Google Home Mini preview hardware defect causes near-constant listening, uploads to server...

2»

Comments

  • Reply 21 of 36
    avon b7avon b7 Posts: 2,642member
    gatorguy said:
    gatorguy said:
    Rayz2016 said:
    With the feature enabled, Google sends not only what the user says when a keyword is spoken or a device is activated -- but the sound recorded a few seconds before the actuation

    How can it record sounds before it was activated?
    Another example of the "always listening but not sending to a server" is the song recognition feature in the new Pixels. Yes that on-board microphone is always "listening" but without a connection to Google servers. All the recognition in done directly on the device with no need for any data sent to Google. But if the user then wants to get more information, perhaps read up on the artist, see what else they've got, maybe start a radio station using that song as the basis they can actively request a server connection to accomplish those actions. 
    How does the Pixel analyze heard media without talking to a server of audio fingerprints for matching? There’s a local copy of these fingerprints on every Pixel? 
    Yes there is. Approx. 20 thousand of them which will get updated as new music comes out. All song recognition is done on-device with zero data sent to Google.
    Yes. There are at least three angles to the idea. Lack of an internet connection. Privacy. Latency. NPUs will help out a lot with the processing although local mobile AI is still in its infancy.
  • Reply 22 of 36
    Rayz2016Rayz2016 Posts: 4,047member
    gatorguy said:
    Rayz2016 said:
    gatorguy said:
    Rayz2016 said:
    With the feature enabled, Google sends not only what the user says when a keyword is spoken or a device is activated -- but the sound recorded a few seconds before the actuation

    How can it record sounds before it was activated?
    Common sense would say: On board hardware. Some chip is digitizing, recording and examining a second or so of sound all the time. How else to recognize a keyword since it relies on a word/phrase that was already spoken?

    Another example of the "always listening but not sending to a server" is the song recognition feature in the new Pixels. Yes that on-board microphone is always "listening" but without a connection to Google servers. All the recognition in done directly on the device with no need for any data sent to Google. But if the user then wants to get more information, perhaps read up on the artist, see what else they've got, maybe start a radio station using that song as the basis they can actively request a server connection to accomplish those actions. 

    Right. 

    Yes, I got the part about the “listening”, I just didn’t realise it was recording all the time too. I thought it only recorded stuff when you started speaking. 

    Blimey. 

    I doubt it goes back to this morning. :) It's not unlike Live Photos. It keeps the prior second of sound as part of the voice activation process diagnostics. 
    Yeah, but I knew Live Photos was doing it. 
    edited October 2017 airnerd
  • Reply 23 of 36
    gatorguygatorguy Posts: 19,061member
    Rayz2016 said:
    gatorguy said:
    Rayz2016 said:
    gatorguy said:
    Rayz2016 said:
    With the feature enabled, Google sends not only what the user says when a keyword is spoken or a device is activated -- but the sound recorded a few seconds before the actuation

    How can it record sounds before it was activated?
    Common sense would say: On board hardware. Some chip is digitizing, recording and examining a second or so of sound all the time. How else to recognize a keyword since it relies on a word/phrase that was already spoken?

    Another example of the "always listening but not sending to a server" is the song recognition feature in the new Pixels. Yes that on-board microphone is always "listening" but without a connection to Google servers. All the recognition in done directly on the device with no need for any data sent to Google. But if the user then wants to get more information, perhaps read up on the artist, see what else they've got, maybe start a radio station using that song as the basis they can actively request a server connection to accomplish those actions. 

    Right. 

    Yes, I got the part about the “listening”, I just didn’t realise it was recording all the time too. I thought it only recorded stuff when you started speaking. 

    Blimey. 

    I doubt it goes back to this morning. :) It's not unlike Live Photos. It keeps the prior second of sound as part of the voice activation process diagnostics. 
    Yeah, but I knew Live Photos was doing it. 
    And now you know our Assistants do too. ;)
  • Reply 24 of 36
    Just wait 'til it can read lips.
    watto_cobra
  • Reply 25 of 36
    gatorguygatorguy Posts: 19,061member
    The latest word from Google is that the activation button will be permanently disabled, and that includes soon-to-be-available retail units. That's encouraging. Google is not taking any chances with a repeat and acting with an abundance of caution. The left/right swipe action to raise and lower volume will still work tho. Otherwise use the wake phrase (currently OK Google) to start voice actions, which is the way most people would do it anyway. 
  • Reply 26 of 36
    Good code name...

    Mushrooms live in the dark on a steady diet of bullshit.
    GG1watto_cobra
  • Reply 27 of 36
    The first batch of Google Home Mini hardware distributed at assorted release events had a serious flaw that caused the device to listen in on every noise generated in a house --



    So, working as expected...


    After installation, the Google Home Mini provided to the author was turning itself on several times a minute, listened to what was going on in the background or on the television, and attempted to respond -- generally saying that it didn't understand.

    Not working as expected. It is supposed to be silent!
    watto_cobra
  • Reply 28 of 36
    chasmchasm Posts: 784member
    I'm happy to accept that this was a genuine hardware error and has been corrected. "Keyword listening" devices do indeed need to be "listening" all the time on a sampling level only (no uploading) to hear the start of the key phrase (be it "hey Siri" or "Ok Google"). Apple has made it abundantly clear that they don't upload things that don't need uploading, and have detailed how their listening works. Google uploads more than Apple, because it is checking for more keywords (this is their business model folks, don't act surprised). Amazon has stayed *deathly silent* on how much uploading is going on. Make of that what you will. The last sentence in this article is a bit nonsensical: no duh that a $49 Google Home Mini or Echo Dot isn't designed to compete with a HomePod. That's like saying the Dash Button and the HomePod are competing with each other (the GHM and the ED are really just slightly more sophisticated versions of the Dash button anyway). Sorry for the pun, but really: try comparing things that are roughly oranges to other things that are roughly oranges next time. That, I'd be interested in reading about.
  • Reply 29 of 36
    gatorguygatorguy Posts: 19,061member
    chasm said:
    I'm happy to accept that this was a genuine hardware error and has been corrected. "Keyword listening" devices do indeed need to be "listening" all the time on a sampling level only (no uploading) to hear the start of the key phrase (be it "hey Siri" or "Ok Google"). Apple has made it abundantly clear that they don't upload things that don't need uploading, and have detailed how their listening works. Google uploads more than Apple, because it is checking for more keywords (this is their business model folks, don't act surprised). 

    -Your iPhone listens for "Hey Siri" before making a server connection and uploading your voice request for processing. I think that's the only key-phase that works.

    - Google Android phones listen for "OK Google" before making a server connection and uploading your voice request for processing. I think that's the only key-phrase that works.

     So what are these other keywords you thought a Google Android smartphone is listening for to initiate voice actions? Perhaps I've missed some new feature announcement.
    edited October 2017
  • Reply 30 of 36
    I have multiple Echo devices in my house because I like the functionality it offers.  However the Tap I have as the main contact point is a piece of junk and while I trust (for some reason I'm not able to explain) Amazon more than Google, I don't want either really listening to me all the time. 

    Which is why Apple needs to hurry up and sell the HomePod.  Then I hope they sell some cheaper slave units to the HomePod that can be used through the house like I use dots.  I trust Apple to not be listening and selling what I say or do like Amazon and google certainly will.  I don't have Nest cameras or thermostat for this very reason.  I have a smart Honeywell thermostat and other closed system web video cameras that while aren't as cool as the Google offerings I trust a lot more.  


    Zero chance I would put a google listening device in my house. This is a perfect example why.  They patch was probably "oh crap, the light came on so they know we are listening.  let's patch that".
    watto_cobra
  • Reply 31 of 36
    gatorguy said:
    gatorguy said:
    Rayz2016 said:
    With the feature enabled, Google sends not only what the user says when a keyword is spoken or a device is activated -- but the sound recorded a few seconds before the actuation

    How can it record sounds before it was activated?
    Another example of the "always listening but not sending to a server" is the song recognition feature in the new Pixels. Yes that on-board microphone is always "listening" but without a connection to Google servers. All the recognition in done directly on the device with no need for any data sent to Google. But if the user then wants to get more information, perhaps read up on the artist, see what else they've got, maybe start a radio station using that song as the basis they can actively request a server connection to accomplish those actions. 
    How does the Pixel analyze heard media without talking to a server of audio fingerprints for matching? There’s a local copy of these fingerprints on every Pixel? 
    Yes there is. Approx. 20 thousand of them which will get updated as new music comes out. All song recognition is done on-device with zero data sent to Google.
    Sounds like data space that will continue to grow and take up more space the longer you own the phone.  Good thing the Pixel 2 has swapable storage...oops.  


    watto_cobra
  • Reply 32 of 36
    gatorguygatorguy Posts: 19,061member
    airnerd said:
    gatorguy said:
    gatorguy said:
    Rayz2016 said:
    With the feature enabled, Google sends not only what the user says when a keyword is spoken or a device is activated -- but the sound recorded a few seconds before the actuation

    How can it record sounds before it was activated?
    Another example of the "always listening but not sending to a server" is the song recognition feature in the new Pixels. Yes that on-board microphone is always "listening" but without a connection to Google servers. All the recognition in done directly on the device with no need for any data sent to Google. But if the user then wants to get more information, perhaps read up on the artist, see what else they've got, maybe start a radio station using that song as the basis they can actively request a server connection to accomplish those actions. 
    How does the Pixel analyze heard media without talking to a server of audio fingerprints for matching? There’s a local copy of these fingerprints on every Pixel? 
    Yes there is. Approx. 20 thousand of them which will get updated as new music comes out. All song recognition is done on-device with zero data sent to Google.
    Sounds like data space that will continue to grow and take up more space the longer you own the phone.  Good thing the Pixel 2 has swapable storage...oops.  


    How much space do you believe those song fingerprints use on the device? As for swappable storage if you are one of those that requires it in your smartphone don't buy either a Pixel2 or an iPhone. There are models that offer it but those two don't.
    edited October 2017
  • Reply 33 of 36
    airnerdairnerd Posts: 549member
    gatorguy said:
    airnerd said:
    gatorguy said:
    gatorguy said:
    Rayz2016 said:
    With the feature enabled, Google sends not only what the user says when a keyword is spoken or a device is activated -- but the sound recorded a few seconds before the actuation

    How can it record sounds before it was activated?
    Another example of the "always listening but not sending to a server" is the song recognition feature in the new Pixels. Yes that on-board microphone is always "listening" but without a connection to Google servers. All the recognition in done directly on the device with no need for any data sent to Google. But if the user then wants to get more information, perhaps read up on the artist, see what else they've got, maybe start a radio station using that song as the basis they can actively request a server connection to accomplish those actions. 
    How does the Pixel analyze heard media without talking to a server of audio fingerprints for matching? There’s a local copy of these fingerprints on every Pixel? 
    Yes there is. Approx. 20 thousand of them which will get updated as new music comes out. All song recognition is done on-device with zero data sent to Google.
    Sounds like data space that will continue to grow and take up more space the longer you own the phone.  Good thing the Pixel 2 has swapable storage...oops.  


    How much space do you believe those song fingerprints use on the device? As for swappable storage if you are one of those that requires it in your smartphone don't buy either a Pixel2 or an iPhone. There are models that offer it but those two don't.
    Space is space, and if you don't control it then that's not a good thing.  Sure all OS's will take up space but I don't use Shazam and wouldn't be too thrilled with space being consumed for a gimmick that I don't use.  
    And I am not wanting expandable storage and that wasn't my point in any shape, form, or fashion.  In no way did I claim to want that, so let's avoid tangents.  
    watto_cobra
  • Reply 34 of 36
    gatorguygatorguy Posts: 19,061member
    airnerd said:
    gatorguy said:
    airnerd said:
    gatorguy said:
    gatorguy said:
    Rayz2016 said:
    With the feature enabled, Google sends not only what the user says when a keyword is spoken or a device is activated -- but the sound recorded a few seconds before the actuation

    How can it record sounds before it was activated?
    Another example of the "always listening but not sending to a server" is the song recognition feature in the new Pixels. Yes that on-board microphone is always "listening" but without a connection to Google servers. All the recognition in done directly on the device with no need for any data sent to Google. But if the user then wants to get more information, perhaps read up on the artist, see what else they've got, maybe start a radio station using that song as the basis they can actively request a server connection to accomplish those actions. 
    How does the Pixel analyze heard media without talking to a server of audio fingerprints for matching? There’s a local copy of these fingerprints on every Pixel? 
    Yes there is. Approx. 20 thousand of them which will get updated as new music comes out. All song recognition is done on-device with zero data sent to Google.
    Sounds like data space that will continue to grow and take up more space the longer you own the phone.  Good thing the Pixel 2 has swapable storage...oops.  


    How much space do you believe those song fingerprints use on the device? As for swappable storage if you are one of those that requires it in your smartphone don't buy either a Pixel2 or an iPhone. There are models that offer it but those two don't.
    Space is space, and if you don't control it then that's not a good thing.  Sure all OS's will take up space but I don't use Shazam and wouldn't be too thrilled with space being consumed for a gimmick that I don't use.  
    And I am not wanting expandable storage and that wasn't my point in any shape, form, or fashion.  In no way did I claim to want that, so let's avoid tangents.  
    Oh. It seemed to be a concern for you so was curious what exactly it was. Thanks for clarifying you had no concern. 
  • Reply 35 of 36
    adm1 said:
    If it were any other company I'd accept it was a mistake, but not google.

    yeah! I also think like you

    badmonkwatto_cobra
Sign In or Register to comment.