What Software Speech-enabled Linux Distributions Work Today?
amazability at rcn.com
Thu Jun 12 16:44:14 UTC 2008
Amazability, Inc. www.amazability.com is preparing to release its Adept1
product which uses voice recognition for command and control and AT&T's
Natural Voices with Fedora Linux. For those who might be interested there
was a Tech Talk given by Ken Ingham at Google, Inc, in California which
features the Adept1. The URL is:
It will give you some idea of what Amazability, Inc. is trying to do.
From: blinux-list-bounces at redhat.com [mailto:blinux-list-bounces at redhat.com]
On Behalf Of Martin McCormick
Sent: Thursday, June 12, 2008 10:57 AM
To: Linux for blind general discussion
Subject: What Software Speech-enabled Linux Distributions Work Today?
For years, I used UNIX via RS-232 or even sometimes telnet from
a DOS P.C. using Kermit or telnet and an external speech box. I
am still doing that at home, but the old hardware is making
decisions for me by dying and it has been time to move along for
I have been using Debian Linux since about 2001 and my
dream installation would be Debian with software speech that
If one has more than one computer available, the one
that talks can get in to ones that don't if they have a serial
port or are far enough in the conscious world to operate on a
So, I am ideally looking for a non-orphaned Debian
distribution that uses a non-commercial speech engine that
Knowing that we rarely get everything we want in the
real world, I am saying all of this sort of tongue in cheek.
What we usually use is something that is as close as practical.
Somebody said, "Don't let perfect kill good enough."
When I tried GRML, it first appeared to be adequate
until that spelling bug hit. No software synth was going to work
Oralux has mostly worked well but it is orphaned.
I am not set in concrete on anything but I would like to
stick with Debian if possible for historical reasons although
there is nothing really wrong with the other distributions.
What is working now? In the case of this laptop, there
is no native serial port so any installation must either talk or
work through a network.
Some general observations:
Some of the best speech right now is on the new
Macintoshes but they also have a terrible flaw right now.
Speakup is the closest I've seen to getting this thing right,
but what Macs do is reset their speech when new input comes in.
You can't listen to, say, the output of a compiler and get
anything useful because new lines kill what you are already
In my job, we frequently use tail on system logs to
look for output in real time and the Mac just stutters as new
input arrives. You can play back the buffer, but that's no
substitute for listening until you've heard enough.
Speakup only resets when one presses keys so you can
listen as long as you want and then kill the input. This is the
kind of utility I want to continue.
I have been experimenting with computers for 29 years
now and UNIX is without a doubt closer to perfection when it
comes to access than we've ever been before.
Martin McCormick WB5AGZ Stillwater, OK
OSU Information Technology Department Network Operations Group
Blinux-list mailing list
Blinux-list at redhat.com
More information about the Blinux-list