Voice4Me is one of the free programs available at the SchoolFreeware website. It's a very basic AAC program for users who have trouble speaking - users can click on a picture to speak the associated text, or type text into a box and then click on the 'speak' button.

There's a comprehensive introduction to this program available on the YouTube website, including configuration instructions:

[embed width="640" height="385"]http://www.youtube.com/watch?v=qcGindLKhMM[/embed]

The program is very limited in how much it can be configured - you must have both the picture and the text areas, and the number and size of the pictures can't be varied. The program also can't be controlled by keyboard alone if you want to access the pictures, so it's not usable with any combination of switches which is unfortunate. On the plus side it's free and once it's set up it's fairly easy to use. If you have no access to other AAC software this could be helpful.

Website: Voice4Me

- Ricky Buchanan

Did you know that you can change narrators automatically when you're using GhostReader? It's an undocumented feature, and only works with the voices from Infovox iVox or ConvenienceWare, but I've confirmed that it does work.

The biggest catch is that you can only switch between voices which are the same voice type and the same voice category. The voice type is either 'Infovox iVox' or 'ConvenienceWare', the ConvenienceWare voices can only be used within GhostReader whereas the iVox voices can be used with any programs. The voice categories are 'HD' and 'HQ' or 'HQM' (these two appear to be the same). To find out which voices are in which category, look on these pages:

• ConvenienceWare Voices lists the category and gender in brackets for each ConvenienceWare voice.
• Infovox iVox Voices does the same for the iVox voices.

We can see from this, for example, that the American English voices "Heather", "Laura", "Ryan", "Kenny", and "Nelly" are all HQ/HQM voices so we can swap between these.

To do the voice change, you put this command in your text:

\vce=speaker=newspeaker\

Instead of the word "newspeaker" you need to use the name of the voice you want, but the rest has to be exact - including the last backslash. So if you're using American English voices, try using this text:

\vce=speaker=heather\Are you coming along? \vce=speaker=ryan\Sure I am, I'd love to! \vce=speaker=laura\I will come along too.

It will start out with Heather, then switch to Ryan and then to Laura. If you have the British English voices, you could use this text instead instead:

\vce=speaker=Lucy\Are you coming along? \vce=speaker=Peter\Sure I am, I'd love to! \vce=speaker=Graham\I will come along too.

I only own the British English InfoVox iVox voices so I haven't tried to see if this works when switching between languages. If it does, it could be a really useful thing for when you have a document that changes from one language to another.

Remember, this is an undocumented command, and the speech synthesis system doesn't really "know" you've changed voices so it won't automatically change back to your default voice for more speech afterwards. For example, my default system voice is Peter, but after using the snippet above (which finishes with Graham speaking) any speech synthesis afterwards will still use the Graham voice. To fix this, add another switching command for your default system voice at the end of your document. So to switch back to my default Peter voice, the above text becomes this:

\vce=speaker=Lucy\Are you coming along? \vce=speaker=Peter\Sure I am, I'd love to! \vce=speaker=Graham\I will come along too.\vce=speaker=Peter\

It doesn't matter that there's no text after the last voice switch, it just tells the speech synth to switch back to Peter. So when my computer next announces the time, or I ask it to speak a block of text, it will be back to speaking in the Peter voice.

Can you think of any fun uses for this? Leave a comment!

- Ricky Buchanan

We've already seen that you can use OS X's built in text-to-speech just by selecting a key combination to use with it, so why pay money for Narrator? There is one feature that Narrator has that no other text-to-speech software has - easy in-line voice switching. You set a marker and it changes to the selected voice at that point while reading.

Use Narrator to read out a play or story with different voices for each of the parts. It uses speech synthesis to read out marked passages using specified voice attributes. You can choose different voices, rates, pitches, inflections, and volumes for each character in the story. The words are highlighted on-screen, and there are also a couple of silent read-along options for stage directions, or for you to read out your own parts.

Starting with version 2, you can have multiple chapters in each document to help organize the story, use a word replacement dictionary to fine-tune the pronunciation, and can export the story as an AAC sound file or export it to iTunes, so you can listen on an iPod or iPhone, just like your own audiobook! Plus many other appearance and functionality improvements.

In my testing I found that Narrator has some trouble with Infovox iVox voices - the highlighting of words as they are spoken breaks down if it reaches a section set to use an iVox voice and doesn't always get back in synchronisation afterwards, and the "pitch" and "inflection" sliders slide around but don't do anything. To my knowledge, none of the Infovox iVox voices can have their pitch or inflection altered, so the sliders shouldn't be available for these voices. These problems didn't happen with the Apple or Cepstral voices that I tried, but the Cepstral voices are generally of a lower quality than the Infovox iVox voices so I found it disappointing that these didn't work. Perhaps this will be fixed in a future version.

- Ricky Buchanan

Have you ever wanted to make phone calls on the net using a synthetic computer generated voice? If you're an AAC user, then this is a great way to communicate with people remotely. This article will show you how to set up Wiretap Anywhere to send your speech voice to Skype's internet phone system. Almost the same method will also work with other voice chat programs like iChat, Gizmo, and iSoftPhone.

You'll need the Wiretap Anywhere and Skype software to start with. The default service that comes with free Skype software only allows computer to computer connections so if you want to call cellphones and landlines you will have to fund an account with Skype. Skype calls these funds SkypeOut - you can see the SkypeOut panel by clicking on the "€" symbol on the top right hand side of the Skype window:

Here's what happens when you click on it:

Now you can visit the account page to put funds into your SkypeOut account if you want to make calls to regular phones.

Configuring the Wiretap Anywhere software to pick up your system audio is absurdly simple. Open the "System Preferences" panel from the Apple menu in the top-left corner of the screen:

Now click on the "Wiretap Anywhere" logo to open that preference pane:

You want to create a new device, so press the handy "plus" button at the lower left of the pane:

Your new device will be called "Untitled 1". You can rename it if you want - I've called mine "AAC Telephone" here:

You should immediately see audio levels on the Wiretap Anywhere indicators
showing any audio activity in your computer, such as iTunes playing or system/third party voices speaking.

Now go into Skype and open the Preferences via the "Skype" menu. Click on the speaker icon at the top of the pane - where the top red circle is in this picture - and then select your Wiretap device as the "Audio Input" device - that's the lower red circle:

Your first call should be to Skype's built in "test call" service to confirm that skype is receiving your synthetic voice audio. The Skype name to call is "echo123" and it gives you a chance to talk and then plays back to you what it heard.

That's all! You're now ready to make phone calls on the net using a synthetic computer generated voice. Congratulations!

- Ricky Buchanan and NewJack Rasputin

Cepstral have announced version 5 of their text-to-speech voices, but I have found little detail about the improvements. The following quote is from the website:

Version 5 is finally here, and it includes improvements to voice quality and engine efficiency.

This is a paid upgrade with the usual price of US$30 per voice reduced for any voice you already have a license for. When I entered the license information for the US "David" voice that I own I was informed that the upgrade price would be US$10 but no general upgrade prices have been announced, so other voices may cost different amounts to upgrade. If anybody has further information, please leave a comment.

Website: Cepstral Text-to-Speech

- Ricky Buchanan (still working on my health mostly, but couldn't resist posting this!)