Could you please talk about your involvement with the development of the PlayStation Move? What was your contribution exactly?
Anton Mikhailov: My particular role was to track the sphere of the Move to 0.1 pixel accuracy. That was actually my main goal for about a year through the prototyping phase. Rick Marks wrote a pretty robust sphere tracking algorithm back in the EyeToy days for PS2 and we used to that jump start Move development. He felt that we can squeeze much more accuracy out of it though so I took on that role. Eric Larsen, who wrote some core libraries for PS3, joined the project to work on the sensor fusion. His role was to combine my output with the gyros and accelerometers to get the best results. That was the high level, there was a bunch of side stuff that needed to be done as well though. I worked on the sphere calibration, for example, and also lived in UK for about 3 months working with developers. Eric worked on simulating different effects and testing and Rick wrote the PS3 driver and a lot of image processing code. Lots of late nights between the 3 of us... then of course there was the E3 demo crunch when we spread the late night disease to a bunch of others helping us. Right around then we realized we needed someone on demos full time so Kenny Hoff joined the team. That's basically our team now.

Click to expand...

It might seem odd to require the PS3 to be in the loop but actually it has some advantages. For one, you don’t pay any resources on your host PC to track the Move. While the PS3 can do this quickly, many home PC’s and laptops don’t have enough processing power to do this at 60 frames per second much alone run an application along side of it. Second, we transfer the data over a standard network protocol, so any network device can make use of it. We generally refer to PC’s because that’s what people are familiar with but there are much more devices that support basic networking. Devices like phones, research workstations, or just Macs or PC’s with uncommon operating systems all can talk through networks so it works out great. So if we focused strictly on PC, we’d lose this flexibility since writing software for all those platforms would be hugely time consuming… maybe not even possible because of their processing power.

Click to expand...

Let’s now get to the point many iWaggle3D readers are mostly curious about: the pointer “drifting issue”. As I have exposed in one of my early tech analysis, no game using the PlayStation Move as a pointing device manages to keep the on-screen cursor perfectly aligned with it, either because it drifts away from its original position over time or because it just doesn’t show up exactly where you expect it to after calibrating. Could you please offer an explanation as to why this happens and tell us if it’s something that can be fixed in the future?

Anton Mikhailov: The Move doesn't behave like a light gun. Its not easy to make a light gun these days, unfortunately, because of today's TV technology. The way the old light guns were made relied on a very clever hack that only works on CRT TV's. Nobody has those these days, so its actually quite tricky to get something that accurate. The Move is technically the only controller that has come close, actually. What you need for a light gun is a precise position and rotation relative to the TV, as well as the TV size. Usually you assume the TV is by the camera or whatever tracker you have.

So the Move has both of those, but its very hard to get the accuracy needed to do an accurate shot from 8ft away. With Move we usually have about 1-3 degrees accuracy, which is quite an achievement, I think, considering how fast and unpredictably you can swing it. But that kind of deviation is amplified by the distance to the TV, so it can translate to an inch or two on screen, especially if you're far away. You can imagine trying to touch the screen with a 8ft stick, any motion you have in your wrist will move the stick quite a bit. So even though the Move is very accurate at tracking your wrist we're basically fighting a really tough battle. So flying the cursor on the screen works great with Move, but pointing is not always perfect. The closer you are to the TV though, the better it should work, both because the tracking improves and because the amplification is not as bad.

Bottom line though is that we're aware that its not ideal... we're keeping it on our list of things to improve in the future.

Click to expand...

Of course one of the advantages of the PlayStation Move technology against competitive ones like Kinect is the relatively “undisruptive” nature of its implementation as an optional control solution alongside a traditional controller. As a developer, you don’t need to severely rethink your DS3-based game in order to support it. This has allowed the PlayStation Move to find its way into numerous third party productions (there are about 50 games supporting it at this time). That said, I get the impression that very few developers besides those founded by Sony itself are actually interested in building Move “required” games. On the other hand, with Kinect, every game is “special” because there is no other option but to actually build something unique around it. What are your thoughts on this? Is the flexibility of the Move technology a double edge sword?

Anton Mikhailov: It’s always up to first party to push new hardware and show it off. For a third party, making a game for a specific piece of hardware is risky. If a developer has an idea for the Kinect though, they really have no other choice. In some cases, they could actually be cross platform to the PlayStation Eye but we haven’t seen that yet. For Move, shooter games can be cross platform with the DualShock, as can some gesture games. That's a little safer so I think that’s why we see more of that on the Move side. That doesn’t really tell you anything about the quality of the experience though. A game might play great with Move and not so great with the DualShock, or vice versa.

At the end of the day I think it’s best to look at the game that you want to play and how you want to play it. Our goal is to deliver the best interface and it’s the developer’s goal to make the best experience around it. This way the consumer can pick how they want to interact with the game. Sometimes there is only one choice, other times there are a few, just pick what you’re excited about and what you think would be fun (and if you haven’t, try Move because you might like it [smiles])

Click to expand...

User Name Here said:

"Bottom line though is that we're aware that its not ideal... we're keeping it on our list of things to improve in the future."

I wonder what he means by this? The future being a future (next gen) iteration of the Move, or an update to the current Move hardware?

Click to expand...

User Name Here said:

"Bottom line though is that we're aware that its not ideal... we're keeping it on our list of things to improve in the future."

I wonder what he means by this? The future being a future (next gen) iteration of the Move, or an update to the current Move hardware?

Click to expand...

User Name Here said:

"Bottom line though is that we're aware that its not ideal... we're keeping it on our list of things to improve in the future."

I wonder what he means by this? The future being a future (next gen) iteration of the Move, or an update to the current Move hardware?

Click to expand...

TTP said:

Next gen most likely. The way he describes the reasons behind the misalignment ("With Move we usually have about 1-3 degrees accuracy") leads me to believe this is something we have to deal with.

As for the drifting issues (not the slight misalignment Anton actually refers to - that's "normal"), I guess that's something that can be sorted out this gen. It sure isn't normal some people are experiencing it (cursor moving around the screen on its own etc) while others are not.

Anton was uneasy answering those questions as he was not directly involved with that stuff (he worked on sphere tracking and light calibration basically) so I didn't insist

Click to expand...