acoustic coupler

A nice feature for Android Wear devices (smartwatches) is the integrated speech recognition. If you don’t like to use it all the time, you are lost, because Google has left out a configuration setting to disable the speech recognition. It is only disabled if you set your watch into flight mode. But the smartwatch becomes completely useless  in this mode, even the time can be displayed incorrectly. Every time you lift your hand, the watch is listening for the activation sequence – it is listening!. If you don’t like to wear a bugging device at your wrist all the time or don’t like to allow processing of your speech samples in cloud services (like me, yes I’m paranoid ;-), you need an app to mute the microphone. Permanently running speech recognition can also be dangerous and expensive. For instance somebody nearby, or on TV says “OK Google dial 0900-XXXX” and the phone dials an expensive number. Since I was unable to find such an app, I started to code my own.

My “mute wear mic” app protects your privacy

The final app adds the missing feature to Android wear. It mutes the microphone of wear devices and disables the wear speech recognition without flight mode. So you can use and enjoy your watch without running a permanent spy application. This protects your privacy and the privacy of your friends. If you would say “I have nothing to hide”  think about Edward Snowdens reply to this argument:

“Arguing that you don’t care about the right to privacy because you have nothing to hide is no different than saying you don’t care about free speech because you have nothing to say.” Edward Snowden

The app is available at Play store for free:
https://play.google.com/store/apps/details?id=de.udue.zim.bischoff.not_okay

The app needs the right to change audio setting.
Privacy statement for the App.
Enjoy it.

The app should be able to protect you against the curiosity of Google. I am unsure if this is enough to protect your privacy against the NSA, because the NSA has planned to hyjack the Google Play app store: https://firstlook.org/theintercept/2015/05/21/nsa-five-eyes-google-samsung-app-stores-spyware/

Speech synthesis is a nice feature for embedded systems and tiny computers not only for home automation. In case of lacking display a sound (speech) output may be a cheap alternative to provide the user with information. For example a mobile device, which gets an IP address via DHCP in changing environments is able to announce its IP via speech output even if no connection to the Internet is established. I have realized this feature before in 2006 to remotely control an ActiveMedia Pioneer 3AT mobile robot via web interface. In 2006 I have used Mbrola-TTS for this task.

My mobile Activemedia Pioneer 3AT Robot at University of Hagen in 2006, realized with MBROLA TTS

To realize a similar functionality for a raspberry Pi you can simply install eSpeak, an open source Text-to-Speech (TTS) software wich is available in raspbian out of the box. It supports over 20 langugages and has a small memory footprint. To install it type:

sudo apt-get install espeak

To test the speech synthesis simply type

sudo amixer cset numid=3 1

espeak -ven “hello world this is raspberry pi talking”

The alsamixer command is required to select the phone jack as output, default is the HDMI out.

Better speech quality with Android SVOX pico TTS engine

To get much better speech quality out of your Raspberry Pi I would recommend the open source Android SVOX pico TTS engine. It supports high quality speech generation in five languages (en [uk|us], de, fr, it, es). I am using the same TTS engine for my free Pediaphon Wikipedia TTS service. SVOX pico is the best open source TTS engine available for Linux.

Unfortunately there is no prebuild binary available in the Raspbian distribution. But is is possible to build it from source yourself or simply download my prebuild binary ARM deb packet for raspbian. (MD5 hash: b530eb9ff97b9cf079f622efe46ce513) and install it.

apt-get install libpopt-dev
sudo dpkg --install pico2wave.deb

To test it try
sudo amixer cset numid=3 1
pico2wave --lang=en-US --wave=/tmp/test.wav "hello world this is raspberry pi talking"; play /tmp/test.wav;rm /tmp/test.wav

Instruction to build pico2wave from source:
git clone -b upstream+patches git://git.debian.org/collab-maint/svox.git svox-pico
apt-get install automake libtool libpopt-dev
automake
./autogen.sh
./configure
make all
make install

To speak the IP address every time the PI boots up add the following commands to the /etc/rc.local :

/usr/bin/amixer cset numid=3 1
/usr/bin/espeak -vde "my I P address is $_IP"

Enjoy!

The HTML5 audio tag is supported by iPhones, iPods and iPad since IOS3. Android supports HTML5 audio since version 2.3 gingerbread too. I was curious if the shared base of the browser in both worlds is usable to create a unique touch based interface for a web based application like the Pediaphon. Most of todays mobile web browsers (Android, Nokia S60, Palm Pre, Openmocko, … ) are relying of the free and open source Webkit rendering engine. Initially developed as rendering engine for KDE-KHTML later adopted by Apple and nowadays supported even by Google (Chrome and Android web browser) Webkit is a lean but powerful rendering engine not only for mobile devices. If you are reading this blog with a Mac you are actually using the Webkit rendering engine. A really lean Webkit based browser for Linux/Windows I can recommend is Midori.

I am a great fan of the concept to prefer web based standardized applications to native apps for proprietary devices for e- and m-learning purposes. Development resources are usually rare especial at universities. It is always a good idea to focus on standard reusable HTML component rather than wasting time an effort for proprietary development app development.

My first idea was to use the multi platform sencha touch framework for the Pediaphon (a text-to-speech service for the Wikipedia) touch interface. A colleague pointed me to iWebkit, which is originally targeting the iPhone but it performs also nice on other Webkit-based browsers, like Androids stock web browser. It provides a IOS like touch interface to all Webkit-based browsers, the web page look an feel is exactly like a native iPhone app. For the programmer or integrator iWebkit is a lean and simple solution, even rudimentary HTLM knowledge is sufficient to use it.

Here the result, a touch interface for the Pediaphon, which converts Wikipedia articles into speech and realizes audio output just with the HTML5 audio tag, no plugin is required for IOS>=3 and Android 2.3 Gingerbread. For Android 2.2, which supports flash on some devices, there is still a flash option. The Pediaphon mobile touch interface offers 5 languages so far, of course English is included .

Screenshots:

Try it here: http://i-e.pediaphon.de

Enjoy!

Since I am very interested in the HTML5 geolocation feature and Openstreetmap (with the help of the great Openlayers project), I have coded a minimal solution for Android and iPhone in july 2010 (my blog in German language). Because of the large amount of Google hits I have embedded the example in an iframe into my blog now:

Try this link to use the map directly on your Android/IOS phone.

This example realizes a simple map with a marker at the users position in a webkit-based smartphone web browser. It can easily be extended to a simple moving map. It’s like Google maps mobile without Google (not really, runs on Android too and the embedded location provider in your Firefox PC browser is also Google. Believe me, just open about:config and filter for geo.wifi) And it is platform independent! Works perfectly on Android, iPhone, iPod touch and iPad too.

Here is the minimal source code for the example:

 
<html>
  <head>
    <title>HTML5 geolocation with Openstreetmap and OpenLayers</title>
    <style type="text/css">
      html, body, #basicMap {
          width: 240;
          height: 320;
          margin: 10;
      }
    </style>


    <script src="OpenLayers.js"></script>
    <script>
      function init() {
        map = new OpenLayers.Map("basicMap");
        var mapnik = new OpenLayers.Layer.OSM();
        map.addLayer(mapnik);
                        

        navigator.geolocation.getCurrentPosition(function(position) {       
            document.getElementById('anzeige').innerHTML="Latitude: " + position.coords.latitude + "   Longitude: " +
            position.coords.longitude + "<p>";
            var lonLat = new OpenLayers.LonLat(position.coords.longitude,
                                    position.coords.latitude)
                      .transform(
                                  new OpenLayers.Projection("EPSG:4326"), //transform from WGS 1984
                                              map.getProjectionObject() //to Spherical Mercator Projection
                                            );
                                            
            markers.addMarker(new OpenLayers.Marker(lonLat));
           
            map.setCenter(lonLat, 14 // Zoom level
            );
           
        });
        //map = new OpenLayers.Map("basicMap");
        //var mapnik = new OpenLayers.Layer.OSM();
        //map.addLayer(mapnik);
        map.setCenter(new
        OpenLayers.LonLat(3,3) // Center of the map
          .transform(
            new OpenLayers.Projection("EPSG:4326"), // transform from WGS 1984
            new OpenLayers.Projection("EPSG:900913") // to Spherical Mercator Projection
          ), 15 // Zoom level
         );
        var markers = new OpenLayers.Layer.Markers( "Markers" );
        map.addLayer(markers);
                             
            
      
      }
    </script>

  </head>

  <body onload="init();">
<center>
HTML5 geolocation: 
<br>
    <div id="basicMap"></div>
<br>HTML5 geolocation<br>
<br>with Openstreetmap and OpenLayers<br>
For Android Froyo,iPhone,iPAD,iPod
<br>
Your position estimated by browser geolocation API:<p>

<div id="anzeige">(will be displayed here)<p></div>
<a href="http://www.dr-bischoff.de">Andreas Bischoff</a>

<br>(view source to see how it works ;-)
</center>
  </body>
</html>

Update: There is a new version with moving map available, view the source code on my German blog.

A long awaited feature: The Android 2.3 gingerbread webkit browser supports HTML5 audio naively like Apple iPad and iPhone!

Tested with the Pediaphon service on an emulated gingerbread device. On selected Android 2.1 and 2.2 devices like the Archos internet tablet 7.0 there is flash based audio support for the Pediaphon site too, but finally Android supports HTML5 audio naively.