kaizoku said:
....You really love this shit don't you?
Why did you phrase your post as if you are disagreeing with me before embarking on some huge bout of speculation? We can all speculate on how it works, but none of us actually knows unless we're game developers for Rev.
.
Well you are right in that it can work like a laser pointer, but it doesn't have to; that is what I was trying to get across. Other people seem to be confused as to whether or not it works like a light gun, once again the answer is yes and no.
The key idea is that as far as software is concerned, the revmote is a bundle of position and orientation data, that works however you program it to. For a light gun, the software is sent information about what color was displayed on the screen (revmote does not do this)
kaizoku said:
To entertain you, I'm just trying to think of how it will work, it really is baffling how they could get it to act as a pointer without calibration. I'm working with the idea that I can sit on my couch and hit a target in the middle of the screen, then decide to sit on the floor closer to the TV and still be able to simply aim at the target and hit it, much like happens in the demos.
Obvously you now have a different set of angles and distances on various axis. I don;t like to think of it as a 3D mouse although it can be programmed to be one. If it was a 3D mouse, when I stand up, it would think I wanted to raise my controller and move accordingly. I would then have to center it again?
On the other hand if its programmed as a pointer I can aim at the screen and stand up and it would still be aiming in the same place, despite different controller positions i.e. it wont interpret me standing up as me aiming up.
I'm confused by this wording. Basically if you stand up, the position of the controller will translate up an amount equivalent to your height. When you stand if you continue holding the remote facing straight forward, you will no longer be aiming at the center of the screen. If you tilt the remote downward while standing, you can continue aiming at the center of the screen, just like a laser pointer. This is all assuming the software is programmed to trace a vector in front of the remote and identify your target as the intersection with the screen.
The mouse analogy was a bad one... I didn't mean the revolution remote is like a 3D mouse, I was trying to describe how it sends data to the software. I don't have a revmote, but I'm basing this assumption on how other virtual reality software, such as vrjuggler, interprets the data from similar 6 DOF devices:
http://www.vrjuggler.org/
Sorry if I am not coming across right, I am just trying to shed some light on how the remote works based on my experience with other similar devices. It seems a lot of people are confused about what it is capable of.
As far as calibration is concerned, a graphical representation such as a cursor could eliminate the need for calibration, associating a predefined amount of arm motion with the edges of the screen (like a mouse). But how would this account for people of different sizes and arm lengths? Angle of movement? But what if the person starts in the wrong position?
Perhaps this is the situation, but using this tracking system, other games could require calibration and allow for better control. As I stated, I would love to see Nintendo come up with a system that doesn't need calibration, and if you have ideas for how this can be accomplished I'd like to hear them as well.
antipode said:
I think claiming that the Revmote can function like a light gun and let the user accurately point at pixels on the screen without any crosshairs or feedback is strectching it. Pkunk - you admit above that it's difficult to figure out how they can obtain precise depth, but wouldn't you need really, really precise depth to extrapolate out to a point on the screen? How else would you do it? I mean, yikes, just stand in front of your TV screen with your thumb out and shrug your shoulder back enough to be pointing at a completely different part of the screen... I find it hard to believe that in a consumer device you can tell from the amplitude of an infrared or RF signal absolute depth to centimeters. Besides that, you have the "no calibration" comment, the "TV size doesn't matter" comment, and the "television can be at an angle relative to the user just fine" comment. .
These are good points and this is what I want to know as well, also why I want to see a picture of the sensor bar, and probably why Nintendo hasnt shown it just yet. What I was referring to when I mentioned using the light receptors to find depth, was something like the P5 glove does. This works well for depth, even though it isnt quite the same as triangulation:
http://www.mts.net/~kbagnall/p5/p5 dissassembly.html
Orientation for the revmote is obviously done differently than the glove, but Im thinking position will be similar. Howevr, there could be other ways.
antipode said:
I think when it comes to aiming, Nintendo designed it to be used relative to the user, irregardless of the screen. If you raise your arm 45 degrees it'll map to the same position on any screen given the same sensor arrangement
The thing is, the same device could be programmed to do both this and actual calibrated position tracking with different software packages. This method would require a visual cue to move around the screen so the user understands how a 45 degree arm tilt corresponds to the size of their TV screen (this is truly like a mouse). But how does this method know where the revmote starts in relation to the user? Did they have their arm at a 45 degree angle already when the power turned on? Was the remote on the table? This is why I believe some form of non-invasive calibration is going on, but I have no idea to what extent, if any.
Probably wouldn't be that easy to obtain, though.
