Text-To-Speech & Voice Interface for Puppy?

[RetroTechGuy]

there is an espeak pet that is smaller and has a few useful wrapper scripts.

http://www.murga-linux.com/puppy/viewto ... 96&t=72589

Excellent!

[mt0ne]

@gcmartin YES! ->http://www.murga-linux.com/puppy/viewtopic.php?t=88095

@ETP is WIN! This distro is my new 'MAIN' puppy!

I had been hanging out in slacko and racy puppy land and had yet to 'really' try precise. ... DUH!
Not only is this distro THE place to start for all things voice/accessibility, but it also uses whatever xvesa stuff bypasses the nvidia insta-crash of liney doom that has plagued my primary machine! My Vostro is now windows free! Not just safely running linux on a partition free, but owning the whole hard-drive with 10G swap partition, uber storage on sda3 with frugal install on sda1!

Both of my PCs are now in puppy zen! FREE at LAST!
To Win is not to Win(dows).


Also Following the steps for setup has revealed my first OBVIOUS USER ERROR!
Yaaay!
The wires in my headsets microphone are toast!! Explains a lot! But since I am no longer waiting for my old graphics card to die and be replaced, I can work toward getting a replacement. In the mean time I have a mono headset that's workable and will be testing fully soon!


@ted dog YES! ->http://www.murga-linux.com/puppy/viewto ... 96&t=72589
Works instantly! Still sounds like Joshua/Hawking but unlike ALL windows speakers technosaurus has STARTED OUT able to use punctiation and pauses in his program! Check out the truly awesome prose...
Save the following to text and target espeak pup from DOCUMENT heading in start menu...

No plain fanfold paper could hold that fractal Puff --
He grew so fast no plotting pack could shrink him far enough.
Compiles and simulations grew so quickly tame
And swapped out all their data space when Puff pushed his stack frame.
CHORUS:
Puff the fractal dragon was written in C,
And frolicked while processes switched in mainframe memory.
Puff the fractal dragon was written in C,
And frolicked while processes switched in mainframe memory.
Puff, he grew so quickly, while others moved like snails
And mini-Puffs would perch themselves on his gigantic tail.
All the student hackers loved that fractal Puff
But DCS did not like Puff, and finally said, "Enough!"
(chorus)
Puff used more resources than DCS could spare.
The operator killed Puff's job -- he didn't seem to care.
A gloom fell on the hackers; it seemed to be the end,
But Puff trapped the exception, and grew from naught again!
(chorus)


I <3 pups with WORKING menu items that work with fire and forget perfection! @technosaurus is WIN!

Now all that this disto needs is a voice creation app and some sexy voices, and we will be so far ahead of the rest of the planet, it won't even be fair.

Recruit at will!

OPEN SOURCE it before the 'pay to play' folks bother to get it done right.
As someone who tried to PAY voice dragon, cepstral and a couple fly-by-night mail-order folks, tried N+1 downloads got rooked EVERY time, IMHO 'done right' isn't out there or even a priority.

I've never been closer than I am now with puppy!

For all windows refugees!!! (like myself)
The term 'manually run' in linux means click your way to the folder with what you want to run in it, press the ` button (just below Esc button) type ./ (period slash) and the name of what you want to run with NO space between the / and the name. Like so...
./configure mudos.cfg
In this example modos.cfg is the command I'm running and configure is the command linux seems to use to RUN IT. Alot of linux instructions leave this out! So if make, make install and configure aren't working try ./ It's like a magical incantation or something, but it WORKS

Also kildclient works instantly on this distro! (note some lost time on my part, what few off hours I've had this month went straight to Lars... *sigh*)
The client has perl stuff in it, and I might be able to link it to espeak.
With some prodding and poking it's response to hyperlinks might just work to pass webpage text to espeak as well... I have alot of catching up to do...

As I begin to catch up, http://lpmuds.net/files/deadsouls/ds3.8.1.zip worked when I followed instructions CAREFULLY, yielding some experience with urxvt and I now have a cheat-sheet started for stuff like 'make' and 'make install' and configure...(see above)
I think my linux skill is approaching 002!

[mt0ne]

anyone know perl for 'select text' in the context of espeaks speak selected text automatically? initial web-searches are FAIL so far lolz. A handy cheat sheet for this function in any and all languages should trickle into the docs for this stuff.

Oh yes, that reminds me, where do docs go in puppy-stuffs, is there a guide in some thread as to where everything in puppy lives? I'm getting there, but random clicking thru file trees with no idea what icons/filetypes are is risky as I can click espeak select text type things to start them, but linux things run pretty quitely and I have yet to figure out how to close things 'politely' when I can't find them.

I'm so close to being self sufficient to learn! I can feeeel it!

[mcewanw]

You could try Vinux, which is Ubuntu for the visually impaired.
fokSyf Eye R was by mcewanw, whose site is now down. It is CLI only (text mode apps, but can be run in a GUI environment), but is a very well put together pet. Originally ran on 301, but will work on later pups -- sometimes YASR or speech-dispatcher complain (forget which). Usually a recompile of the offending libs works.
I do have a copy I can post here if you think you might want to try it.

I'm considering reviewing that old dotpet of specially configured commandline programs (including specially configured versions of espeak, yasr, and speech dispatcher). I may also upload the old websites (and maybe tidy them up in the process) once I find some time to do so. In the meantime, the old foksy dotpets can also be accessed via this link:

https://drive.google.com/folderview?id= ... sp=sharing

For Puppy 4 series and above you probably also need the patched version of krecspk, which is also in that download folder. If you plan to try foksy, you would need to make sure your existing Puppy does not include any other versions of the above programs; perhaps best to test it on a clean Puppy install. You need to do a few things to set it up and get it running and carefully read the documentation on usage. These are documented in the main murga linux thread for it here:

Make X apps speak text too! fokSyf Eye R (Puppy and Ubuntu):

http://www.murga-linux.com/puppy/viewto ... 1163210554

Today I would probably re-write the krecspk mp3 voice recording component to either use precord for recording (or a modified version of that), which I'm thinking of looking into. krecspk uses sox (rec) for recording whereas precord uses arecord and lame (or ffmpeg for some audio types). arecord, lame and ffmpeg are provided by default in modern Puppies. No voice interface was provided in foksy, though I did try out an old programe called cvoice which I stumbled on somewhere.

Cheers, mcewanw

EDIT: I've now uploaded some of my old project work including wiak and fokSy Eye R to website: http://arkitektia.heliohost.org. Will be accessible via arkitektia.com once DNS sorted out. My software is currently found there under Products -> Software Designs (which is also accessible from right hand pane of the website)

[RetroTechGuy]

I just ran across this project:

Chatterbox STT / TTS / TTA project
(Speech-To-Text / Text-To-Speech / Text-To-Action)

http://www.murga-linux.com/puppy/viewtopic.php?t=89258

[mt0ne]

I have now entered the time of year officially described as "panic season".
Between now and Xmas I'll be working non-stop with little/no time for anything else...
I shall return!

[Ted Dog]

I have now entered the time of year officially described as "panic season".
Between now and Xmas I'll be working non-stop with little/no time for anything else...
I shall return!

He He, as you get older and family kids are no longer kids, this time of year time just magically opens up. That young generation, with jobs, corp partys, kids 'talent' shows, shopping... are so over whelmed that we who served the time, lived that hectic life, before are free to send a card with a store gift card and be 'DONE'

[mt0ne]

I have figured out that the GNOME-Mplayer running music will run side by side with Pmusic doing the speaking as default media player
speak clipboard is now firmly planted on my desktop and I can listen to any webpage with ctrl-a and not interrupt my tunes

AT LAST I can start taking in information at will!
I could use a pointer to where the speed control etc is for speech but hey, it's keeping me on my toes...

[mt0ne]

Thought I would check in and update!
The frenzy of the season is about a week away from being over and I've been downloading and testing what I can in my spare time. (details to come later)
Along the way I came across an awsome webpage that had concise instructions for linking several different mud clients to several different screen-readers / TTS engines. BUT I CLOSED MY BROWSER AND LOST THE LINK!!!

It had a black backround and bold yellow text in the menu and as paragraph headers. If you see it please E-mail the link to mt0neg@gmail.com
it had script for linking espeak thru kildclient in perl, and underneath that it had instructions for similar procedure with festival / festvox! I was following them when I accendently closed chrome and flushed my history, I have since disabled that function... I think the reason I can't find it again is that I followed a secondary link to it and it may have been coded in frames so it's not showing up in google search... (tried and tried and tried and cried...)

On the good news side of things this thread has started showing up in google search pretty high up! Even when I use different machines/connections to search!

There IS A LOT of interest out there, especially in TTS + mud client searches! Lots of people are looking, but when they get their solution built, they AREN'T posting FOLLOWUPS...

In January I'll be reorganizing everything and distilling all that I've found into my first solid solution (hopefully) and will be testing on a fresh install of Obedient puppy (this time in secure mode with proper non admin status as default post build). In the mean time I'm still using the espeak 'speak clipboard script' and in kildclient I have a copy of clipboard.pm and it sorta works with
/use Clipboard
and an OnReceivedText hook
/Clipboard->copy($hookdata)
You'll need to set your line length pretty long though as multi-line chunks are still an issue as are blank lines right after good ones...
screen 300 35
with as small a font as your client will manage.

Although the 'speak clipboard' script is the most basic, it's been the most robust, it works in any window with CTRL+A and I can copy text into a text file for repeated listening. It is also the only solution I've found that doesn't cause things to hang until it's stopped talking. I have no idea why though...

In the mean time, I've been downloading and testing every linux speech system and screen reader I can find and as time allows, testing them.

While many linux projects rise and fall, I've been noticing a strange "Iced out" pattern with regard to voice interface and particularly TTS/screenreading projects... I'm begining to think someone out there has been quashing them.
I hope it's just me, but just in case I'm planning to develop my final solutions entirely from scratch...

[mt0ne]

So I had the day off and figured I'd try again...
UGH!
The problem with the puppy package manager is it installs things that I can't find...
and then if I do find them even with a console I can't get them to do ANYTHING. Look on the website of a linux project and you get install instructions (sometimes) but I can never find step 1
now that I've installed
flite
gnomespeech
emacs
festival
festvox
yasr
I can't get any of them to do a THING, they have no docs for how to USE them and the scripts do nothing...
they give NO instructions for any command line thing to make them work...
test-gnomespeech
works, but there's no information on how to use ANY other commands...
*INSERT ANGRY NAP HERE*
"How is the name of yasr 'Yet another screen reader' when I can't get anything but technosaurs' speakclipboard to acctually DO anything...

So like, where do I get this all powerfull linux knowledge that fills this painfull gap? I ~would~ RTFM, if I could find one...

Result day off wasted 16 hours, accomplished nothing, learned nothing and made no progress
</rage>

[Trobin]

Could I ask what it is that you are trying to do?

I've read this tread three, or more, times and have no idea. Are you trying to get console applications to speak.

If so, then you need to install espeak, speech dispatcher, and yasr. Read through this thread:

http://murga-linux.com/puppy/viewtopic. ... 41&t=24571

Be warned, I ramble a lot through it.

If you can het Foksyfeyer from the link that mcewanw gave you then the above is done. Once installed.

But they are console programs and won't read windows (GTK) apps.

If you are trying to get windows apps speaking, well I don't know if anyone has successfully with puppy. At least not beyond the experimental stage.

[mt0ne]

Your absolutly right!
My ultimate goal is to have a speech engine up and running that will speak text piped in from all applications, encluding the desktop and voice command allowing me to "do everything".
and have 0 java on my machine... [personal hatred of java update badness left over from windows days]
An insanely tall order, I know...

But! I'm slowly making progress! OP V3 seems to have installed both festival AND Technosaurous' speak pets and can run them both! (OP2 seemed to balk at festival pup on my machine)
espeak (standard ppm package) gives -w (no sound) only error, I need to to some work there...

I also found

http://hak5.org/episodes/haktip-51

and now have a hotkey to speak selected text!
[For puppy users the hotkey controls are in Menu>Utility>PupControlPanel>Window Manager Settings>Keyboard Shortcuts ]
Or!
/root/.jwm/jwmrc-personal {or personal2 }
You'll need to click the eye to see hidden if you are using ROX...
look for
<!-- Key bindings -->
in that file and add
<Key mask="CS" key="z">exec:/usr/bin/talkit.sh</Key>
on a new line right below it.

save and exit
then in a terminal window type these lines

cat << eof > talk.sh
#!/bin/bash
xsel | festival --tts --pipe
eof
cmod +x ~/talkit.sh

install xsel in puppy package manager

and poof! CTRL-SHIFT-z will now speak selected text!
(even though we did this in the wrong order)

If you run into the error Linux: can't open /dev/dsp
echo "
(Parameter.set 'Audio_Command "aplay -q -c 1 -t raw -f s16 -r $SR $FILE")
(Parameter.set 'Audio_Method 'Audio_Command)
" >> ~/.festivalrc
fixes that!
(and I hope to have one just like that for espeak so the -w error goes away, not sure how technosaurous' speak works fine but espeak doesn't...)

The speech rate can be configured in /usr/share/festival/voices/english/kal_diphone/festvox/kal_diphone.scm
under the setting "Parameter.set 'Duration_Stretch" -- the lower the value the faster the speech rate.

YAAAAAY HakTip!
Believe it or not "puppylinux.com" was a link on the site in haktip-10
I'll have to update my linux skill soon

I have yasr installed but yasr seems to do NOTHING??
yasr --help = error
info =same
man -k yasr = deadend search 0 results...
it's in Pprocess manager a couple times but nothing seems to get it to speak...
Speech dispatcher hangs my puppy package manager but Gespeaker seems to have installed ok...
I seem to still be missing the go between part that pipes output to the speech from bash/urxvt ,

(also the desktop>urxvt settings thingy doesn't seem to be changeing forground or background values in console even after restart...)

on the mud front I have a thread at
http://www.kildclient.org/forum/viewtop ... f=3&t=4020
but it's still a work in progress, hopefully soon my now working tk bindings for perl will allow us to get a finished work.

I have so many partially working bits installed it's getting hard to keep track already, and it's only week 1 with the new distro...

[Trobin]

Speech dispatcher hangs my puppy package manager but Gespeaker seems to have installed ok...
I seem to still be missing the go between part that pipes output to the speech from bash/urxvt

How are you installing Speech Dispatcher, and with what version of Puppy?