Quote:
| My concern would not be those fixes but the actual quality of the deinterlacing algorithms on the chip written by the Terranex folks, some of whom are no longer with SO. How good these are will have to await full testing after product release. Assuming they are less than fully ideal as some who have seen deinterlacing performed by various Terranexes suggest, will SO want to improve them and then be able to improve them and transfer those improvements to the chip. Let's just say certain things on the chip for one reason or another are rumoured to be less than fully accessible. |
Mark: it's Teranex, not Terranex. ;)
"Less than fully ideal"? Teranex's video processing was nothing short of
amazing from the very get-go. Don't take my word for it, check out this post from way back in 2001:
The New Standard
The original Xantus was an overnight success because it delivered the best PQ to those who absolutely required it (i.e. broadcast, post-production, etc). Teranex actually
guaranteed superior PQ and was able to do so because of the Xantus' software-updatable (and fully scalable) architecture. So when Mr. Hollywood Editor discovered a problem, he sent the content to Florida where the Teranex guys looked at it, fixed it, and then sent the updated software right back. Thanks to years of this kind of intensive content verification, the de-interlacing algos have been continuously refined much like what Porsche has done with the 911 (our favorite metaphor).
The Realta was born with these exact same algos which, as you suggested, will continue to be improved upon based on feedback provided by our customers in both professional and consumer markets as well as you guys. Keep in mind that these are improvements to the de-interlacing
algorithms. Software updatability (i.e. true programmability versus mere firmware updates) is one of Realta's major differentiating factors.