Catch up on stories from the past week (and beyond) at the Slashdot story archive


Forgot your password?

Slashdot videos: Now with more Slashdot!

  • View

  • Discuss

  • Share

We've improved Slashdot's video section; now you can view our video interviews, product close-ups and site visits with all the usual Slashdot options to comment, share, etc. No more walled garden! It's a work in progress -- we hope you'll check it out (Learn more about the recent updates).

Chrome Software Google News

Google Releases Chrome 25 With Voice Recognition Support 93

Posted by timothy
from the say-what-you-mean dept.
An anonymous reader writes "Google on Thursday released Chrome version 25 for Windows, Mac, and Linux. While Chrome 24 was largely a stability release, Chrome 25 is all about features, including voice recognition support via the newly added Web Speech API and the blocking of silent extension installation. You can update to the latest release now using the browser's built-in silent updater, or download it directly from" But if you're more interested in the growing raft of Google-branded hardware than running Google OSes, some good news (via Liliputing) about the newly released Pixel: Bill Richardson of Google posted on Thursday that the Pixel can boot Linux Mint, and explained how users can follow his example, by taking advantage of new support for a user-provided bootloader.
This discussion has been archived. No new comments can be posted.

Google Releases Chrome 25 With Voice Recognition Support

Comments Filter:
  • by KiloByte (825081) on Friday February 22, 2013 @10:59PM (#42987421)

    It looks like they're making a marketing mistake: they make it sound as if they added recognition of arbitrary text.

    There are only two things voice recognition is useful for:
    * taking a small number of distinct commands
    * producing nonsense poetry that keeps rhythm and rhyme with input voice

    A small corpus of words can be distinguished between pretty easily -- as long as no two are similar to each other. In a real language, with many thousands of words, even a human has a hard time without understanding the subject matter and filling the gaps from context. In fact, what you hear is mostly gaps -- just try to transcribe a series of random words with any real speed. Or, for another example: in a written text, randomly permute all letters except the first and last in every word -- it will still be pretty understandable if you recognize its sense or not at all if you don't. And recognizing the sense is an AI-hard task.

  • by ChunderDownunder (709234) on Friday February 22, 2013 @11:02PM (#42987429)

    Aside from being '64 bit clean', why would you care about RAM?

    Doesn't each Chrome tab run in a separate process, i.e. say each tab addresses 2GB, if you're have 8 tabs open you're maxing out your 16GB workstation??

    Running a 32bit browser on a 64bit OS can be a blessing - Running Chrome on Windows means I don't have to disable (for security reasons) the 64bit Java Plugin the JDK installs.

  • by fuzzyfuzzyfungus (1223518) on Friday February 22, 2013 @11:19PM (#42987497) Journal

    The proposed API itself is agnostic, it just provides a way for a page to ask for mic access and a 'plz speech-to-text-this-audio' mechanism.

    Google's implementation, unshockingly enough, phones right back home to the mothership for speech recognition services. I would assume that(if this proposal makes it out of the cradle) implementations will vary: Google will phone home, Apple will 'siri' home, Microsoft might be awfully tempted to phone home on consumer SKUs, but not on enterprise ones; copies of Dragon NaturallySpeaking will probably include a browser plugin that brings your existing recognition training over to web text-to-speech, etc.

You can not win the game, and you are not allowed to stop playing. -- The Third Law Of Thermodynamics