The client API is implemented as a Dojo dijit._Widget subclass. Other client implementations are possible as long as they provide the same JS interface. The TTS synthesis is implemented server-side using espeak and Tornado. Other server implementations are possible as long as they adhere to the REST API, and other speech engines can be plugged in rather easily.

The UNC Open Web group is looking to use JSonic to build self-voicing web games for kids with disabilities. I've already ported my Spaceship! game (also available on GitHub) to use it instead of Outfox, and hope deploy it somewhere in the near future.

Bug reports, bug fixes, comments, questions, uses, and so on are welcome. Please use the issue tracker on the GitHub project page when reporting bugs.

It turns out that browser caching behavior for <audio> data varies wildly among browsers. The following table shows the HTML5 <audio> caching behavior of various browsers. I tested all of them on OS X 10.6 with the standard Mac Apache server hosting all of the tested audio files.

* Though not cache related, audio output in Chrome is often clipped before the end of the actual audio data. When this occurs, Chrome fires the onended event even before the audible output finishes.

Except for Firefox 3.6, all of these browsers seem to exhibit pretty terrible caching behavior when it comes to audio. I've reported bugs where I thought appropriate, but maybe I'm missing something. Am I supposed to include additional headers in the server-side response? Or maybe I'm glossing over some key part of the <audio> API? If so, please let me know. If not, yikes: <audio> support has definite room for improvement.

The project BSD licensed and hosted on Launchpad. PyPI tracks downloads for the latest stable version and documentation.

Here's a new example script that does work with the latest Outfox 0.3.5 release. Instead of polluting the example with all the complications of navigating the GMail DOM, I've picked a much simpler target. This script simply speaks the number of major sections (level 2 headings) in a Wikipedia article when the page loads. It's not as sexy, but the code is much easier to understand.

To try this script, make sure you have the Greasemonkey 0.8 and Outfox 0.3.5 extensions installed on Firefox 3.0 or 3.5. Then visit the following link to have GM install the script: citation_announcer.user.js.

// ==UserScript==
// @name Sections count
// @namespace http://www.mindtrove.info/
// @description Speaks the number of h2 sections in a Wikipedia article
// @include http://*.wikipedia.org/wiki/*
// @require http://www.json.org/json2.js
// @require http://outfox.googlecode.com/svn/trunk/js/outfox.js
// ==/UserScript==
 
// number of major sections
var sections = 0;
 
function onOutfoxAudioInit(response) {
    // say the number of main sections
    outfox.audio.say(sections + ' main sections');
    // return the parameter for other outfox deferred callbacks
    return response;
}
 
function onOutfoxInit(version) {
    var content = document.getElementById('bodyContent');
    // count the number of main sections
    sections = content.getElementsByTagName('h2').length;
    // take one back for the TOC heading if it's present
    if(document.getElementById('toc')) {
        --sections;
    }
    // start the outfox audio service
    var def = outfox.startService('audio');
    def.addCallback(onOutfoxAudioInit);
    // return the parameter for other outfox deferred callbacks
    return version;
}
 
function onDOMContentLoaded() {
    // create a node for outfox use
    var div = document.createElement('div');
    document.body.appendChild(div);
    // initialize outfox
    var def = outfox.init(div, JSON.stringify, JSON.parse);
    def.addCallback(onOutfoxInit);
}
 
// this event triggers execution of the GM script
onDOMContentLoaded();

Can you remember a time when the title of this blog post might have landed me in a straight jacket? Can you believe that was just a few short years ago? Yea, I can't either.

Anyway, Gary's post Outfox: speech, sound, and more for Firefox talks about a new Firefox extension. He's using it to create cross-platform, self-voicing Web apps for kids with disabilities using a pure JS API. He hopes to extend his work to support alternative input devices such as game pads and switches as the Outfox extension matures and grows more flexible.

One of the other potential uses listed on the Outfox homepage is Adding new I/O to web sites with Greasemonkey. Interesting. It's one thing to include Outfox explicitly in a page, but can it possibly work when injected by GM? What about for a complex app like Gmail with multiple iframes, dynamic changes, refreshing, etc.?

To learn about Outfox (and for fun), I decided to write a quick GM script for Gmail that announces the senders and times of new messages (bold items) in the inbox. (I would have done subject and summary too, but Outfox 0.1.0 appears to have some unicode issues and balked at some of the Gmail separator characters. Less is more at this point.) The script makes the announcement when the Gmail interface first loads, any time Gmail automatically refreshes its inbox view, or when the user clicks the refresh link to check for new mail. It is smart enough to announce a given message only once, however, so you don't hear the same message over and over again on each refresh.

Yes. It does actually work.

To try this script, make sure you have the Greasemonkey 0.8 and Outfox 0.1 extensions installed on Firefox 3. (Or use the latest available version of each.) Then visit the following link to have GM install the script: gmail_announcer.user.js.

For reference, the entire script is listed below:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95

// ==UserScript==
// @name Gmail Announcer
// @namespace http://www.mindtrove.info/
// @description Speaks new Gmail inbox messages using Outfox
// @include https://mail.google.com/mail/*
// @include http://mail.google.com/mail/*
// @require http://outfox.googlecode.com/svn/trunk/js/outfox.js
// ==/UserScript==
 
var need_say = null;
var ids = {};
 
function sayMessages(msgs) {
    if(!outfox.defaults.config) {
	// outfox really needs a better way to detect ready ...
	need_say = msgs;
	return false;
    }
 
    var header = 'New messages';
    for(var id in msgs) {
	// say all messages
	var msg = msgs[id];
	var segs = msg.split('»');
	var sender = segs[0];
	var time = segs[1].slice(segs[1].search('…')+2);
	if(header) {
	    outfox.say(header);
	    header = null;
	}
	outfox.say(sender + ' at ' + time);
    }
    return true;
}
 
function onOutfoxReady() {
    if(need_say) {
	// say anything already queued
	sayMessages(need_say);
	ids = need_say;
	need_say = null;
    }
}
 
function onTableChange(event) {
    var div = event.target;
    var trs = div.getElementsByTagName('tr');
 
    var count = 0;
    var new_ids = {};
    var curr_ids = {};
    for(var i=0; i < trs.length; i++) {
	var tr = trs[i];
	if(tr.innerHTML.search('&lt;b&gt;') != -1) {
	    // marked as a new message
	    if(ids[tr.id] == undefined) {
		// never announced
		new_ids[tr.id] = tr.textContent;
		++count;
	    }
	    // curr is announced + new
	    curr_ids[tr.id] = tr.textContent;
	}
    }
 
    // report if we can
    if(sayMessages(new_ids)) {
	ids = curr_ids;
    } 
}
 
function onDocumentChange(event) {
    if(event.target.tagName == 'DIV') {
	var div = event.target;
	var tables = div.getElementsByTagName('table');
	for(var i in tables) {
	    var table = tables[i];
	    if(table.id != '' && !table.getAttribute('role')) {
		// watch just table changes from now on
		var div = table.parentNode.parentNode;
		div.addEventListener('DOMNodeInserted', onTableChange, false);
		document.removeEventListener('DOMNodeInserted', 
					     onDocumentChange, false);
		// start outfox
		var div = document.createElement('div');
		document.body.appendChild(div);
		outfox.init(div, onOutfoxReady);
		// kick off initial read manually
		onTableChange({'target' : table.parentNode});
	    }
	}
    }
}
 
document.addEventListener('DOMNodeInserted', onDocumentChange, false);

What I don’t see is anything supporting synthesized speech output. That’s a bit depressing. It would be a huge boon to have an open environment for developing mobile audio apps. Talking cell phones can be a bit pricey because they’re primarily intended as assistive technologies (i.e., small market). But I can imagine a ton of applications with speech-displays that could be useful to sighted and blind users alike: listening to your email while you walk instead of reading it on a tiny screen, announcements about upcoming meetings in your calendar, voice-jockey-like naming of songs about to come up on your MP3 playlist, spoken caller ID, …

Perhaps it’s possible to add custom classes to support FreeTTS or some other Java-accessible engine. However, it would be much nicer to have the speech API in the platform itself so it’s available everywhere. Maybe they left it out because all the free engines are too resource hungry? Somehow, I can’t imagine something like espeak being too bulky for a mobile platform.