<!DOCTYPE html>
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
</head>
<body>
<p>Hi all,</p>
<p>I want to first thank Josh for sharing on-list several days ago
about the Piper voices for NVDA add-on. The purpose of this
message is to share my initial impressions of Piper with members
of this email list. Of course since every message sent to this
list if automatically archived and publicly readable, I am also
sharing these impressions with the entire Internet after the send
button is pressed.</p>
<p><br>
</p>
<p>I have downloaded the Piper voice NVDA add-on from GITHUB and
tried some of the voice packages available after choosing and
installing them. Overall I am impressed with quality of some of
the AI-powered voices, especially since Piper is currently in
Beta. The Piper voice manager which the developer added recently
makes it very easy to find, preview and download voices.</p>
<p><br>
</p>
<p>For anyone who has not tried Piper voices yet and interested in
doing so in NVDA: After installing the Piper add-on, I suggest
downloading the Arctic voice package through Piper voice manager.
In my experience, the other English-speaking voice packages
available don't currently speak letters or words correctly when I
navigate by individual letter or word. For whatever reason, the
Arctic voices don't have that extremely annoying problem most of
the time. My favorite Arctic voice is BDL. The Piper version of
that voice is, to my ears, more realistic/human-sounding compared
to the non-AI version available through<br>
</p>
<p><a href="https://rhvoice.org/">RHVoice</a>.<br>
</p>
<br>
<p>I have bookmarked the Piper GitHub page for future reference. I
am excited about trying additional Piper neural voices and
versions in the future as the TTS project evolves!</p>
<p><br>
</p>
<p><br>
</p>
<p>Blake<br>
</p>
</body>
</html>