There's a lot of ground to cover with this week's installment of Xbox Adaptive Controller hacking, so let's dive right in. First up, I recruited some professional help with soldering. Alex, whom some of you may know as the bubble guy from the TR BBQ, solders for a living and is about 10 times faster than I am at the task. He helped me whip up two of the experimental switches shown below. Thanks, man.
The first switch is this tilt switch. I figured if I'm going to experiment with this stuff, why not explore DIY motion controls? It's super simple: just a ball bearing that completes the circuit when it touches the internal contacts. I've got a couple ideas for how to use this bouncing around in my head, but they haven't coalesced quite yet. I can't shake the idea of an oversized joystick, but that wouldn't really be bringing anything new to the table. At just $6 for ten of them, though, it feels like the sky's the limit for crazy ideas.
The second switch is this magnetic contact switch. It has a lot of potential. When I was testing it, I found that it triggered the button a few millimeters before making contact. Cool. I need to run some more tests and try to trigger it with something other than the magnet that it came with to see if I can extend that distance. I'm spoiled for choice by the concept of a touch-free button that costs less than $2 a pop.
The third experimental switch is based on the Sanwa switches I picked up earlier. I got to thinking that a ball pit ball would make a good topper for the switch if I could mount it securely. The ball is light enough that the switch can pop it back up, and it also makes the switch a much easier target to hit. A few strategically placed drill holes in the button cap and a single zip tie was all I needed to bring things together. I added some Gorilla Glue in an attempt to hold air in the ball, but it's not structurally important otherwise. I'm very happy with the result and plan to follow up the red button with orange, yellow, green, blue, and purple ones. Hey, speaking of colors...
Even though I have a couple options for games to have Ellie try with her controller, I couldn't shake the feeling that a custom-made game was the right place to start for learning about how Ellie would interact with it. When it comes to programming problems, the person I turn to is my nephew (sorry, Bruno). He recently helped me with renaming my saved .wtv files and he's the one to blame for the abomination that is Jellies vs. Whales.
What we would later call Sundoodler started as my request for him to make a game that was like an "RGB Etch A Sketch" where "the cursor could roll off the edge of the screen, Pac-Man-style." Just over an hour later and with the only additional input consisting of links to some rainbow-colored yarn and a sunglasses-clad sun, he was finished. The kid is good.
Ellie's custom Xbox Adaptive Controller rig - Spherical Image - RICOH THETA
With Sundoodler fired up on the floor projector and the switches of my contraption plugged into the A, B, X, and Y ports on the back of the Xbox Adapter Controller (directionally corrected for portrait-mode play), it was time for the big test! Ellie was acting pretty sleepy but she did spin the rollers a bit, both with her left arm and her feet. This was the first time she'd spun them with the switches adding their resistance to the rollers, and it didn't slow her down much at all. Phew! She wasn't big on watching the ceiling at the same time as spinning, but she watched a little while I was playing via the secondary inputs on a keyboard.
As for the frame itself, I've wrapped it up to add some color and also to cover up the screws I put in the bottom of most of the joints to hold everything together. I'm reasonably happy with its appearance, but I still want to switch to pool noodles at some point. My plan to use Q-tips and Gorilla Glue to lock the switches into their calibrated positions worked perfectly. I would estimate that the keys register a press correctly 99% of the time, and when they don't, it's because my ribbon-covered key stem technique proved a bit too hard to push back against. I expect this will improve with time and wear.
This post probably marks the end of my weekly cadence for the series. I need some extended time to observe and adjust what I have so far, as well as to work out just exactly what I want to create with the newest experimental switches, before I'll have content for another post. It's exciting just to know that the new switches are an option, though, and I'd appreciate your help brainstorming about them in the comments. Thanks to all of you that have contributed your thoughts and ideas so far. I'll be back when I have more to share or if Ellie does something awesome with what I have so far—whatever comes first.Building a custom Xbox Adaptive Controller peripheral, part four
If you're just joining this series, I encourage you to flip back to parts one, two, and three to get yourself up to speed. They are quick reads with short videos and should give you all the context you need. Last time, I introduced my final pick for the switches I wanted to use in the controller and shared a rough idea of what I wanted the frame to look like. I had a lot more time this weekend to work on the project, and I'm very happy with the progress I made and the results so far.
I started with the assembly of four of my switch holders. I'm still amazed by the coincidence of how well these things fit together. Between the PVC/switch duo and the massage roller-paint roller combination, this project has proved far simpler to build than I ever imagined. After the switches were together, I plugged them into the Xbox Adaptive Controller (now paired with the NUC-in-a-box), and made sure they were all registering button presses as expected. Everything was all good.
With the switches secure in their mounts and functioning correctly, I tested out an idea that came to me just a couple days earlier. I wanted to know if using an audio splitter would allow multiple switches to control the same button press on the XAC. I was already dreaming up all kinds of ways that could be used before I confirmed that, yes, it worked exactly as I'd hoped. Two different physical switches could trigger the same input on the controller. Nice. This discovery is going to add a lot of flexibility to future designs. I now expect to take a modular approach to some of the additions I have in mind. It'll be easier than breaking down and reassembling a single gargantuan contraption. Allowing for multiple methods of triggering the same button on the controller is also a huge boon to creativity.
Perhaps what I'm most excited about when it comes to the input splitting is that I can build an entirely separate controller to test out my other switch ideas and use both controllers with the XAC at the same time. There's a lot more stuff I want to try out, but as you'll see below, my original concept is coming together quite nicely, and I just don't feel like slapping extra bits and bobs on is necessary or wise at this point. But I love bits and bobs, so I expect to build a second controller full of my crazier ideas and aimed at allowing our 17-month old, Wren, to join in on Ellie's experience, whatever that ends up being. In my head, I'm picturing a reversal of the trope where an older sibling helps a younger one get past a level. We'll see. At any rate, let's check out what her experience is looking like so far.
I've got to say, wow, the Kailh Speed Bronze switches were a game changer. Thanks to their fast and audible triggering, it was super easy to calibrate the rollers for minimum resistance while also being confident I was getting a solid button press. I haven't glued anything into place yet, but when I do, I won't have to worry about whether I'm locking them into a bad position or not. After I take a couple days to think things over, I'll probably use a combination of glue and screws to fix everything into its ideal position. After that, some zip ties, pool noodles, and athletic wrap should take care of the rest, and I'll finally be able to try it out with Ellie for a real play session.
First up will be Proteus, a game that TR gerbil Pbryanw suggested early last month when my XAC first arrived. Beyond that, I have precious few ideas for what kind of experiences Ellie could actually use the controller to enjoy. The difference between her having fun spinning some rollers and realizing that spinning them is influencing what is happening on her ceiling is not something we're taking for granted. It's a pretty big leap for Ellie, but on the cusp of feasibility given her abilities.
My wife offered up an interesting idea that I'll be noodling on for awhile. Maybe Ellie's interaction with the controller could be thought of less traditionally. Don't have her play a game, but let her participate in her own way. Maybe Ellie could be an element of chaos in a co-op game, like the world's cutest RNG, or perhaps a rogue chef in Overcooked. Or maybe there's a role for her to play in an asymmetrical game—something that puts us at her mercy, like the player that controls Bowser in Mario Party for the Wii U. I think she'd like that. Now, we just need to find such a thing. Any ideas?Building a custom Xbox Adaptive Controller peripheral, part three
In part two of this blog series, I introduced my idea of using paint rollers as the mechanism for allowing easy rotation of the foam massage rollers that trigger keyboard switches connected to the Xbox Adaptive Controller. If that perfectly logical sentence isn't crystal clear to you, you may want to start from the beginning before reading further. As things worked out, I didn't have near the free time over the last weekend that I'd hoped for, and I only made a fraction of the progress I'd intended. However, a couple helpful suggestions from the comments on last week's post lead to the acquisition of some new hardware that's worth talking about, let's start there.
TR gerbil llisandro read my mind and suggested Kailh Speed switches as an option for a switch that would actuate with a very small amount of travel. That's something I've discovered is especially important for the method by which they'll be triggered in my project. I ordered Bronze and Copper switches from Novel Keys, and they arrived in just two days. I think I'm going to be very happy with them, and they fit inside my threaded PVC adapters even more snugly than my original switches. Another helpful post came from TR gerbil massbart, who suggested I take a look at Sanwa arcade buttons. I ordered a six-pack from Amazon and I'm excited by their potential. Check it out.
As you can see by the end of that video (or just the thumbnail if you didn't watch it), I do have the rough shape of the "controller" coming together. I wanted to start by getting the outside dimensions established before splicing in the other connectors to hold the switches and legs of the device. With the assistance of a century-old rubber mallet that I got from my grandfather, the frame is holding together ever better than I expected without gluing or pinning. I think the hose clamps on the paint rollers are actually compressing the ends of the tee connectors enough to get a better grip in the pipe inside. It may end up that force is all that's needed to stop undesired torsion in the frame.
I do need to make the frame at least a few inches longer, though. The top of Ellie's head hits the tube running across the top as things are now, but it's close. I'll probably also narrow it by a few inches after some additional testing. Beyond that, I obviously need to cover up any pokey bits with something soft (possibly pool noodles) and fish the wires through the tubing and out one of the fittings. Now, speaking of testing...
I can't express how stoked I was to see Ellie start spinning that roller, even if it didn't have a switch on it quite yet. Clearly, the paint roller was much more successful than my first attempt with trying to have her spin the roller directly on the PVC tube. She was so proficient at it, so quickly, that it has me wondering if I should buy two more of the little rollers and just use them for the foot switches as well. The big pink roller is starting to look pretty clunky and difficult to spin by comparison.
If I used small rollers for the feet, I could set one for forward movement and the other for jumping, that's pretty tempting. Without the big pink roller, I wouldn't even need to add legs to the frame because the green rollers have plenty of clearance just from the height of the fitting they are strapped to. It would also be nice to keep the profile of this contraption very low for easy storage. I think I'm talking myself into this idea. Surely, I can find another use for the big roller...Building a custom Xbox Adaptive Controller peripheral, part two
Last week, I officially introduced you to my project to build a custom input device for my daughter Ellie that she could use with the Xbox Adaptive Controller. I've had some time to think and fiddle since that post went up, and I'm ready to catch you up on the latest developments.
First up, I want to share a look at the screen Ellie will be viewing her gaming sessions on. It was no great feat to put together, but I'm happy with the outcome. After a busy day of sitting up at school, Ellie spends most of her afternoons chilling in her swanky new bed. Logically, that meant projecting a screen onto her ceiling. To get there, I zip-tied an old Haswell NUC and a cheap LED projector I had laying around into a small plastic basket with an extension cord running out of it. Behold.
That certainly puts a whole new perspective on The Magic School Bus. Part of me regrets not buying a Gigabyte Brix PC with an integrated projector when they could be had for a couple hundred bucks a few years back. That would be pretty slick to use, but this solution isn't bad. The plastic shell keeps everything tidy and protected from overzealous vacuum robots, and it's stable enough to place nearly anywhere.
Back to the actual building of the controller. I've decided to continue pursuing the spin-actuated mechanical keyboard switch plan. An errant paint roller in the back of my car turned out to fit perfectly inside the rollers I picked. That made a huge difference in ease of construction and proved easy to integrate into my design. However, I may still need to source a switch that actuates at a shallower depth, so I can leave a larger gap and potentially reduce spinning resistance even further. A hair trigger is desirable in this case.
A bag of hose clamps and $10 worth of paint rollers later, and I arrived at what you see above. I'm very pleased with how securely the hose clamps hold onto the rollers. They all but eliminate motion on all three axes thanks to the tee connector. That means the integrity of the finished controller will come down to the PVC connections themselves. I'm still trying to decide if I want to glue them or pin them in place. I know I don't want to use actual PVC cement to permanently fuse them, though. I'm just not that confident I'll be able to avoid breaking the contraption down for fine-tuning after assembly. Now, let's take a look at the rollers in motion.
I'm confident enough with the outcome demonstrated in the video to move ahead with trying to complete the first full prototype this weekend. I just need to double-check some measurements to make sure Ellie can reach everything before putting the frame together and routing all the cables through it. I suppose I also need to put some more thought into making sure nothing will tangle up in Ellie's hair, even though this will obviously be a supervised toy.
In the next update, we'll be looking at the entire device all put together and how it works out on the software side of things (I hope). Please don't be shy about sharing your thoughts and advice on what you've seen so far. I'm just making this up as I go along, and I'm open to testing out other ideas.Building a custom Xbox Adaptive Controller peripheral, part one
A few weeks ago I unboxed and started experimenting with the brand new Xbox Adaptive Controller and its auxiliary 3.5mm headphone jack inputs. Ultimately, my goal is to build a peripheral that my daughter Ellie can use to have an interactive experience on a PC. Ellie has some fairly significant physical and mental limitations, courtesy of her syndrome. But make no mistake, she's plenty feisty, and curious enough that I'm confident we'll be able to figure out something that will at least be a source of amusement for her.
For this project, the name of the game is accessibility. I don't want to do anything overly complicated or that anyone would have a difficult time replicating if they were so inclined. I won't be hitting up my friends with 3D printers to produce parts for me and I won't be spending a lot of money either. Improvisation, creativity, and flexibility are the most important things. I'm adapting a controller to suit Ellie's abilities, not the other way around. It's innovation by way of restriction.
I don't know about you, but the keywords above make me think of a few specific construction materials. PVC pipe and fittings are near the top of the list. Of course, for the actual switch, I needed something slightly more sophisticated than my first proof of concept. I chose something familiar to most gerbils: mechanical keyboard switches. I figure there are dozens of fun ways to actuate keyboard switches and I know they're built to last. I got started by soldering one up to the leads I'd stripped off the audio cable in the previously linked video. Once I confirmed that worked, well, I'll let myself take it from here:
The design in the video is predicated on the fact that Ellie likes to spin things. This is one of her favorite toys. The spinning action that I know she's comfortable with and capable of seemed like a good place to start. The next step was going to be making a rectangular frame large enough for her to fit inside of while laying in her bed. The rollers would have been located where her unique anatomy would be afforded the most leverage on them.
I say "was" because, so far, Ellie does not seem impressed by dad's handywork. I have a couple ideas for making the rollers easier to spin and more interesting to Ellie, though. Even if those ideas don't pan out, I might go through with my original plan anyway, just to see what I learn and to have a completed platform to build on for whatever comes next. Plus, I'm attached to the roller idea because it reminds me more than a little bit of a player piano, or punch card programming, and could lead to "physical macros" where moving and jumping can be rolled into one motion or something along those lines.
That said, I'm already thinking of other ways to trigger the switches, especially for her arms. I like the idea of something she could flick up and would hit the switch when it comes down again. One of the nice things about the keyboard switches is that they are so easy to trigger, you don't have to overcome a lot of force. I feel like focusing purely on motion and not leverage is probably a better course, and a flaw of my original concept. Or maybe she'll like it better if I let the rollers move loosely in a channel and treat the switches as speed bumps. I need to run some more tests.Twin eagles: building matching Ryzen systems for teens
A wise man once said, "Fish, you idiot, Ryzen is almost here! You should have waited." It took a while, but all the way back in mid-June I finally got my hands on a Ryzen system—two of them, in fact. With the school year wrapped up, and an entire summer looming with kids in need of entertainment, two of my coworkers asked for my help replacing their sons' PCs.
Last time I built computers for these guys, it was four years ago, and the budget was much smaller. Back then, all those kids were doing on their PCs was playing Minecraft and Team Fortress 2, and two-core, two-thread Haswell Pentium G3220s and GeForce GTX 750 Tis were plenty good enough for their needs. Times have changed, though, and these now-teenagers use their PCs for a wide variety of tasks, including video editing, music creation, 3D modeling, live streaming, and much more demanding gaming. "What's a computer?" indeed...
Anyway, back to building these kids' new PCs. I said yeah, I'll build them, but on two conditions: one, that everyone participates in the process and two, that the kids document everything on their phones to produce a video I can share on TR after it's done. It took two and a half months, but they eventually got around to it. Amusingly, and in a fashion befitting teenagers since time immemorial, the video was haphazardly slapped together, rendered out on one of the new systems, and delivered long after it was originally expected. Before we share that treat with you, though, let's take a look at the TR-staff-approved specs.
Originally, the identical PCs were going to be Intel Core i5-8400 and GeForce GTX 1060-based builds. However, the first decent video card deal of the summer got the entire budget bumped up to a level that seemed like a better fit for a CPU with a bit more oomph (especially since the boys aren't strictly gaming on them). The builds went smoothly enough, and I had no major complaints about any of the components. It was odd to have to install motherboard standoffs, though. I haven't done that for a while. Oh, and if you ever build in a Focus G, hook up the front-panel audio connector before you install the PSU, because it can't fit through the gap otherwise.
Once the PCs got to their new home, our intrepid builders found that their chunky HDMI cables collided with the slots on the back of the case, preventing them from plugging in all the way. Many of you have probably seen that before. I suggested a DisplayPort adapter they could pick up locally, and that saved their first weekend of playing with the new systems. It was pretty much radio silence from them after that.
Now, brace yourselves for the video that was decidedly not crafted carefully during all that summer free-time (and got rejected by TR staff three times before they let me embed it).
That, my friends, is an entire afternoon of education and explanation as seen by the mind of a 14-year old. You can almost hear the Charlie Brown adult-voice over the bass line. I'd like to think that some of the finer points of PC building sunk in, but I'm happy just having the kids participate in building their own rigs and getting introduced to the basics.
Fast forward to today. Since we now have the perspective afforded by an entire summer of use, I'm happy to report that there have been no problems with either system. I was a bit nervous given the stock cooling and purported high temperatures of the attic-adjacent gaming room the PCs and their Xbox One cousins reside in. It shouldn't come as a surprise, though, because the boys asked for help installing Gigabytes's RGB Fusion software pretty early on and it wasn't much longer after that before they wanted to tweak other things, including fan-speed profiles. I guess that makes RGB LEDs the gateway component. Doesn't that just warm the heart?The Tech Report attends Silicon Valley Virtual Reality 2016
I spent a few days last week at the Silicon Valley Virtual Reality Expo in San Jose. SVVR is a fairly small convention that showcases the latest developments in VR. The Turris VR chair that we covered last week was at the convention, but since it got its own article already, I won't talk about it here other than to say that it was subjected to nearly non-stop use for three days and got plenty of positive feedback. Turris aside, here's what I did and saw at the event.
Palmer Luckey from Oculus presented the event's keynote. Unsurprisingly, he's bullish on the future of VR. My plan to use Cortana to dictate notes failed epically (and humorously), though, which means I've got nothing to share that I'm confident he actually said. You'll have to settle for a recap of the other experiences I had at the show instead.
The first VR device I saw at the show was the somewhat imposing-looking Infinadeck. The Infinadeck is a true VR treadmill that allows for freedom of movement in any direction, thanks to a treadmill deck comprising individual sections of belt that can move in a perpendicular direction to the main belt. After signing a personal safety wavier, I hopped on and buckled in. There was no VR aspect to this demo, and I quickly found out why. It was plenty tricky enough to stay balanced with the floor effectively moving underneath me as it was. The deck moves when your waist pulls on the belt you're strapped into. The direction of the deck's motion (which can be any angle) is dictated by the armature above your head.
The deck itself delivers its promised functionality, but the whole system was fairly loose and sensitive to unintentional motions. The sensitivity left me in a constant state of unbalance, which explains the safety wavier and the lack of a Rift or Vive to goggle in with. You need all of your real world faculties just to stay upright on the Infinadeck. I suspect that with some adjustment and practice, this system would be something you could get better at operating, though. The Infinadeck appears to be for professional use only—I don't even want to know what one would cost.
Mobile VR was a huge deal at SVVR. There were probably 10 Gear VRs or similar devices for every wired Vive or Rift (and about a 5:1 ratio of Vives to Rifts, by the way). The second demo I tried was from the folks at Interactive Lab. It used a Galaxy Gear with a Galaxy S7 clipped into it. The demo was a room-scale setup that used reflective balls on the headset and a gun controller. Those reflective balls were tracked by emitters and cameras. In the demo itself, your character is placed in a space station that is being invaded by robots. I could move around and duck under cover to avoid being shot, but there didn't seem to be any penalty for getting hit.
The graphics quality of the demo was fairly impressive for a mobile solution—with a smooth frame rate, to boot. It was also extremely nice to try room-scale VR without being tethered at all. I didn't get any hint of motion sickness, either. As PC enthusiasts, it's easy for us to focus on the the high-end stuff, but this demo proved to me how compelling mobile VR at room-scale can be, as well.
The next demo I tried was Talaris Technologies' Stompz. Stompz are small devices that clip onto your shoes and translate the motion of walking in place into movement in a VR experience. Stompz don't use positional tracking, though. Instead, the associated software translates accelerometer signals from the devices into either a key press for walking forward (like W in our time-honored WASD) or a modifier and a key press to run (like Shift + W). The thresholds to trigger each control are adjustable to accommodate however much real world effort you want to exert. Talaris used a CV1 Rift for its demo, and movement direction was controlled by the direction the headset was facing.
Not to pick favorites here or anything, but I have to compare the experience of using Stompz and using the Turris. Mine is an extremely niche case, but coming from the Turris, I immediately felt uneasy about my walking direction being solely dictated by where I was looking—especially because I was walking in place instead of sitting. I do think that simply walking in place is probably a better solution to VR movement than some of the stationary treadmill options that are out there, though, because Stompz-like devices should require less effort and space to set up. If Stompz or something similar was upgraded to work with the Vive's lighthouse trackers, it could become something pretty special, though (come on room-scale goalie simulator!).
The only demo I had read about before getting to SVVR was Tactical Haptics'. The Tactical Haptics device is an add-on that clips onto an HTC Vive controller to take advantage of the absolute positioning data the controller provides. The add-on has a trigger you activate with your index finger, but the special sauce is actually built into the grip that the rest of your hand wraps around while holding it. The front and the back of the grip are rubber pads that can actually slide up and down the handle. When you're holding the Tactical Haptics device, this motion is translated into surprisingly convincing interactions with objects in VR.
In my favorite part of the demo, three virtual bungee cords hang in the space in front of you. Each one is a different thickness, and each responded differently when I reached out and stretched them. The sliding of the grips comes across to your hand as almost a torque-type sensation as you stretch and twist the bungies. There's obviously some pretty fancy software magic going on to make the interaction of the VR simulation, controller feedback, and real-world movement all sync up so well. It reminded me of a VR version of the Novint Falcon's abilities, but that's so obscure I doubt it will help many people understand. All told, Tactical Haptics' device is something I would definitely want to include in my VR experience if I could, but as of right now there is no release date or price for the product.
Life of Lon
One other demo I tried was for the VR game Life of Lon by Block Interval. When I tried it, I didn't realize I was playing a demo for an actual game instead of a tech demo that was showing off some particular VR innovation. I have to say that the demo probably lasted about five times longer than it needed to, thanks to an extended and non-interactive opening sequence. Maybe it was made with VR rookies in mind. After the demo finally gives you control, you meet a cute alien sea-otter type character and (hopefully) your curiosity compels you to follow him. The gameplay consists of pressing the A button on an Xbox controller and then tilting your head left and right. It's pretty intuitive and, with the right game design choices behind it, could be a viable control mechanism (but perhaps a bit of a limiting one, as well).
There are an insane number of people and groups all trying to work out this whole VR thing. I seems like everyone has something interesting to contribute, but no one could possibly say which ones are going to stick at this point. There were a ton of presenters at SVVR, and I only had time to check out a small number of them. I'm told that the Project Alice demo from Noitom and partners Lenovo and Nvidia was particularly impressive at combining different elements of VR, but I wasn't able to try it.
There were also dozens of different solutions for capturing 360 video for use in VR as well. It was clear that VR is not just the realm of real-time rendered images. I really need to get working on something like the 360-degree video drone above for the BBQ—the show gave me lots of new ideas for bringing that event to the world.
I'll wrap up my thoughts about SVVR with this. If you think hats in TF2, decorations in The Sims, horse armor, Counter-Strike skins, and cosmetic DLC in general are ridiculous already, just wait until social VR takes off and all those digital goods become virtual goods. It's only going to get weirder from here, folks.
|All the Nvidia GeForce GTX 1660 Ti cards so far||26|
|Bargain basement: a Ryzen 5 2400G with a Gigabyte mobo for $210 and more||11|
|Nvidia's GTX 1660 Ti : Turing sans RTX||62|
|What is Intel’s graphic “Odyssey”?||44|
|Adata's SU750 SSD serves up TLC, hold the DRAM||23|
|National Grain-Free Day Shortbread||38|
|Intel teases discrete graphics community "Odyssey"||87|
|All of Noctua's coolers now include mounting hardware for socket AM4||8|
|Tuesday deals: an Intel 660p 512-GB NVMe drive for $67 and much more||28|
|Every review of an SSD without DRAM cache that I have ever read reinforces the same, tired old conclusion. DON'T RUIN $100 of NAND by saving $2 on DRA...||+36|