If you’re a gamer, lag is one of your worst enemies. But what would it be like if you experienced lag in real life? Imagine how frustrating that would be!
Introducing Living With Lag — a cute experiment put on by an internet provider called Ume. Using an Oculus Rift development kit, a Raspberry Pi, noise cancelling headphones and a webcam, Ume’s thrown together a fun social experiment. The webcam captures both audio and video and repeats it to the Oculus Rift via the Pi at a variable delay to show the effects of slow internet speeds.
They attempt four different scenarios. Ping pong is pretty much impossible. Dance class is just embarrassing. And attempting to cook or eat is absolutely hilarious. They even try bowling, which also proves more difficult than you could imagine!
Stick around to see for yourself.
In the end it’s still a commercial, but heck — good job Ume!
[Thanks Brandon!]
And the hack is?
The technology is the first hack (putting it all together to create the effect).
The experience is the second hack. You can experience a virtual-world problem in the physical world.
It is a hack of the human visual system. By the decoupling/extending the normal vision processing delay (13ms – 150ms) of normal event-related potentials (ERPs), you can understand the interplay of even simple motor motions (just like the guy trying to pour pancake mix into a non-moving pan) that sometimes reveal signs of neural processing well before the motor output. i.e. You would do better in a lot motor actions by closing your eyes when ‘lag’ becomes an issue. It also relates why closing your eyes during mental doubt or during the final decision making stage equates to ‘better’ decisions being chosen.
Plus is it just funny to watch… ;-)
How would they simulate the kind of lag I most often experience, rubber-banding? My actions are timed correctly, but the latency from me to the server interferes, rather than the other way around.
With a huge rubber band of course!…well.. ok then maybe not.
some of their actions in the video seem to be more for comical response than actual real life.
I’m actually curious as to the affect on the body if they did a similar test to those of george stratton and his inverted glasses. I am curious if like those the body would adapt and eventually accommodate the delay, at least for things like cooking. dancing would still be off because the sound vs image would not be right but general orientation might be corrected.
nah, the sound would get synced to the image in the brain, brain is magical like that
now where did I read about it hmmm, there it is
http://blogs.scientificamerican.com/observations/2011/09/15/time-on-the-brain-how-you-are-always-living-in-the-past-and-other-quirks-of-perception/
the 80-millisecond / 30 meters rule
coincidently this http://singularityhub.com/2012/03/29/speech-jamming-gun-from-japan-silences-people-from-up-to-34-meters-away/
has similar range and I suspect it has something to do with part of the brain doing the syncing
Nice read buddy, I dont know how much I spent reading these articles!
There was a study (that I can’t seem to google up) about perception of lag when using GUIs. The GUIs had buttons that flashed slightly after being clicked.
The interesting thing is that abruptly removing the delay between click and flash caused the next couple flashes to be perceived as happening *before* the click, like the computer read their mind.
This isn’t really “in real life” like you meant it, but it’s interesting because it shows perception is so malleable that even apparent ordering can change. Maybe it’s just me, but this seems really freaky. Like, cyberpunk “what if none of this is really real maaaan” level freaky.
I’m left with the impression I get from most informercials, that they’re making it look a LOT harder than it really is. You don’t crack an egg by holding it a foot away then slamming it on the bowl. Likewise, bowling is more about the fluidity of motion than knowing where your feet are, don’t move your head around and you should have a general idea where the lane is.
Dance would be fine, if she knew the steps and could hear the music. But it seems like she was just trying to follow along – which is hard enough when you’re not impaired by latency.
You should try it yourself and then say that again. Recently I’ve tried to speak with friend who was on loudspeaker. I heard strong echo with about 0.5 – 1s delay. Seems like nothing and I should be able to speak normally, but I’ve had problems even with thinking. It was impossible not to stutter even when I know about effects of such echo already (I’ve read about it in several studies). After about a minute I’ve had enough and ended call.
https://www.youtube.com/watch?v=oU9EGeMP5n4
I get migraines that cause some sort of visual lag. It is super easy to correct for it by not using your vision as a primary sense and relying more on a mental image of your area.
Cracking an egg, cooking it, flipping it, then walking back to the dining room to eat it like any normal person while I am blindfolded would be trivial for me. I assume people with bad or no vision are the same way.
I guess my point is, the human mind adapts well with experience. Yours ( and mine ) just does not have enough experience with audio lag to compensate.
This. My migraines make my vision spotty and unreliable for depth perception so you learn to build a mental image from other cues. Be willing to be the same would happen here with prolonged exposure.
You crack an egg by holding it above the edge of the bowl, applying sufficient force to accelerate it to “breaking velocity”, continuing the trajectory and then applying deceleration force just before the egg hits the edge in order to avoid completely severing the egg into two halves in your hand. The deceleration command has to happen early in order to account for nerve delay and inertia of the muscle and bone.
Guidance toward the bowl edge when viewed from the side may also require dynamic correction during the trajectory.
A similar break in time-sense can be experienced by trying to toss and catch a ball in a dark room with a fast disco strobe for light.
I noticed it used a single webcam as input, so no stereo vision and less depth perception. Plus the FOV is probably different too.
So this is not just lag. But seeing it’s a commercial I guess we should not expect honest science.
This is a good point. If optical feedback is present, we couple to it. They probably would have done better just taking there bearings for a second and closing their eyes, but that’s not the natural reaction, even with lag, I would think.
I saw a video on that the other day, namely the McGurk Effect: http://youtu.be/G-lN8vWm3m0
It’s pretty easy, in my case, to decouple and hear the sound for what it is, but still, it’s a neat illusion.
not to mention raspi was produsing something like 15fps from that webcam
Yeah, see, I finally figured out that bowl-edge cracking is horrible. I make less mess and get fewer eggshells in my eggs when cracking on a tabletop, which as other posters have mentioned, can be done almost entirely by feel alone. My vision is quite poor without my glasses, and my hearing quite acute (I have developed the famed ‘sixth sense’ that blind people get, can tell when someone is coming up behind me by the sound of the room) so I imagine that some of the difference would be in reliance on vision vs multiple senses.
Except that the sound is also lagging behind completely nullifying your “sixth sense”. They’re wearing noise cancelling headphones with sound playback. The sound is lagging behind as well.
That’s because it is an infomercial.
infomercial
noun – chiefly North American – an advertising film which promotes a product in an informative and supposedly objective style.
So no. This is just a normal commercial not an infomercial.
The cooking bit looked more like the problems were caused by mono-vision and the camera not being where your eyes would be.
Neat hack, too bad it’s a commercial for an ISP (sweedish?). Was on reddit yesterday or the day before. Looks like they intentionally stacked the deck against the users, by doing things like, using a single webcam to run both eyes, and setting the lag to 3 full seconds. I never even saw that sort of lag on my 14.4 modem playing Doom 20 years ago.
3 second lag is not as uncommon as you might think.
Reply from 8.8.8.8: bytes=32 time=3117ms TTL=42
Reply from 8.8.8.8: bytes=32 time=3116ms TTL=42
Reply from 8.8.8.8: bytes=32 time=3127ms TTL=42
This is my connection right now, in my state’s capitol while using a “4g” connection. (US)
It is a miracle that your phone carrier’s network got packets across at all.
And that jitter. It is such a small ratio. I smell buffer bloat.
Your cell tower network has massive, huge, soul crushing overpopulation. They tried to fix it by increasing the memory on equipment and increasing the number of cheap routers. What they got was reliably barely usable service that is going to be even more expensive to fix later on.
I suppose a decent analogy would be checkout lines at christmas. Instead of having more checkouts, they put in bigger checkout counters. Instead of hiring more checkout clerks, they hired security guards to keep people in line.
This is cellphone networks everywhere, ever. 2G, 3G, 4G… they all work on the bare minimum number of equipment because actually providing a proper service would eat into your profit margin.
Is “4G” LTE or HSDPA(+) (which is really 3G)? Although that is even quite slow for 3G – I’m in Australia, and I typically get around 1500ms on HSDPA+. LTE on the other hand, gives me <50ms pings typically (I assume this is because it's using OFDMA rather than TDMA).
It's a shame, in Australia our LTE networks are actually pretty damn good (I've hit 80Mb/s – compared to around 4Mb/s on ADSL2+ at home), but on mobile we only get tiny data allowances (1.5GB a month is typical for a phone plan).
I know of too many people that experience this naturally. :D
Why all the hardware, there are several popular chemicals that will produce the same effect with no hardware.
here is the http version
http://www.youtube.com/watch?v=_fNp37zFn9Q
I think a lot of the problem is simply introduced by the video camera not able to honestly reproduce a 180 degree video as seen by human eye. I bet the person can crack an egg even without watching.
And if that was the point of the commercial you’d have a great argument! :)
Everyone has a visual lag to muscle action of about 200ms. If you don’t believe:
http://www.humanbenchmark.com/tests/reactiontime/
You are just used to it, so you don’t realize.
That’s why aimbots are nearly unbeatable when they are playing with hitscan weapons (e.g. instagib in Q3A/Openarena/QuakeLive).
Interesting to compare peripheral vision to direct vision with this one. Peripheral vision is a survival trait to react quickly to change without directly observing.
Your mouse is polling at a rate of 125 Hz assuming it’s USB and not wireless, so that’s 4 ms additional delay on average before it even registers a button press. Your video card’s frame delay is about 16 ms and the video card is buffering three frames, so that’s 50 ms more lag introduced by your very own hardware. Your monitor is also buffering at least 1 frame, usually 2 frames, giving you 16-32 ms lag or more.
Then there’s the processing delay of running java/flash/shockwave etc. in a web browser to run the app you’re using to measure the reaction time, which puts it easily up and over 100 ms, but more importantly, you don’t know exactly how much delay you’re getting because your operating system and web browser are not hard real time software.
That’s why the attempt to measure reaction time using a web app on a PC is folly. Youre measuring more how slow your computer is rather than how slow you are.
Ironically, older computers have less display lag because CRT monitors dont buffer the incoming signal, so there’s just the average refresh rate lag, which was typically 5-6 ms, and because the video card was running at a higher refresh rate and not necessarily even double buffering the output, you could get output from the keyboard onto the screen in less than 25 ms.
That effect is btw. apparent in the above reaction time test data. If you look at the monthly averages going back to 2006, you’ll see that reaction times are consistently getting worse. Why? Because at that time many people were still using CRT monitors and Windows XP, while today we’re all running LCD monitors and Windows 7/8 which has more abstraction layers between the program and the video output, giving more lag.
Actually USB Interrupt polling rates for HID devices (e.g Mouse) are either 10 milliseconds (100 times a second) for the low-speed devices or 1 millisecond for full-speed devices.
No, the possible settings are 1,2,4,8,10 ms which correspond to 1000, 500, 250, 125, 100 Hz. The Windows default polling rate for mice is 125 Hz.
The goPro wifi output has quite the lag, sufficient to reproduce the effect in the video without additional hardware . It’s fun to play with: https://www.youtube.com/watch?v=teHNantMl_A
You can get used to anything in life. I remember when a 250ms ping was considered good using a dial up. You just learned how to play the lag, and lead everything a bit.
If you’re comparing video games to this then you’ve got it all wrong. You won’t learning how to play the lag, video games have since the Quake era had all sorts of incredible lag prediction. Extreme lag on the clients end didn’t actually result in you not being able to kill the opposition but rather resulted in the opposition killing you when you were already past the window or had already ducked due to your version of the events not happening at the same time as the other person’s version of the events.
Valve did a fantastic video demonstration years back of what happens when a game optimised for dialup is played on cable, and when a game optimised for cable is played on dialup. Both were playable but produced weird results like being able to kill another player without ever hitting them. In the most extreme cases neither client placed the target in the line of fire but the server saw the kill.
This is used in the drunk driving simulators, just on the brake and gas. I am not sure of steering in these rigged cars. It gives teens a chance to experience a taste of danger without the holeywood scarey shows often used in schools.
Can’t they just get drunk and play mario kart like we did?
The films we watched were no Hollywood ones, they were the real deal. The corpses you saw on the screen didn’t get up. They were for real dead.
I’ll never forget this kid I went to school with in Drivers Ed class watching one of the crash and burn films. It was called Underride. The gist of it was passenger cars that went under the back of semi tractor trailers. What a mess! Anyway this kid turned an honest to god shade of green. When we turned the lights on in the classroom everyone looked at him and we were all like, Reto, you’re green! I’ve seen a few folks turn white, but it was the only time I’ve ever seen someone turn a shade of green.
I’ll say this for him, he didn’t hurl. Guy must have been scarred for life though, or something. He was an immigrant from Europe. Quite the delicate type I suppose. Thinking back now those were some ultra sick films. I mean I know they were made to shock, but did the producers have to do that good of a job?
There was this other kid I went to school with who was a real horror buff, he couldn’t get enough of those films.
Now I’m doing a web search looking for the video. I can’t find it but this one sounds familiar, I’m pretty sure they subjected us to it too
https://www.youtube.com/watch?v=Rzr7CHX_0b0
Nice but I’d like to see it working at 100 up to 500 ms lag and see how that affects playing table tennis or something. 1 sec was just too much
It reminds me, a few years ago, I gently installed and modified a copy of google’s pacman doodle at http://cloud.gaming.free.fr. It shows how latency can remove any pleasure to play a game.
At the time I did this, many people believed that cloud gaming will become the future of gaming systems, and I was not convinced at all, hence this mini-demo.
They need an Enclosed helmet covering the ears also giving the same Synchronised Lag lol Synchronised lag
They’re wearing noise cancelling headphones and ARE receiving lagged sound.
I’m convinced! Sign me up for your low-lag, high speed optic fiber connection!
Oh, by the way – I’m in Australia.
Just don’t forget that now Facebook is the company behind the Oculus.