Re: yasr and screen

Re: yasr and screen

[pj]

Greetings,

Kyle wrote:
> I did most things with a single text console that ran YASR
> automatically at login and did all my work in Screen,
> which allowed me to have a nearly unlimited number of "windows"
> open on a single console, all under a single YASR instance.

The combination of yasr and screen is a VERY good one,
because screen provides a really good cut-and-paste mechanism
which is the most important thing that yasr lacks.

> The trade-off is that you will get no speech prior to login,

If you are able to log in, you have dmesg and /var/log/messages,
and dmesg provides a record of things that happened long before
the speakup module gets loaded.

A big problem for installs is the BIOS;
that, I guess, needs a separate computer with camera and O C R ...
Perhaps the ideal solution would be to build speech into a
touchscreen monitor, and not into the operating system ?

My big problems with yasr are :

One: yasr can't use espeak except through emacspeak,
and I can't get emacspeak configured to use espeak properly
(it seems to use eflite commands even when talking to /usr/bin/espeak).
But my eflite chops off the last part of every word,
which makes a lot of things unrecognisable.

Two: I can't understand how keystrokes are specified in yasr.conf
for example, to me, 0x1b6c means Escape l, not alt-L

If I were rewriting yasr, I'd probably do it in perl or lua,
which would reduce the code-bulk by several times,
and make it more maintainable and portable.

Regards,  Peter Billam

P.S.  I am a long-time unix programmer, and still sighted,
though this is not improving, so I'm wanting to select a speaking interface
and learn it while it's still easy for me.

http://www.pjb.com.au      pj@pjb.com.au     (03) 6278 9410
"Follow the charge, not the particle."  --  Richard Feynman
 from The Theory of Positrons, Physical Review, 1949

Re: yasr and screen

[Trevor Saunders]

> A big problem for installs is the BIOS;
> that, I guess, needs a separate computer with camera and O C R ...
> Perhaps the ideal solution would be to build speech into a
> touchscreen monitor, and not into the operating system ?

or you can just unplug the hard drive and whatever so the first bootable
thing the bios finds will be your usb drive.  Then plug the drive back
in to do the install of course, linux usually handles hard drives
showing up after boot gracefully ime.

> My big problems with yasr are :
> 
> One: yasr can't use espeak except through emacspeak,
> and I can't get emacspeak configured to use espeak properly
> (it seems to use eflite commands even when talking to /usr/bin/espeak).
> But my eflite chops off the last part of every word,
> which makes a lot of things unrecognisable.

you can also have yasr talk to speech-dispatcher.  The config for that
is in the stock yasr.conf but commented out.

> Two: I can't understand how keystrokes are specified in yasr.conf
> for example, to me, 0x1b6c means Escape l, not alt-L

off hand I'm not sure how that works either.  Maybe you could use the
keyboard config thing that's built in (ctrl - alt - k by default)?

> If I were rewriting yasr, I'd probably do it in perl or lua,
> which would reduce the code-bulk by several times,
> and make it more maintainable and portable.

I think some of it will need to stay in C / C++ but allowing it to be
scripted with lua sounds interesting.

Trev

> 
> Regards,  Peter Billam
> 
> P.S.  I am a long-time unix programmer, and still sighted,
> though this is not improving, so I'm wanting to select a speaking interface
> and learn it while it's still easy for me.
> 
> http://www.pjb.com.au      pj@pjb.com.au     (03) 6278 9410
> "Follow the charge, not the particle."  --  Richard Feynman
>  from The Theory of Positrons, Physical Review, 1949
> _______________________________________________
> Speakup mailing list
> Speakup@linux-speakup.org
> http://linux-speakup.org/cgi-bin/mailman/listinfo/speakup

Re: yasr and screen

[Gregory Nowak]

On Sat, May 11, 2013 at 09:54:39PM +1000, pj@pjb.com.au wrote:
> A SpeakingVGA monitor would let you get work on the BIOS, though,
> and on things like RAID cards where you have to press Ctrl-A
> before boot to configure them or add and remove discs etc.

I agree that a "talking monitor" -- something that speaks output from a
vga/hdmi port would be really, really nice. I seem to remember a
segment on ACB radio's Main Menu a while back from one of the
technology conferences. In it, a guy from a company in Italy was
demonstrating a multi-purpose OCR device, and one of the things it had
was a vga input. That's all I remember though, and haven't heard
anything else about it since. Wonder if this device is still around?

Greg


-- 
web site: http://www.gregn.net
gpg public key: http://www.gregn.net/pubkey.asc
skype: gregn1
(authorization required, add me to your contacts list first)

--
Free domains: http://www.eu.org/ or mail dns-manager@EU.org

Re: yasr and screen

[pj]

Hi :-)

I wrote:
> A big problem for installs is the BIOS;
> that, I guess, needs a separate computer with camera and O C R ...
> Perhaps the ideal solution would be to build speech into a
> touchscreen monitor, and not into the operating system ?

Trevor Saunders wrote:
> or you can just unplug the hard drive and whatever so the first
> bootable thing the bios finds will be your usb drive.  Then plug
> the drive back in to do the install of course, linux usually
> handles hard drives showing up after boot gracefully time.

Yes, nice trick :-)

A SpeakingVGA monitor would let you get work on the BIOS, though,
and on things like RAID cards where you have to press Ctrl-A
before boot to configure them or add and remove discs etc.

I had been thinking of a touchscreen, like a tablet computer,
that would speak any text you stroke a fingertip over.

More practical; if you could add a video-INPUT socket onto some old
computer, and also give one of its U S B ports a driver to make it
emulate a U S B keyboard; then you could use a video cable and a
U S B cable to plug the main computer into it, and have it work
as a virtual SVGA monitor with a speakup-like keyboard-interface.
It might even be possible to add a video-input socket to a
raspberry pi, and make the whole thing small and light.
Ambitious project, though, way beyond me...


I wrote:
> If I were rewriting yasr, I'd probably do it in perl or lua,
> which would reduce the code-bulk by several times,
> and make it more maintainable and portable.

Trevor wrote:
> I think some of it will need to stay in C / C++

Well, Perl and Lua can easily invoke routines written in C

> but allowing it to be scripted with lua sounds interesting.

After the fork and dup, in Perl you could use the existing
  http://search.cpan.org/perldoc?Term::Terminfo
module to intercept all the right escape codes, and the
  http://search.cpan.org/perldoc?Speech::eSpeak
module to speak with.  In Lua, best first step might be to translate
those Perl modules into Lua and get them onto the luarocks server...

Ideally (back in Perl), you would use Jouke Visser's
  http://search.cpan.org/perldoc?Speech::Synthesis
module to provide a generic interface for different TTS Engines,
but it was last updated in 2005 and only supports:
  MSAgent, SAPI4 and SAPI5 (Win32), MacSpeech and Festival
  (connects to the default port on localhost)
so it doesn't yet know about espeak. It should be possible
to use the Speech::eSpeak module to give it espeak, though...

We have too many of these half-abandoned projects :-(

Peter

http://www.pjb.com.au      pj@pjb.com.au     (03) 6278 9410
"Follow the charge, not the particle."  --  Richard Feynman
 from The Theory of Positrons, Physical Review, 1949