DIY Head Tracker For A Tenner
- Thread starter Dead Fred
- Start date
I don't believe you. The small print isn't fine enough. lol
Now, about that wireless head tracker...
Now, about that wireless head tracker...
Yep. If you're doing it on protoboard you can always put the button somewhere else if you want - it's only on the custom PCB that the button point is covered.
If you need another piece of protoboard I've got loads of the blasted things here so I can happily pop one in the post for you if you need
I'll post a write-up of how it went tomorrow, but at the moment I'll say I got it working with little difficulty, and very much appreciate the offer good sir. Your YouTube videos were invaluable and I thank you for such amazing.. well everything. Everyone involved with Edtracker I owe a great deal of thanks.
And again, the protoboard offer is very generous, thank you again.
Hi Brumster,
In the event that FD adds FPS to ED, do you have a plan to add translation to EDTracker?
EDIT:
I was just reading that the 9250 has an accelerometer. How come it isn't used for translation?
In the event that FD adds FPS to ED, do you have a plan to add translation to EDTracker?
EDIT:
I was just reading that the 9250 has an accelerometer. How come it isn't used for translation?
Last edited:
MEMS gyros (generically; I mean the gyro & accelerometer) cannot accurately track translational movements/position of your head, certainly not over any useful amount of time for this application. You need an external frame of reference.
Don't believe me? Look at the combined intellect of Oculus, HTC and Sony. They all rely on external cameras and point models/head recognition to track the position of a persons head in their VR headsets. Any belief that 3 blokes in a shed in the UK can somehow master accurate head positioning from a £5 MEMS gyro needs to end now ;-) !! We do have the facility on the EDTracker Pro board to run LEDs for point models but it's not something we are prioritising at the moment - it's a bit of future-proofing but current priority is wireless.
Don't believe me? Look at the combined intellect of Oculus, HTC and Sony. They all rely on external cameras and point models/head recognition to track the position of a persons head in their VR headsets. Any belief that 3 blokes in a shed in the UK can somehow master accurate head positioning from a £5 MEMS gyro needs to end now ;-) !! We do have the facility on the EDTracker Pro board to run LEDs for point models but it's not something we are prioritising at the moment - it's a bit of future-proofing but current priority is wireless.
Well, that blows. Ok then, throw an IR emitter on the module and three IR sensors on the monitor (or wherever it will fit) and calculate translation from the varying differences received from the sensor.
And is power consumption what is holding up the wireless version? I don't mind mounting a battery pack on my headset.
And is power consumption what is holding up the wireless version? I don't mind mounting a battery pack on my headset.
So you are using 9250? How's your build?
My one using 9250 will have to re-calibrate just after a few days, bcoz the Yaw axis will start jumping slightly after a few days, but the Pitch axis do not have any issue...
Also I find that if I unplug the tracker from the USB and reconnect to the same port, the UI cannot connect back to it (but okay in the game), unless I plug it into another USB port. But then if I unplug again, that USB will not work also, until I reboot my PC.
Sure, but then the costs ramp up massively. We need to develop a camera, plastic mouldings and construction/assembly of all the parts, plus the software development costs. I bet we wouldn't be far off TrackIR money at that stage.
More likely is you get a device with IR points on it and you go the whole Freetrack/Opentrack route yourself and provide your own camera/etc.
No. It has an integral lipo; battery life isn't a problem. It's wireless certification and some antenna tuning; all time and money, that's all!
Last edited:
Unfortunately, Opentrack doesn't allow you to specify two trackers. :-\
I had the brilliant idea of using the camera for translation and EDTracker for rotation, but alas, there is no option. I tried it on FacetrackNoir, but it only recognized the EDTracker input.
I had the brilliant idea of using the camera for translation and EDTracker for rotation, but alas, there is no option. I tried it on FacetrackNoir, but it only recognized the EDTracker input.
Finally ordered the custom PCB from EDTracker and DIY kit from Hobby Components a while ago. Still waiting for the PCB to come in, but can't wait to put it together!
Is there any advantage to the 9250 compared to the 9150? Or did they just replace it due to discontinuing the old one?
Is there any advantage to the 9250 compared to the 9150? Or did they just replace it due to discontinuing the old one?
Noob alert. I'm in the beginning stages of trying to put together a DIY tracker. I've sourced my parts, and I'm attempting to assemble the components on a protoboard using Bartybee's guide. I'm okay with soldering, not an expert. Mostly I've just done capacitors, resistors, headers, other through hole components. So I don't have any trouble soldering the headers to the board or components to the headers. What I am having trouble with is connecting all the small wires to the headers. I'm using some cat5 24 gage wire that does pass through the board, and just can't seem to get the hang of making a good connection from the wire to the header pin.
Does anyone have some tips / techniques on the best way to do this and still have a compact assembly? Any links to videos of the technique or assembly videos that followed Bartybee's wiring?
Also, I'm using the MPU-9250 with 10 pins, and I understand I can remove the tactile switch and re-center the device using a hotkey. Can someone point me to a wiring diagram that doesn't include the switch? I'm guessing it's just as simple as joining the 3 grounds on the Arduino with the ground and AD0 on the MPU. Then not connecting pin 10 on the Arduino to anything.
As a last resort I'll just order the custom PCB, but I'd like to give it another try with the protoboard and hopefully learn something new. Thanks in advance!
Does anyone have some tips / techniques on the best way to do this and still have a compact assembly? Any links to videos of the technique or assembly videos that followed Bartybee's wiring?
Also, I'm using the MPU-9250 with 10 pins, and I understand I can remove the tactile switch and re-center the device using a hotkey. Can someone point me to a wiring diagram that doesn't include the switch? I'm guessing it's just as simple as joining the 3 grounds on the Arduino with the ground and AD0 on the MPU. Then not connecting pin 10 on the Arduino to anything.
As a last resort I'll just order the custom PCB, but I'd like to give it another try with the protoboard and hopefully learn something new. Thanks in advance!
Hi Brumster,
I have an idea so fantastic and ludicrous that you might want to slap me twice and call me Shirley. But hear me out.
When you rotate your head around a point your eyes actually move through an arc and undergo a tiny bit of translation. In this, merely rotating your head can generate parallax and artificially simulate 3d. Would it be possible, or even worth it, to calculate this translation in EDTracker and inject it into the (x,y,z) axes to simulate parallax? Or is this idea so ludicrous it should be laid out in plaid?
And speaking of which, why doesn't ED have configurable axis for rotation (sans yaw and pitch) and translation?
I have an idea so fantastic and ludicrous that you might want to slap me twice and call me Shirley. But hear me out.
When you rotate your head around a point your eyes actually move through an arc and undergo a tiny bit of translation. In this, merely rotating your head can generate parallax and artificially simulate 3d. Would it be possible, or even worth it, to calculate this translation in EDTracker and inject it into the (x,y,z) axes to simulate parallax? Or is this idea so ludicrous it should be laid out in plaid?
And speaking of which, why doesn't ED have configurable axis for rotation (sans yaw and pitch) and translation?
Last edited:
I think you just answered your own question: ED only supports 2 axis for head tracking so 3D movement is pretty much impossible.
LOL
When your real head and eyes are doing it, there's nothing artificial or simulated
.Is it possible? Well, technically, yes although everyone's eyes and head sizes are different. Is it worth it though? I would say no.
If you essentially want to introduce some lateral movement with rotation, then there's nothing stopping you from doing this with something like Opentrack, and indeed I have done a video for Cliffs of Dover users on how they can use this technique to introduce some zoom and lateral movement and get a "pseudo-6DOF". But it will very much be up to the end user how much feels right and suits their environment (monitor size, distance, etc)
That you'd have to ask Frontier
but I suspect again, it's all about "is it worth it". You're on about roll I assume? You have to remember that when they started developing this game there were no products around that reported head movement as joystick axes, the only head tracking solution out there with 3DOF+ was TrackIR. So they embedded that, and included a simple 2-axes headlook for people with thumbsticks on their throttles (which was quite common back in the 90's in flight sims).Would be my guess
I doubt it'd be enough more likely going to need 230+ !! Lead-free solder takes a bit more heat than the older, leaded stuff (which we can't use these days, boooo )....
With a basic multimeter you can obviously check the VDD and GND lines to it but realistically to check the data and interrupt lines you'll need an oscilloscope, which is obviously beyond most people's repertoire of kit. But I think faulty 9250's boards are pretty rare to be honest, I would suggest 9 times out of 10 it was a build error (wiring, bad solder joint, etc) or otherwise just an incorrect firmware...
more likely going to need 230+ !! Lead-free solder takes a bit more heat than the older, leaded stuff (which we can't use these days, boooo )....With a basic multimeter you can obviously check the VDD and GND lines to it but realistically to check the data and interrupt lines you'll need an oscilloscope, which is obviously beyond most people's repertoire of kit. But I think faulty 9250's boards are pretty rare to be honest, I would suggest 9 times out of 10 it was a build error (wiring, bad solder joint, etc) or otherwise just an incorrect firmware...
Following up on Brumsters response, I think it's worth trying to map a bit of head roll into the mix. When you roll your head the on-screen view rolls already so no need to map that, but you could add a bit of sideways translation (via opentrack). That would mirror what happens in reality when you c-o-ck your head.
I'm not an opentrack user myself but if it's possible you could add yaw in as well (as show in Brum's Cliffs of Dover video) so that when you yaw past 90 degrees you get some additional lateral motion which would mimic the physical 'looking over the shoulder' motion.
Now I'm also wondering about throwing pitch in to add a hint of forwards/backwards motion
Just a smidge, and only at extremes to again mimic how the head translates when looking up / down.Rob
Finally got the customer PCB in the mail yesterday. Got right to it and put the thing together. It did take me quit some time to get the calibration right though, the yaw keeps drifting away and it seems that there are some spots I simply can't hit when calibrating the magnetometer.
Oh well, I'll have to put some more time in it I guess! It really adds a lot more fun to the game!
Oh well, I'll have to put some more time in it I guess! It really adds a lot more fun to the game!