kennyc260
New member
- Messages
- 2
- Reactions
- Points
- 3
hi this is my first post, i am here to report/confirm that PICO 4 works with Leap motion v1, wirelessly, alongside Virtual desktop.
i have a pico 4 global version, firmware 5.8,
native hand tracking deactivated,
high power OTG activated.
i have virtualhere USB server android 4.6 APK, on pico 4 installed,
and purshased 25$, purchase can be done through the windows VH client,
license received in Gmail, and pasted in VH client,
license for android is not free, and bound to pico 4 for life,
notes: i have tested numerous remote VirtualHubs nothing works aside from virtualhere,
usb network gate, flexihub, uSB/IP server android both versions, all of them cannot pick up the leap motion camera when plugin in pico 4 usbC/usbA with samsung otg adapter (since usb c can be flipped be aware that only one way will work)
virtualhere picks it up although not showing it the list pico4/server side, but it shows in the windows client as leap motion camera.
virtualhere server is free for linux/windows/mac as long as you use it to stream one device, therefore i have tested the linux version with a raspberi pi zero 2 w, it did work but the signal was so weak it only lasted one second and connection was dropped by the ultraleap control panel/service. and also did try with an android phone galaxy but the signal was also very weak, since antennas on phone are not designed for maximum troughput but for ability to pick up a wide range of wifi signals, so they are still useful for detecting surounding wifi interference, much better that usb wifi dongles.
at first i thought that the weak signal was caused by inadequate power received by the leap controller v1, so i tested various ways to feed more power through an amvr powerbank 3A/wired power adapter, to the raspberri pi and also directly to pico4, BUT nothing same weakness in the signal. it was maybe not related to power but to wifi signal, i was mislead by ultraleap message it seemed to me.
from there i went on doing some test in a windows/laptop to windows/desktop envirement with a bunch of usb wifi dongles, the best data link i could establish with my 5ghz old router was a 300mbits/s and the native laptop wifi interface was amongst the worst; 72mbps.
i picked on the hint an bought another router, a cheap huawei ax3 dual core, wifi 6 5ghz quad bands version. from router/desktop to router/laptop and for the first time it Worked.. i had a clue that i needed to establish a strong signal from Pico4 to router ax3, and i was able to establish an AX 1200mbps/s, on channel 36 and 40mhz/80mhz width, i had the luxury to achieve this since wifiinfoview32 reported vacancy on consecutive channels, no neighbor are using them, therefore avoiding intereference.
i had previously seen someone reporting that it works with a quest2/raspberri pi 4, but i highly doubt that this kind of wifi signal would work, no raspberi pi is capable of wifi 6, and would need two separate router one for pico and one for raspberri and also two 1gpbs LAN interfaces on desktop since vam through virtualdesktop would eat up all the bandwidth.
i further tried numerous usb diagnostic tools and techniques to try to understand how much bandwidth virtualhere would need to function properly, to my understanding and reports it and needs at least one third of bandwidth to stream, since it’’s a camera feed. it shows on windows resource monitor 60MB/s. not sure, but i put a tmeter throttle on VD around that.
i also test the pico real transfer rate using a local wifi speed test APK on pico, you need to install a client side on windows, disconnect internet, and give a go. WiFi - Internet Speed Test, zoltan pallagi.
from there i went on and tried netlimiter to throttle the virtual desktop since it’s very aggressive at eating bandwidth, but i settled for the free tmeter with a rule/filter for VD, lettiing the remaining bandwidth to virtualhere.
///////////////////////////////////////////
i also edited the ultraleap hand_tracker_config.json config to make it eat less bandwidth with the addition of these lines to the json file, but i am not sure if this is taken into account:
EDIT:
"allow_images": false,
ADDED:
"low_resource_mode_enabled": true,
"power_saving_adapter": true,
"robust_mode_enabled": false,
the robust mode needs testing, leave it at false for your first testings.
///////////////////////////////////
in vam i used the leap motion LEAPC.dll replacement in the vam folder , not the github technique with the replacement in steamvr folder, deleted those. since they are interefering with vam, steamvr console should not detect leapmotion controllers upon vam launching, the vam version 1.22 but i have a creator version.
i checked leap motion in VR settings, and pushed the UI button to activate leap motion.
and voila,
i can have sessions up to 15 minutes with passthrough and without leap motion hands dropping into the floor-less voids, i am working on alexa / cortana setup to make a routine for rebooting leap motion service without having to do it mannualy.
////////////////////////////////////////////////////////
requirements:
pico 4 / usb otg / rigid usb micro-b cable / tmeter rule filter 60MB limit / json edit / huawei ax3 800-1200mbits connection / VD low res / VD automatic codec / vam leapc.dll replacement / virualhere usb server purchase andoird / VH usb server apk / ultraleap gemini 5.16 service.
others: WiFi - Internet Speed Test, zoltan pallagi / wifiinfoview32 / resource monitor windows / win 10 / vam / velcro stick bands to attach leap to headset.
setup: nvidia 1650 / ryzen 5 / huawei AX3. (the GC is obviously a bottleneck but perfect for this test).
next setup: better GC/better router/leap motion 2.
////////////////////////////////////////////
i am sorry for the way i have written this could be confusing, also very thankful to all the esteemed members of this community, i felt like this needed to be reported since the informations are so scarce.
my next interest is FBT i bought steamvr standables (an fbt estimation trackers), but i am still a noob i coulnd’t figure out yet how the 8 trackers are visible in vam but not useable, i also bought miscreated plugin to activate the balance/plugin along with possess mode. i am still trying to make it work. if not i would go the route of slimevr, i have a kinect360 but that is for mocap as i understand ill give it a try.
i have a pico 4 global version, firmware 5.8,
native hand tracking deactivated,
high power OTG activated.
i have virtualhere USB server android 4.6 APK, on pico 4 installed,
and purshased 25$, purchase can be done through the windows VH client,
license received in Gmail, and pasted in VH client,
license for android is not free, and bound to pico 4 for life,
notes: i have tested numerous remote VirtualHubs nothing works aside from virtualhere,
usb network gate, flexihub, uSB/IP server android both versions, all of them cannot pick up the leap motion camera when plugin in pico 4 usbC/usbA with samsung otg adapter (since usb c can be flipped be aware that only one way will work)
virtualhere picks it up although not showing it the list pico4/server side, but it shows in the windows client as leap motion camera.
virtualhere server is free for linux/windows/mac as long as you use it to stream one device, therefore i have tested the linux version with a raspberi pi zero 2 w, it did work but the signal was so weak it only lasted one second and connection was dropped by the ultraleap control panel/service. and also did try with an android phone galaxy but the signal was also very weak, since antennas on phone are not designed for maximum troughput but for ability to pick up a wide range of wifi signals, so they are still useful for detecting surounding wifi interference, much better that usb wifi dongles.
at first i thought that the weak signal was caused by inadequate power received by the leap controller v1, so i tested various ways to feed more power through an amvr powerbank 3A/wired power adapter, to the raspberri pi and also directly to pico4, BUT nothing same weakness in the signal. it was maybe not related to power but to wifi signal, i was mislead by ultraleap message it seemed to me.
from there i went on doing some test in a windows/laptop to windows/desktop envirement with a bunch of usb wifi dongles, the best data link i could establish with my 5ghz old router was a 300mbits/s and the native laptop wifi interface was amongst the worst; 72mbps.
i picked on the hint an bought another router, a cheap huawei ax3 dual core, wifi 6 5ghz quad bands version. from router/desktop to router/laptop and for the first time it Worked.. i had a clue that i needed to establish a strong signal from Pico4 to router ax3, and i was able to establish an AX 1200mbps/s, on channel 36 and 40mhz/80mhz width, i had the luxury to achieve this since wifiinfoview32 reported vacancy on consecutive channels, no neighbor are using them, therefore avoiding intereference.
i had previously seen someone reporting that it works with a quest2/raspberri pi 4, but i highly doubt that this kind of wifi signal would work, no raspberi pi is capable of wifi 6, and would need two separate router one for pico and one for raspberri and also two 1gpbs LAN interfaces on desktop since vam through virtualdesktop would eat up all the bandwidth.
i further tried numerous usb diagnostic tools and techniques to try to understand how much bandwidth virtualhere would need to function properly, to my understanding and reports it and needs at least one third of bandwidth to stream, since it’’s a camera feed. it shows on windows resource monitor 60MB/s. not sure, but i put a tmeter throttle on VD around that.
i also test the pico real transfer rate using a local wifi speed test APK on pico, you need to install a client side on windows, disconnect internet, and give a go. WiFi - Internet Speed Test, zoltan pallagi.
from there i went on and tried netlimiter to throttle the virtual desktop since it’s very aggressive at eating bandwidth, but i settled for the free tmeter with a rule/filter for VD, lettiing the remaining bandwidth to virtualhere.
///////////////////////////////////////////
i also edited the ultraleap hand_tracker_config.json config to make it eat less bandwidth with the addition of these lines to the json file, but i am not sure if this is taken into account:
EDIT:
"allow_images": false,
ADDED:
"low_resource_mode_enabled": true,
"power_saving_adapter": true,
"robust_mode_enabled": false,
the robust mode needs testing, leave it at false for your first testings.
///////////////////////////////////
in vam i used the leap motion LEAPC.dll replacement in the vam folder , not the github technique with the replacement in steamvr folder, deleted those. since they are interefering with vam, steamvr console should not detect leapmotion controllers upon vam launching, the vam version 1.22 but i have a creator version.
i checked leap motion in VR settings, and pushed the UI button to activate leap motion.
and voila,
i can have sessions up to 15 minutes with passthrough and without leap motion hands dropping into the floor-less voids, i am working on alexa / cortana setup to make a routine for rebooting leap motion service without having to do it mannualy.
////////////////////////////////////////////////////////
requirements:
pico 4 / usb otg / rigid usb micro-b cable / tmeter rule filter 60MB limit / json edit / huawei ax3 800-1200mbits connection / VD low res / VD automatic codec / vam leapc.dll replacement / virualhere usb server purchase andoird / VH usb server apk / ultraleap gemini 5.16 service.
others: WiFi - Internet Speed Test, zoltan pallagi / wifiinfoview32 / resource monitor windows / win 10 / vam / velcro stick bands to attach leap to headset.
setup: nvidia 1650 / ryzen 5 / huawei AX3. (the GC is obviously a bottleneck but perfect for this test).
next setup: better GC/better router/leap motion 2.
////////////////////////////////////////////
i am sorry for the way i have written this could be confusing, also very thankful to all the esteemed members of this community, i felt like this needed to be reported since the informations are so scarce.
my next interest is FBT i bought steamvr standables (an fbt estimation trackers), but i am still a noob i coulnd’t figure out yet how the 8 trackers are visible in vam but not useable, i also bought miscreated plugin to activate the balance/plugin along with possess mode. i am still trying to make it work. if not i would go the route of slimevr, i have a kinect360 but that is for mocap as i understand ill give it a try.
Attachments
Last edited:
