The Emerging Vocabulary of Haptics: Vibration vs Texture

Ed Colgate
Feeling Disruptive
Published in
4 min readJan 2, 2017

I’ve written before of my disdain for the word “buzz.” Perhaps I should feel differently since buzz speaks to the earliest commercial incarnations of haptic interface (e.g. one type of buzz for an incoming call from my wife, another for my son, etc.) It is true that buzz helped get some legs under haptics, proving that it could do something of value. But with legs you can walk, maybe even run. Now it’s time to push on.

To frame the discussion, let’s talk about two other words: vibration and texture. Clearly, they ain’t synonyms.

  • Vibration encompasses buzz, but also the resonant purr of a cat, or the throbbing hum of a five liter V8. Something is actually vibrating, pushing air around, so we can both hear these and, if we touch, feel them.
  • Texture is a little less aggressive. Silk sheets or pebbled leather are content to sit quietly, awaiting our touch. When we touch them, the texture is distinct, recognizable. Often, we can hear the sound of our fingertips sliding over the texture, but that sound plays a secondary role — it reinforces what we feel.

With traditional vibrotactile approaches to haptics, vibrations are the only choice. That’s because they work by getting a little slug of mass to vibrate, and there’s just no way to sync the movement of that mass to the movement of the fingertips on the touchscreen.

Actually, it is a little worse than this most of the time. To be energy efficient, a lot of vibrotactile devices are designed to vibrate preferentially within a narrow range of frequencies. And what’s the most common choice of frequency? Well, it’s the one that is easiest to feel, about 250 Hz, which is — you guessed it — buzzy. When companies talk about HD, what they are really saying is that they can create vibrations over a wider range of frequencies. But, hey, they’re still vibrations.

With surface haptics, on the other hand, both vibration and texture are possible — either separately or at the same time. You may recall from my previous posts that surface haptics is a localized effect caused by modulating the friction underneath each individual fingertip. Roughly speaking, if you modulate friction as a function of time — independent of the particular movement of the finger — you feel a vibration. If you modulate friction as a function of finger position, you get a texture.

In practice, modulating friction as a function of position is a little more easily said than done. You have to start by measuring the position of the finger (that’s what touch screens do), and then you have to modulate friction in accordance with that position. Sounds simple, but it turns out that your finger position measurement needs to be really good (low jitter), and the total time required to 1) make a measurement, 2) compute a friction value based on that measurement, and 3) update the friction, needs to be really short. We call this round trip time latency. If either the jitter or the latency is too high, the texture starts to feel … you guessed it … buzzy. And I am not a fan of buzzy.

So, at Tanvas, we’ve worked really hard to solve the jitter and latency problems. And we’ve done a pretty good job, IMHO. The jitter is sub-pixel, and the latency is typically a couple of milliseconds. What that means is that our tech can display some pretty cool textures.

And actually not just some, but instead a near infinitude. Because those textures depend on position, they’re kind of like images — an image is a pattern of colored pixels, a texture is a pattern of friction values. With Tanvas’ surface haptics, you can more or less paint any texture you want.

Just to summarize:

  • Vibrations and textures are distinct classes of haptic interactions.
  • Vibrations can be created by vibrating (!) a little slug of mass inside a device, or by varying the surface friction underneath a fingertip as a function of time.
  • Textures can be created by varying the surface friction underneath fingertip as a function of position. But to do this right, you need low jitter and low latency.

One final note: Earlier I said vibrations depend on time and textures on position, roughly speaking. The reason that’s only rough is that the world of textures — one that we’re just beginning to understand deeply — is in fact much more expansive.

For instance, last year our lab published a paper showing that, if you add speed-dependence to a class of wavy surface textures, you can make them feel more rubbery. And that’s just the beginning. One of my doctoral students, Craig Shultz, is working on what he calls “dynamic textures” that combine position and time dependence in some really interesting ways. So, to say that textures depend on position is a pretty good start, I think. But, only a start.

--

--