Well I bought into HD DVD and have about 20 discs but looking around I've probably purchased 500 standard DVDs.

I was considering how much money it would take to replace all of my discs in HD, assuming that they ever all came out when I came across some talk about the toshiba spursengine and upconversion. It apparently uses super resolution which is a technique that uses multiple frames in a video to create a resolution higher than the original still frame. I would love a SD player that can produce true HD from my old DVDs.

Super resolution is a technique used in astrophotography so I know that it exists. Typically, it takes about 9 frames to double the resoluton of the sensor. I also know that the software is very computer intensive and what I've seen produces one picture at a time. Did anyone see the demonstration. I'm curious about how much the resolution is actually increased and if there any noticable artifacts.

I really don't think I will be able to talk my wife into another HDM player anytime soon but something that will make our old DVD's look better would probably be an easy sale.

I'm assuming that this is the right area to post.

I can see taking multiple video frames and coming up with a higher resolution still frame. But I'm really skeptical about taking x number of video frames "look ahead" and constructing a single video frame and keep doing this constantly to produce "upconverted" video. At the very least it would take a lot of processing power, like you said. Do you have a link that describes in detail how this is done?

larry

There is no link from Toshiba that I have seen. I've seen a couple of videos and still pictures that demonstrate the effect but as you can imagine internet video is less than usefull.

Apparently, the processor is cell based and will be made at the plant they bought from Sony.

Their market is computers as it offloads a number of video functions, DVD players and even built into HDTV's to convert SD material.

I can picture how it works, it converts the info that is stored temporaly and converts it into a static image. If your computer were fast enough, theoretically you could do it 30 FPS at basically realtime. I can imagine that there would be some diference in resolution within a picture as objects move on and off screen.

I was curious if anybody here went to the Toshiba booth. It probably wasn't the most visited place with the recent turn of events.

I guess, if it works, it would help make SD look better, but I don't see any substitute for more bits in the source. Maybe this tech could be used to show HD (2K) on future 4K displays. 4K sources will take a lot of space.

larry

I guess, if it works, it would help make SD look better, but I don't see any substitute for more bits in the source. Maybe this tech could be used to show HD (2K) on future 4K displays. 4K sources will take a lot of space.

larry


Astronomers have been using the technique for years. When you supersample, take multiple pictures, there is more information recorded than each single picture holds. I've seen some very nice supersampled pictures using a webcam as the source.

I have a little understanding of information theory and the concept makes sense but I would really like to see a true example of what it can do with video.

OK, I may be talking to myself at this point. I down loaded a program that us supposed to use super resolution on video.

I does better that upconversion on straight lines but honestly, I could perceive no increase in real resolution.

If Toshiba does no better, I am dissapointed.

I saw a web cam video of Toshiba showing a demonstration video at CES. I was really hoping that someone here had seen the presentation and could give their opinion.