Excellent clear British SAPI5 TTS voices for short animations and tutorials are Ivona Voice 1.6 Amy 22kHz (aka Ivona 2 Amy). The other is the less common Neospeech Voiceware Bridget (demo). Yes, they’re still robo voices, but I’d say they’re the most pleasing and human of the female British voices released in the 2010s for SAPI5. And you can’t argue with free and offline. No subscriptions or data-gouging needed (though Eleven Labs are the best, if you want that). Both voices are now free on Archive.org, and work with the free Balabolka. Here’s how to control them with some simple markup.
1. Amy can be controlled using the normal SAPI markup (SSML) in the Balabolka editor. She also supports phenomes (visimes) symbols, a prosidy markup tag, and others.
|
1 |
<phoneme alphabet="ipa" ph="ɺʡʜʑʢ"/> |
|
1 |
<prosody rate="-20%" volume="40">prosidy</prosody> |
|
1 2 |
You say, <sub alias="tu-mah-toe">tomato</sub>. I say, <sub alias="to-may-toe">tomato</sub>. |
2. Neospeech Voiceware Bridget however does not support phenomes, prosidy etc and requires the following XML tags to add pauses and emphasis…
|
1 2 3 |
Here we insert a very slight pause. Here <silence msec="01"/> we insert <silence msec="01"/> a <silence msec="03"/> <emph>very<emph> slight <silence msec="03"/> pause?! |
Add ! or ?! to inflect and emote a word at the end of a sentence…
|
1 |
gosh! |
|
1 |
gosh?! |
Also working as XML tags with Bridget…
|
1 |
<emph>emphasis</emph> |
|
1 |
<volume level="50">quieter voice</volume> |
|
1 |
<rate absspeed="-6">slower speed</rate> |
|
1 |
<pitch absmiddle="6">higher pitch</pitch> |
No closing tag is needed for the following milliseconds pause tag…
|
1 |
<silence msec="300"/> |
Incidentally, Neospeech Voiceware Bridget shows up in Balbolka as “VW Bridget English (United States)”, when she should be “VW Bridget English (British, upper class)”.