are watching ESPN using your STB output at 1080i You have the interlace and scaling in the box, pass througn in the Yamaha with at nest no effect, then the set has to take that and de-interlace a signal that was not supposed to be interlaced to start with, AND scale it. Now consider a movie, mastered to an interlaced signal, deinterlaced and scaled to broadcast at 720p, interlaced and upconverted in the STB or the Yamaha to 1080i, then deinterlaced and scaled again in the set. Or take a test disk, mastered to DVD, de-interlaced to output a progresive signal, scaled and interlaced in the Yamaha, then deinterlaced and scaled again in the set.
Now sort all that processing out and are you surprised that some strange effects are occuring?
Lets not jump to too many conclusions without some clear facts and some actual test data to evaluate. We have neither in any significant quantities, so for now I suggest that people WATCH the sets and keep what they seen in the proper context, considering the processing path from the source. The bottom line is that we want to avoid conversions as much a possible uless one has a fairly sophisticated VP. Even then, it won't be perfect.
Obviously, there are trade-offs in processing in all sets. Overall, Mitsubishi seems to have made some reasonable ones based on my experience. The sets look quite good on most signals compared to competition,particualrly when calibrated.
Unfortunately, the answer to most questions on the matter come down to the one generic answer that can apply to most video systems. It depends...