Group items tagged

The pictures come from all around the world; as individual photographers and organizations license their high quality photos under Creative Commons, the book will only grow as a community contributed and shareable resource.

The biggest photo book ever… growing everyday with only high quality and 100% relevant pictures due to our community-based curation process.”

First, face-to-face conversation is a dynamic interaction between two people. Talking is a kind of turn-taking, and turn-taking is a rhythm that organizes how and when things unfold. When people talk, they modify their postures and tempos so that they correspond. (There is fascinating research on this sort of "coordination dynamics" by, among others, J.A. Scott Kelso and colleagues.) And moreover, they tend jointly to pay attention to what is going on around them. One upshot of this is that when you are talking with someone, you actually experience that they are listening. Talking, in this sense, is like dancing. You can tell whether someone is dancing with you or whether they have detached and lost interest.

It is often very difficult to establish this sort dynamic engagement or coupling to someone over a cell phone. The connection is often just not good enough for that. One has the experience of speaking, but one lacks the corresponding experience of being heard. You literally lack the feedback — the coordinated breathing, the modulations of tones of voice, etc — that signals the other’s focus on you. And so you shout.

When a cell-phone connection is excellent, you and your remote partner to start up a dance together. But crucially, you dance to music that only you and your partner hear. That is, you succeed in establishing the kind of dynamic integration and entrainment that is typical of face-to-face conversation. And that has the effect of transporting you right out of your physical environment into one that you share with your conversation partner. And this in turn has the effect of guaranteeing that your tone of voice and volume will be inappropriate for your actual physical conditions

ouTube, Google, Stanford, Berkeley, and the California School for the Deaf (CSD) are about to speak on YouTube and accessibility.

Now the Google Speech Technology team is speaking about the challenges they faced to get auto-captioning operational. Their vision was to create accurate captions for all videos in all languages, but had to deal with huge vocabularies, background noise, poor recordings, accent variability, and distinguishing between song and speech.- Google’s approach is to deliver captions from the cloud, given them the ability to rapidly iterate and model at a large scale.

You can only caption your own videos — you can’t just caption someone else’s videos.