Skip to content

Index Knuckles

"The best we've got, but not good enough."

Drafted 2023-10-31 | Posted 2023-11-01 | Last edited 2023-11-01

Background and Context

Earlier this year, I purchased some Valve Index Controllers (often nicknamed Knuckles). Actually had to check my Valve Purchase history to be sure as to when I bought them, it was the fourth of June, 2023. While taking the time to look back here, I also purchased my first base stations on the sixth of May, 2023 -- at the time, I was daily-driving an HP Reverb G2 (v2) with WMR controllers and was wanting to get into Full Body Tracking (FBT), so I purchased some Tundra Trackers. Essentially, I was in the Base Station ecosystem for about a month before getting my hands on Knuckles.

For context, I was coming from an Oculus Rift S before the HP Reverb G2. The controllers on the Rift S were great, fine for most of my use-cases and overall rather pleasant to use as far as controllers go (perhaps a little creaky on the spring for the trigger and grip, but nothing functionally wrong). However, when I switched to the Reverb G2 and its appropriate controllers (as the Rift S controllers were incompatible with the G2), my experience overall degraded very quickly based on many factors:

  • The overall battery life, despite having twice the amount of AA batteries in the enclosure, was significantly worse than that of the Rift S.
  • The build quality of the plastics felt very flimsy, cheap, and rattley (for the lack of a better word).
  • It felt significantly heavier than Rift S controllers.
  • The area in which the controllers were trackable was less than my field of view. (In other words, if I put my hand in a certain area of my view, the controller would stop being tracked due to the cameras losing consistent sight of the tracking ring and light constellation being emitted.)
  • As a VRChat player, the lack of capacitive buttons and thumbsticks made expressions very difficult to consistently control

Essentially, from the time I first got my G2 and full body tracking, I decided this wasn't nearly enough. I wanted something else. Obviously no other controllers would really work with my current setup unless I went with light house tracked controllers... so I decided to go with the best out there for PCVR: Valve Index Controllers (Knuckles).

The overall experience from receiving and setting up the controllers was nothing short of great -- Valve has an impressive end-user experince when it comes to unboxing hardware, as I've found with the base stations a month prior. Even the shipping box having call-outs to the Portal Companion Cube was something I genuinely found cute -- I really appreciated the little details they put into the unboxing experience. But the unboxing only happens once.

Pairing the controllers took me a bit of a process as I was already on a G2 (which didn't have integrated Watchman dongles like the Index). I purchased the Tundra Tracking 4 pack, which provided me with an SW4 dongle. However, using three of the four possible connections for FBT meant that there was only room for one of the two controllers to be used at any given point in time. That was, until I found the firmware on my Steam Controller dongle could be flashed to work on the Watchman protocol. Following some guides helped me figure out how to do this and it was rather easy from here to pair my second controller.

At this point, I managed to pair two Knuckles, three Tundra Trackers, and get the G2 working within SteamVR -- I had to calibrate to make sure everything was perfect (it wasn't perfect... but my rants for the G2 may wait for a future post). Once everything was where it should be and working nicely, I tested that the fingers were working as expected in the SteamVR Home as well as Valve's The Lab experience. Overall, an immediate thing I notiecd was the build quality and consistency of materials being used was generally higher, inspiring a lot more confidence in the construction and care put into the design overall. I also noticed the sensitivity of the finger tracking leading to much more immersive experiences. However, these experiences weren't enough. I wanted to get the whole "shabang".

I jumped into VRChat and did as any VRChat player would -- stood in front of a mirror. I just took some time to just look at my hands, look at my avatar's hands, and try to see how these related to my hands. It clicked. It felt great. It felt natural. It felt intuitive. It was amazing, I loved it!

But I noticed just a few things straight from the first time using them. Particularly the calibration on the fingers didn't seem completely accurate to my fingers. Not too bad of a problem, just recalibrate by drumming my fingers on the grips (I learned this trick a little later; a very clever means of on-the-fly calibration, hats off to you, Valve). But then I noticed that the fingers only worked for my index, middle, ring, and pinky fingers. The thumb was subjected to simply the capacitive "Stick", "Touchpad", "B", or "A" buttons for different thumb positional tracking. However, last I recall, thumbs can go over the individual fingers to form a fist, instead of simply resting on the outside as if it's holding a controller -- and thumbs have a much larger range of motion than all other fingers, yet it's got the least available on this controller style... Perhaps I was missing something though! I kept using it because overall, it's a great package.

Issues

Over time though, some more quirks were revealed to me. Particularly, when I set aside a little bit of money and purchased myself a Bigscreen Beyond (on the twelfth of July, 2023), I was able to interact with many others of the like mind who decided to also pre-order the Beyond headset. Such community members had experiences varying from similar to me (having only used VR consistently for about a year to 14 months at that point) to those who would be considered VR Veterans, having pre-ordered the first Rift Development Kit, working hand-in-hand with some of the earliest creators of VR in general. They've run the VR gauntlet and come back home all before breakfast, as it were... so having the ability to interact as one of them in a community of "We all shelled out a paycheck for a headset" was rather valuable in getting insight as to various things with VR in general as well as certain experiences in specific.

Valve Knuckles were a very specific topic that has been, and (as of writing) continues to be a topic to talk about in this particular community of PCVR enthusiasts. From much of my interactions and lurking, I found some of the most annoying things on Knuckles were the following:

  • Quality Assurance (QA)
    • Overall, the Knuckles are fine. However, there have been some who've expressed much disappointment in ensuring the product worked out of the box without RMA.
  • Long Term Reliability
    • "Stick Drift" being one of the biggest complaints I've seen overall. At the time of the Index release (2019), hall effect joysticks hadn't yet been wide-spread and easily attainable through a supply chain. Since its first release, the Knuckles haven't seemed to change significantly, if at all, and so the sticks continue to use potentiometer-based angle sensing
  • Durability
    • Essentially, the idea of "if I hit something because I have a screen on my face, I should expect that the thing in my hand shouldn't break instantly". Fortunately, I personally never had any experience of this sort of thing happening to me, but I could definitely see a few points in the construction where a non-trivial amount of mechanical stress may affect the structural integrity and cause a few things to snap.
  • Size
    • "Knuckles are too big for my hands". While ergonomics is a very difficult challenge to overcome for all customers, I can at least respect Valve for trying to accommodate a wide variety of audiences by making these controllers too large rather than too small (and potentially unuseable for big-hand havers).
  • Gesture Control
    • While Knuckles provide a wide range of gesture control with individually poseable fingers, in games such as VRChat where each finger gesture may change an expression on the avatar's face, or trigger certain actions, having consistency in these actions is a necessity.
    • Unfortunately, due to the on-the-fly calibration working the way it is, sometimes the algorithms may be thrown off and, on occasion, a ring-finger on the left hand would pop out randomly. Or perhaps the proximity sensor under the trigger enclosure may think there's an object close-by (when there really isn't), and the index finger in-game would be half bent.
    • Overall this has been a very large pain point for me personally as I not only rely on gesture control for accurate VRChat face expression puppeteering, but also phantom sense.

So in essence, while these issues are much easier to live with than, say, the ones detailed above in my segment of complaints about the WMR controller provided with the G2, they are issues regardless. So what alternatives do we have compared to the Knuckles?

Really, if you're in the Light House ecosystem, that's about it. It's either Knuckles, Vive Controllers (nicknamed Wands), or Pimax Swords -- and neither of the other two have joysticks or proper button control like the Knuckles. Of course, Etee controllers, at this current point of writing, have been released. However, from what I've seen and heard, they provide users only finger tracking and a single trackpad on the top of the controller with (also) no useable buttons. This makes it very difficult to use as a versatile replacement to the Valve Knuckles in all gaming and social situations for VR applications. Overall, the Knuckles are the most versitile, available, and complete package. They are great, amazing pieces of technology provided by Valve. And really, they're the best we got. But based on the issues above, it's become clear to me that Knuckles are simply not enough.

That all being said, while the Knuckles are within warrantee, I've heard great things about Valve Support. Personally, I've only needed to RMA one of my base stations (out of four, at the time of writing), and Valve Support was prompt, kind, and able to get everything sorted out for me -- within less than a week and a half, I had everything set up again and ready to go. While a great customer service could go a long way, from a user-perspective, the fact that individuals are aware that they may need to reach out means that the product inherently has flaws that should have been addressed earlier in the design and engineering stages.

Okay, enough complaining!

Fine, fine -- I do like complaining though, sometimes it's fun! Anyway...

So what now? Do we, both casual enjoyers and deep-diving VR enthusiasts continue using Valve Knuckles? What could we do? Well currently unless you're open to using something like Vive Wands, Pimax Swords, or Etee controllers... Knuckles may be the only option currently, and the best option out there overall for finger tracking controllers with full functionality and features. But let me raise the question: is this the best we can do?

I would argue that the potential for VR controller technology improvements is massive. And to demonstrate this, I've decided that I would like to document, if not build, a few ideas that I've been having for various use-cases and designs for alternatives and potentially viable competitors to the Knuckles. In fact, at the beginning of October, 2023, I was talking with a few community members in the Bigscreen Beyond channels and we agreed that there should exist more alternatives and competitors to the Knuckles -- in terms of ergonomics, input accuracy, reliability, and ease of use. It was from such conversations that I began this exact project: https://github.com/altunidev/SteamVR-Controller .

In other words, it was through our collective dissatisfaction that I began to develop a few ideas for a product that, to begin with, may be an open source and open hardware design. However, I've also heard quite a bit of dissatisfaction with DIY options as these all felt "cheap", or support would be very much a "Oh this broke, hot glue it back together!" rather than having a refined, well-produced item in hand. From these sorts of comments and feedback, I currently have a personal goal of selling these such controllers as products for end-users to purchase at some point in the future. The sooner, the better, but I am but one person in a vast world of opportunity.

That all being said, I do hope that through my projects and goals like these, I can help others be more immersed and, perhaps, enjoy their virtual experiences just a little more -- either together or in their own worlds, however they would like to use this technology!