Ringu wrote: As for older controllers, I meant PS4 controllers made before the PSVR was in development, which wouldn't have had IR LEDs added to them with no actual reason, and never offered the kind of tracking that PSVR uses for controllers (relying instead only on the internal controller gyros).

after checking release dates its also very possible that the controllers had IR position leds in them, as they were made after playstation move and have similar movement tracking abilities.
and a single additiona LED in the already existing lightbar is no huge feat.

Cornflakes_91 wrote:

Ringu wrote: As for older controllers, I meant PS4 controllers made before the PSVR was in development, which wouldn't have had IR LEDs added to them with no actual reason, and never offered the kind of tracking that PSVR uses for controllers (relying instead only on the internal controller gyros).

after checking release dates its also very possible that the controllers had IR position leds in them, as they were made after playstation move and have similar movement tracking abilities.
and a single additiona LED in the already existing lightbar is no huge feat.

You're being very silly, stop it.
There is zero purpose in having a visible light AND an IR LED if you're not going to use the visible light; it's going to get in the way, be an additional expense and power drain, and require different hardware design to allow it to be visible and consistent etc.

On the other hand, having ONLY a visible light - especially a circle such as the PS Move controllers - is easy to spot with a simple camera and some computation to track. It's not perfect, but it's reasonable.

You're already having to jump through hoops to explain how an IR LED could be made to work, and the visible light just puts obstacles in your way - not to mention that the whole point of IR is that it's Infra-*RED*, so confusing that with a blue coloured object would require one weird-ass filter that would be extremely noticeable on the camera output.

It's not tracking with an IR LED, it's tracking with visible light sources.

Dunno about the rest, but

Ringu wrote: not to mention that the whole point of IR is that it's Infra-*RED*, so confusing that with a blue coloured object would require one weird-ass filter that would be extremely noticeable on the camera output.

No, actually, camera's have tiny filters per pixel - red, green and blue. However, they also receive IR. Which colour the IR appears as purely depends on the transmittance of those filters in the IR band, which depends on the camera and can perfectly well be blue or purple. (after all, they're designed to block the visual spectrum except for 1 colour, not the rest of the spectrum)

I mean, how can you say it's silly when you can literally see it on Corn's photo?

Ringu wrote: You're being very silly, stop it.
There is zero purpose in having a visible light AND an IR LED if you're not going to use the visible light; it's going to get in the way, be an additional expense and power drain, and require different hardware design to allow it to be visible and consistent etc.

DUALSHOCK®4 features a light bar on the top with three color LEDs that illuminate in various colors. The light bar illuminates to match the color of characters in a game to offer a simpler, more friendly way to identify players, even when playing side by side.

http://www.sie.com/en/corporate/release ... 0221b.html

the visible lights dont have to have a technical purpose.
there have been stupider things mounted in devices which served less purpose.

Ringu wrote: You're already having to jump through hoops to explain how an IR LED could be made to work, and the visible light just puts obstacles in your way - not to mention that the whole point of IR is that it's Infra-*RED*, so confusing that with a blue coloured object would require one weird-ass filter that would be extremely noticeable on the camera output.

define "noticeable"? the filtered throughput the user sees doesnt have to be the output of the filter the tracking system used

also: look at the stupid photo i took
it appears purple, and i tried with another camera (the one on my brother's phone) and it appeared white.
i dont know what it appears in the move camera, could very well be that that interprets it as blue.
i have no idea what the specific color filters (the ones that make the camera RGB capable in the first place) do with IR light

Dinosawer wrote: I mean, how can you say it's silly when you can literally see it on Corn's photo?

which incidentally was taken by a sony-made camera so will have similar characteristics with a significant likelyhood.

Cornflakes_91 wrote:http://www.sie.com/en/corporate/release ... 0221b.html

the visible lights dont have to have a technical purpose.
there have been stupider things mounted in devices which served less purpose.

From that page you link: "PlayStation®4 Eye senses the color of DUALSHOCK®4’s light bar to judge the positions of multiple players"

From the PS Move wiki page: "Based on the colors in the user environment captured by the camera, the system dynamically selects an orb color that can be distinguished from the rest of the scene. The colored light serves as an active marker, the position of which can be tracked along the image plane by the camera.[11][12] The uniform spherical shape and known size of the light also allows the system to simply determine the controller's distance from the camera through the light's image size, thus enabling the controller's position to be tracked in three dimensions"

Wow, this thread exploded while I was sleeping.

When I was playing with my home-made VR solution, the head-tracking was one of the reasons I initially shelved it. One of the things that I played with was FreeTrack which allows you to make your own head-tracking via different lights. One way they recommended was to use a PS3 Eye as your camera source. First gen ones used filters so that it could only see IR lights (like Dino said). Second gen ones did not. This is also a $4 camera used at this point. The instructions they provided were on how to remove the filter because it was later realized that there was no reason to use just IR since you could tweak the software so that it filtered out all but the 'brightest' sources, which were IR (since IR overloads most standard sensors; hence why it appears 'white' most of the time). Also, the PS3 Eye camera was chosen initially because it supported 120fps.

This means that the most 'reliable' home-grown methods via FreeTrack didn't really care about what the light source was, just that it was easiest to identify/brightest to be.

Also, from my experiences with diving deep in FreeTrack, it doesn't matter if it's visible or IR light. It's more about the point source. The LEDs used in the PS4 devices (DS4 and move controllers) are a 'big' source, whereas the Vive and Oculus are much more pointed sources. This was explained in FreeTrack that you can't just tape an LED on unless it's a point source.

That would mean that if the light source tilts even a little bit away from the camera, it would disappear since it is a pointed light. This is also why the Vive and Oculus have so many lights.

Did a little bit of looking, and it looks like the Oculus and Vive both have gyros, magetometer, and accelerometers. That would mean that it matters more on software than what's actually in the device.

DWMagus wrote:The LEDs used in the PS4 devices (DS4 and move controllers) are a 'big' source, whereas the Vive and Oculus are much more pointed sources. This was explained in FreeTrack that you can't just tape an LED on unless it's a point source.

That would mean that if the light source tilts even a little bit away from the camera, it would disappear since it is a pointed light. This is also why the Vive and Oculus have so many lights.

Did a little bit of looking, and it looks like the Oculus and Vive both have gyros, magetometer, and accelerometers. That would mean that it matters more on software than what's actually in the device.

As far as I can tell, the Vive doesn't have lights at all: the boxes at the corners are IR emitters that have an array of IR LEDs each on a slightly different channel, and the HMD (and the controllers and the new Track widgets) seem to read IR signal strength or something -- at least, that's how I'm figuring it since the boxes have no connection to any compute source, and you can't put cameras in the controllers cos that would just be silly.

In my head, I could see how two diametrically-opposite lighthouse boxes on different channels, each with an array of IR LEDs, could be used to triangulate and orientate with just a couple of sensors in the device and some really simple maths; and that would certainly explain the accuracy too since signal strength is consistent and analogue, so you can have as much precision as you want. Using gyros, magnetometers and accelerometers on top of that would certainly explain it, to me anyway.

Ringu wrote: As far as I can tell, the Vive doesn't have lights at all: the boxes at the corners are IR emitters that have an array of IR LEDs each on a slightly different channel, and the HMD (and the controllers and the new Track widgets) seem to read IR signal strength or something -- at least, that's how I'm figuring it since the boxes have no connection to any compute source,

yeah, something along those lines according to wikipedia. some time-of-flight / phase difference measurements from the IR pulser boxes.

Ringu wrote: and you can't put cameras in the controllers cos that would just be silly.

thats how the wii controllers work.
the stationary part is simply 4 IR leds.

Ringu wrote: In my head, I could see how two diametrically-opposite lighthouse boxes on different channels, each with an array of IR LEDs, could be used to triangulate and orientate with just a couple of sensors in the device and some really simple maths; and that would certainly explain the accuracy too since signal strength is consistent and analogue, so you can have as much precision as you want. Using gyros, magnetometers and accelerometers on top of that would certainly explain it, to me anyway.

the description i read was that its cyclic pulsing and that they can take orientation out of the different arrival times ("it arrived at corner 2 half a microsecond earlier than at corner 3 thus its so and so far more left)

its not an intensity measurement as far as i gathered.
which would also be pretty meh because you dont know the positions of the pulsers in the room :V

Cornflakes_91 wrote:yeah, something along those lines according to wikipedia. some time-of-flight / phase difference measurements from the IR pulser boxes.

Makes sense. Easy and quick to calculate.

Ringu wrote: and you can't put cameras in the controllers cos that would just be silly.

thats how the wii controllers work.
the stationary part is simply 4 IR leds.

By 'cameras' I meant something that captures a frame, optically - the wii controllers don't have cameras per se, they are mostly gyro/accelerometer-based for the motion detection, with an IR optical sensor for the hand pointer function (as you say, reading the IR LEDs in the 'sensor bar')

the description i read was that its cyclic pulsing and that they can take orientation out of the different arrival times ("it arrived at corner 2 half a microsecond earlier than at corner 3 thus its so and so far more left)

its not an intensity measurement as far as i gathered.
which would also be pretty meh because you dont know the positions of the pulsers in the room :V

Sure, the flight time is the same thing as intensity in terms of the positional calculations.
Also, the initial calibration of the Vive essentially records the initial sensor data and that becomes "centre" from then on, so it doesn't need to care about the actual position of the lighthouses, just the change in data values.

Also also, it is possible to calibrate it so that in Standing or in Room-scale mode, you trace the boundaries of your free movement area and it records these extreme values to ensure it doesn't force you out of them.

Ringu wrote: By 'cameras' I meant something that captures a frame, optically - the wii controllers don't have cameras per se, they are mostly gyro/accelerometer-based for the motion detection, with an IR optical sensor for the hand pointer function (as you say, reading the IR LEDs in the 'sensor bar')

well, the IR sensor in the wii does capture a frame. it just doesnt return it but has a bunch of internal processing and filtering that only returns bright IR points in its FoV and the captured frames dont get out of that monolithic module.

Stumbled across little tidbit in regards to the PSVR;

...the "drift" happens it's because the calibration isn't properly coded. When connected to PC they can see the x,y,z parameters, the chip calibrates it fine however the processor unit seems to make a synchronising mistake causing the two to contradict = drift.

This information is coming from PS4 owners...which means the PSVR does have drift, but ymmv. I may just not have used it long enough to have an issue.

But it also means that we desperately need some sort of firmware update so that we can start reverse engineering it. From what I gather than the processing unit is doing, with a little bit of trickery, it would be possible to perform the same things that SteamVR is doing in that unit, meaning all you'd need is a barrel-distorted sbs image and the rest would be just fine.

I may have to start sniffing some data on it. This is getting me a bit more interested than anticipated.

Shout if you want anything I can provide, DWMagus, always happy to run tests or diagnostics or whatever.

Ringu wrote:Shout if you want anything I can provide, DWMagus, always happy to run tests or diagnostics or whatever.

Actually, you could hopefully answer something for me.

I keep seeing references to a Day 1 patch on the PSVR. Since I don't have a PS4, I'm guessing I have no easy way of checking this myself. Was there a day 1 patch, or any firmware update for the PSVR you needed to do?

DWMagus wrote:

Ringu wrote:Shout if you want anything I can provide, DWMagus, always happy to run tests or diagnostics or whatever.

Actually, you could hopefully answer something for me.

I keep seeing references to a Day 1 patch on the PSVR. Since I don't have a PS4, I'm guessing I have no easy way of checking this myself. Was there a day 1 patch, or any firmware update for the PSVR you needed to do?

I think there was: I had difficulty getting it to work initially and I was patching a lot of things so I don't recall exactly everything that I had to do, but I seem to remember one for the PSVR, one for the camera, and I had a system OS update and a few games updates to go through as well...
I'll try to find the current version number tonight.