Voice command

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • philippe_44
    Senior Member
    • May 2008
    • 9243

    Voice command

    Does somebody have plans to integrate things like Echo or Cortana into the LMS system ? I have an Echo and I see easily how I could use it for some controls :-)
    LMS 8.2 on Odroid-C4 - SqueezeAMP!, 5xRadio, 5xBoom, 2xDuet, 1xTouch, 1xSB3. Sonos PLAY:3, PLAY:5, Marantz NR1603, Foobar2000, ShairPortW, 2xChromecast Audio, Chromecast v1 and v2, Squeezelite on Pi, Yamaha WX-010, AppleTV 4, Airport Express, GGMM E5, RivaArena 1 & 3
  • EricBergan
    Member
    • Mar 2006
    • 75

    #2
    Personally, I'm hoping Amazon (or Logitech, or some 3rd party) does an integration between Echo and Harmony remotes. Then you could use voice controls for the managing the whole AV system...

    Comment

    • Julf
      Senior Member
      • Dec 2010
      • 2567

      #3
      Originally posted by EricBergan
      Personally, I'm hoping Amazon (or Logitech, or some 3rd party) does an integration between Echo and Harmony remotes. Then you could use voice controls for the managing the whole AV system...
      "Turn down the volume. I said TURN DOWN THE VOLUME!!"
      "To try to judge the real from the false will always be hard. In this fast-growing art of 'high fidelity' the quackery will bear a solid gilt edge that will fool many people" - Paul W Klipsch, 1953

      Comment

      • epoch1970
        Senior Member
        • Apr 2008
        • 2280

        #4
        Originally posted by philippe_44
        Does somebody have plans to integrate things like Echo or Cortana into the LMS system ? I have an Echo and I see easily how I could use it for some controls :-)
        I praise your creativity.
        Your discussion on a 3D printed case for picore (other thread, earlier) made me think. I thought of a player/server that would include an iThing dock. Seemed too ambitious.
        However I went on fantasizing about a player/server that would include Bluetooth "beacon" tech (to get IPeng and others to automatically select the closest player). After a brief search I found this project.

        It appears Siri is in the realm of Homekit, as is iBeacon... Please don't take my word for it and check by yourself if you care. All of this is way over my head.

        On a personal note, I don't believe at all in voice command, but a lot in presence detection.

        Anyways. Always great to see new ideas and initiatives.
        2 SB 3 • 1 PCP 7 • Libratone Loop, Zipp, Zipp Mini • iPeng (iPhone + iPad) • LMS 8.1 (docker) with plugins: CD Player, WaveInput by bpa • Material Skin by Craig Drummond • IRBlaster by Gwendesign (Felix) • Smart Mix, Music Walk With Me, What Was That Tune? by Michael Herger • PowerSave by Jason Holtzapple • Song Info, Song Lyrics by Erland Isaksson • BBC Sounds by Stuart McLean • AirPlay Bridge by philippe_44 • Auto Dim Display, SaverSwitcher, ContextMenu by Peter Watkins.

        Comment

        • philippe_44
          Senior Member
          • May 2008
          • 9243

          #5
          Originally posted by Julf
          "Turn down the volume. I said TURN DOWN THE VOLUME!!"
          Yep ... I've been working with VR for the last 25 years and it drove me nuts, especially because every time, there was supposed to be *the* algorythm that will make it fully usable. But I have to admit that after using Echo for a while, it is surprisingly good even with a lot of background noise. I've not paid a lot of attention, but it seems that neural networks are now working thanks to the insane increase of CPU power. I used to work on that topic in mid 80's and although it was fun, computers at that time where clearly lacking performances
          LMS 8.2 on Odroid-C4 - SqueezeAMP!, 5xRadio, 5xBoom, 2xDuet, 1xTouch, 1xSB3. Sonos PLAY:3, PLAY:5, Marantz NR1603, Foobar2000, ShairPortW, 2xChromecast Audio, Chromecast v1 and v2, Squeezelite on Pi, Yamaha WX-010, AppleTV 4, Airport Express, GGMM E5, RivaArena 1 & 3

          Comment

          • pippin
            Senior Member
            • Oct 2007
            • 14809

            #6
            Oh, voice recognition actually works and it has indeed already worked pretty well since the late 90s.
            The biggest issue is that you need a huge semantic database to compare with if you want to analyze more complex phrases which used to be a problem for mobile solutions but now that they all just send some meta-analyzed voice (or in the case of echo even the raw recordings) back to some server farm and do all the work there it's obviously more feasible, especially since you no longer have to replicate that database to all devices.

            The problem with VR is that it annoys everyone around you if you use it and speech is also pretty slow for a lot of tasks, I think that's more of a limitation than technology.

            And then... call me old fashioned, but I don't want to have the NSA run their bots over every piece of sound generated in my home, recorded by an array mike that even detects the faintest whisper in the most remote corner...
            Last edited by pippin; 2015-08-10, 23:25.
            ---
            learn more about iPeng, the iPhone and iPad remote for the Squeezebox and
            Logitech UE Smart Radio as well as iPeng Party, the free Party-App,
            at penguinlovesmusic.com
            New: iPeng 9, the Universal App for iPhone, iPad and Apple Watch

            Comment

            • philippe_44
              Senior Member
              • May 2008
              • 9243

              #7
              Originally posted by pippin
              Oh, voice recognition actually works and it has indeed already worked pretty well since the late 90s.
              Multi-speakers, dealing with accent, noisy environment, text-to-recognition was still a big challenge till very recently (and still is a bit) and there is nothing that irriates people more than having to repeat the same thing a few times to then realize that something else was recognized.
              The biggest issue is that you need a huge semantic database to compare with if you want to analyze more complex phrases which used to be a problem for mobile solutions but now that they all just send some meta-analyzed voice (or in the case of echo even the raw recordings) back to some server farm and do all the work there it's obviously more feasible, especially since you no longer have to replicate that database to all devices.
              that's right ...
              The problem with VR is that it annoys everyone around you if you use it and speech is also pretty slow for a lot of tasks, I think that's more of a limitation than technology.
              A bit, but it is also very true if you have to speak loud & slow - if you can use normal tune, then the annoying factor goes away a lot. Using voice as a UI can be slow in some cases, but there are many many simple tasks where it is very convenient and where using any other sort of input system is slow & painful - a sort of sweet spot for VR
              And then... call me old fashioned, but I don't want to have the NSA run their bots over every piece of sound generated in my home, recorded by an array mike that even detects the faintest whisper in the most remote corner...
              I understand that although I'm afraid anything we do is already scanned, unfortunately. I've given up hope on that
              LMS 8.2 on Odroid-C4 - SqueezeAMP!, 5xRadio, 5xBoom, 2xDuet, 1xTouch, 1xSB3. Sonos PLAY:3, PLAY:5, Marantz NR1603, Foobar2000, ShairPortW, 2xChromecast Audio, Chromecast v1 and v2, Squeezelite on Pi, Yamaha WX-010, AppleTV 4, Airport Express, GGMM E5, RivaArena 1 & 3

              Comment

              • philippe_44
                Senior Member
                • May 2008
                • 9243

                #8
                Originally posted by epoch1970
                I praise your creativity.
                Your discussion on a 3D printed case for picore (other thread, earlier) made me think. I thought of a player/server that would include an iThing dock. Seemed too ambitious.
                However I went on fantasizing about a player/server that would include Bluetooth "beacon" tech (to get IPeng and others to automatically select the closest player). After a brief search I found this project.

                It appears Siri is in the realm of Homekit, as is iBeacon... Please don't take my word for it and check by yourself if you care. All of this is way over my head.

                On a personal note, I don't believe at all in voice command, but a lot in presence detection.

                Anyways. Always great to see new ideas and initiatives.
                I agree that presence detection is a really nice thing and I would also gladly welcome that on my LMS system. I'm not supporting the idea of full voice command, as pointed by pippin, it is be much less efficient than haptic technology in many cases, but there are simple commands for which it is much faster. I also think that a multi-input UI is the best way to go, taking advantage of the most efficient mode for each situation
                Last edited by philippe_44; 2015-08-11, 02:39.
                LMS 8.2 on Odroid-C4 - SqueezeAMP!, 5xRadio, 5xBoom, 2xDuet, 1xTouch, 1xSB3. Sonos PLAY:3, PLAY:5, Marantz NR1603, Foobar2000, ShairPortW, 2xChromecast Audio, Chromecast v1 and v2, Squeezelite on Pi, Yamaha WX-010, AppleTV 4, Airport Express, GGMM E5, RivaArena 1 & 3

                Comment

                • pippin
                  Senior Member
                  • Oct 2007
                  • 14809

                  #9
                  Originally posted by philippe_44
                  if you can use normal tune, then the annoying factor goes away a lot.
                  I still feel it's awkward to start talking to a machine to control it when you are not alone. It can disturb others or raise "did you talk to me?" kind of confusion etc. I feel it's primarily for situations where you are alone.
                  Here in Berlin's tech community I'm recently seeing quite a number of people using Siri on their Apple Watch in public and I can't help it, I'm annoyed by someone adding a reminder next to me. It's a bit like talking on the phone in public, a pest that fortunately got quite significantly reduced in recent years due to the fact that people now tend to use messaging instead...

                  there are many many simple tasks where it is very convenient and where using any other sort of input system is slow & painful - a sort of sweet spot for VR
                  Actually I believe it's primarily _complex_ tasks where it shines. Simple tasks like changing volume are much easier done by a physical controller, especially given latency and need to repeat the command.
                  But something more complex like a search which requires several steps when done through a haptic UI can get much faster with voice because you can phrase more complex commands in a sentence than a simplified UX will usually directly allow.

                  I understand that although I'm afraid anything we do is already scanned, unfortunately. I've given up hope on that
                  Oh, I'm pretty sure there are no microphones for mass surveillance in my home right now (never can be sure about _individual_ surveillance but that's a whole different topic).
                  And saying "I don't care" just makes it too easy for them. Companies like Apple and recently to a certain degree also Google are trying hard to fight back on the privacy front, let's not spoil all that by not caring about even more severe intrusions.
                  Phones and Watches at least only have near-field microphones, Echo OTOH...
                  ---
                  learn more about iPeng, the iPhone and iPad remote for the Squeezebox and
                  Logitech UE Smart Radio as well as iPeng Party, the free Party-App,
                  at penguinlovesmusic.com
                  New: iPeng 9, the Universal App for iPhone, iPad and Apple Watch

                  Comment

                  • philippe_44
                    Senior Member
                    • May 2008
                    • 9243

                    #10
                    Originally posted by pippin
                    Actually I believe it's primarily _complex_ tasks where it shines. Simple tasks like changing volume are much easier done by a physical controller, especially given latency and need to repeat the command.
                    But something more complex like a search which requires several steps when done through a haptic UI can get much faster with voice because you can phrase more complex commands in a sentence than a simplified UX will usually directly allow.
                    Fair point, but don't you think that it can be very efficient for simple tasks because these require you to find your input device and then use your hands for a simple tap where voice would afford the same result while mobilizing much less "resources" ? I feel the efficency of voice as a UI does not vary linearly against complexity of the tasks. It is very situational, though. If I know what I want to listen to, VR is great. If I don't know exactly, vision + touch helps to embrace a larger amount of information / make choice much for efficiently, no ?
                    LMS 8.2 on Odroid-C4 - SqueezeAMP!, 5xRadio, 5xBoom, 2xDuet, 1xTouch, 1xSB3. Sonos PLAY:3, PLAY:5, Marantz NR1603, Foobar2000, ShairPortW, 2xChromecast Audio, Chromecast v1 and v2, Squeezelite on Pi, Yamaha WX-010, AppleTV 4, Airport Express, GGMM E5, RivaArena 1 & 3

                    Comment

                    • pippin
                      Senior Member
                      • Oct 2007
                      • 14809

                      #11
                      Voice command

                      Yea, sure, if I accept to not have an input device for voice but to allow it to listen to my whole home, that's a case, too.
                      I know what you mean because it's actually the only reason I sometimes wear my Aplle Watch at home: to be able to quickly access basic player controls without having to grab for a device.

                      Still don't use Siri on it, for some reason it's always set to the wrong language. Actually that's indeed the biggest unsolved issue for me: detecting languages. I switch between English and German all the time when writing and my devices are usually all set to English but then if you e.g. Want to search for a navigation destination in Berlin it simply doesn't recognize them because they have German names.... Same the other way around for music: if I set the language to German but then want to search for an Album with an English name you first have to go figure how Siri feels that name might be pronounced in German....
                      ---
                      learn more about iPeng, the iPhone and iPad remote for the Squeezebox and
                      Logitech UE Smart Radio as well as iPeng Party, the free Party-App,
                      at penguinlovesmusic.com
                      New: iPeng 9, the Universal App for iPhone, iPad and Apple Watch

                      Comment

                      • philippe_44
                        Senior Member
                        • May 2008
                        • 9243

                        #12
                        Originally posted by pippin
                        Yea, sure, if I accept to not have an input device for voice but to allow it to listen to my whole home, that's a case, too.
                        I know what you mean because it's actually the only reason I sometimes wear my Aplle Watch at home: to be able to quickly access basic player controls without having to grab for a device.
                        Well, I've not bought yet an apple watch, but the main reason I might one day is to be able to use iPeng
                        Still don't use Siri on it, for some reason it's always set to the wrong language. Actually that's indeed the biggest unsolved issue for me: detecting languages. I switch between English and German all the time when writing and my devices are usually all set to English but then if you e.g. Want to search for a navigation destination in Berlin it simply doesn't recognize them because they have German names.... Same the other way around for music: if I set the language to German but then want to search for an Album with an English name you first have to go figure how Siri feels that name might be pronounced in German....
                        I know exactly what you mean, I have the same problem, with French obviously, and this is really irritating
                        Last edited by philippe_44; 2015-08-11, 04:38.
                        LMS 8.2 on Odroid-C4 - SqueezeAMP!, 5xRadio, 5xBoom, 2xDuet, 1xTouch, 1xSB3. Sonos PLAY:3, PLAY:5, Marantz NR1603, Foobar2000, ShairPortW, 2xChromecast Audio, Chromecast v1 and v2, Squeezelite on Pi, Yamaha WX-010, AppleTV 4, Airport Express, GGMM E5, RivaArena 1 & 3

                        Comment

                        • spedinfargo
                          Senior Member
                          • Feb 2008
                          • 174

                          #13
                          I started a thread a couple years ago and got yelled at by naysayers ;-)



                          The key that I think is important is to make it non-comprehensive. Make it SIMPLE:

                          It doesn't even have to be too complicated - I'm looking for commands
                          like: "Playlist Shane Favorites Random", "Skip Track", "Mute All",
                          "Sync All Players", "Play Artist Pink Floyd", "Play Album Dark Side of
                          the Moon", etc.

                          Sure there will be lots of easy ways to "trip it up", but being able to say something like "play favorite spa music" would even be helpful...

                          Comment

                          • pippin
                            Senior Member
                            • Oct 2007
                            • 14809

                            #14
                            "Sync all players" is a wonderful example for something that doesn't work. Plenty of unanswered questions in that command....
                            ---
                            learn more about iPeng, the iPhone and iPad remote for the Squeezebox and
                            Logitech UE Smart Radio as well as iPeng Party, the free Party-App,
                            at penguinlovesmusic.com
                            New: iPeng 9, the Universal App for iPhone, iPad and Apple Watch

                            Comment

                            • philippe_44
                              Senior Member
                              • May 2008
                              • 9243

                              #15
                              So I started to look a tiny bit at that. Seems that you need to add an Alexa Skill https://developer.amazon.com/public/...-started-guide. Some have done that already for Sonos https://www.youtube.com/watch?v=AlOBJJyulI4. Interesting, I might try my luck later with this although it seems like a lot where I would have preferred more local processing of the request. I'll look at how hue/wemo are integrated, it seems that it is a simple UPnP discovery request that then, when answered properly, adds some basic functions where Echo simply invoke the returned URL's when preset actions are requested later
                              Last edited by philippe_44; 2015-08-18, 05:32.
                              LMS 8.2 on Odroid-C4 - SqueezeAMP!, 5xRadio, 5xBoom, 2xDuet, 1xTouch, 1xSB3. Sonos PLAY:3, PLAY:5, Marantz NR1603, Foobar2000, ShairPortW, 2xChromecast Audio, Chromecast v1 and v2, Squeezelite on Pi, Yamaha WX-010, AppleTV 4, Airport Express, GGMM E5, RivaArena 1 & 3

                              Comment

                              Working...