Interview: Matrox Unleashes RT.X100
Page 3 of 4
DMN: Flex3D architecture is incorporated into this new chip, and it was also incorporated into the RT2500 chip, is that correct?
Legault: Yes.
DMN: So Matrox DigiSuite also has the Flex3D architecture in it as well, Right?
Legault: That's correct.
DMN: So now we might be able to extrapolate that since you've updated the RT2500 to this RT.X100, that we can assume that DigiSuite will also be updated.
Legault: I'm not willing to talk about specific dates or to pre-announce products, but I think it's a fair statement to say that we like to develop technology which is evolving from one product to the other. In other words, if you look at the history of our product -- I think you did a pretty good description already -- we developed DigiSuite initially, then we developed DigiSuite DV, and then took codec technology to our first RT product [RT2000], which we then developed Flex3D architecture which we brought back to DigiSuite through the MAX option. Now, we've developed the "Power of X" architecture, I think it's a safe assumption to expect Matrox to bring the best of that new technology back to DigiSuite, and we will come out with a new DigiSuite product that will leverage that architecture, the architecture being scalable. I think scalability can work in both directions -- it can work upward or downward. So I think you can expect us to come out with lower-end products as well as higher-end products, and when we think of higher-end, we're thinking of DigiSuite products.
Now Flex3D originally on the RT2000 and RT2500 was using a processor called the G400 from Matrox graphics. The latest product, the RT.X100 uses a G550, which has some improvements in terms of performance, allowing us to do, for example, two 3D perspective effects at once. The software-based effects are based on the standard API, which is basically like Direct3D, so it's something that we can leverage going to future 3D graphic chips, because it will run on it right away. So it's something that we will allow all of our users in the future to benefit, so all the code base we've written of all of these effects will migrate to the next products as they come out. And if this 3D graphics chip gets faster, you can do more streams, more layers, or bigger images. And when we think of bigger images, I guess we're thinking of HD, right?
[an error occurred while processing this directive]
DMN: Maybe that's what we have to look forward to, because you mentioned scalability, and then you mentioned DigiSuite. We can only assume that you're taking a look at HD for that.
Legault: Yeah, I think it's safe to say that. You have to realize RT.X100 is doing quite a lot of stuff, so in order to differentiate our DigiSuite product line I think we have to depart from what we have into much more stuff. And I think HD is a natural thing to look at.
DMN: Let's look ahead, way further than that. Say, five, six, seven years from now. What's the long-term trend, do you think, Alain? Do you think that soon all effects will be done via software when host processors just double and triple and quadruple in speed, or do you think there will always be a place for hardware-based effects as well?
Legault: Well, five years looks like a long time, Charlie (laughs). Let's say a couple of years from now. I think things are evolving definitively toward software. I think everything which can be done by a host processor will be done by the host processor, as people demand more sophisticated, higher-quality effects, with deeper pixel depth, such as 10-bit, and larger images such as HD, there will still be a challenge for a long time to have hardware devices in computers. I think those devices will continue to be the I/O technology which we will probably not see from PC manufacturers, other than 1394, I don't really expect computer makers to start putting analog video or SDI or even HDTV I/O on their motherboards. I think there will be a need for us to provide that to the user base.
And like I said earlier, 3D graphic chips are delivering a very high level of performance and are pretty inexpensive. Look at the latest 3D graphics chip Matrox has introduced -- it's an 80 million-transistor processor that has a memory throughput of 16GB of data per second. So, that kind of technology allows you to manipulate an extreme level of images, a very high level of layers all at once, and I think we can leverage that and bring more visuals with more impact in the future. I think for a person who is trying to make a living with his computer, a producer, I think he's going to want to buy an affordable device that can bring more visuals to his production. That needs to be nicely tied up with editing software that can also leverage the dual P4 5GHz that you're going to have, I guess, two or three years from now.
Prev 1 2 3 4 Next
[an error occurred while processing this directive]
