[SOUND] speechd lets your Computer speak

Message

wrs4 · Post by **wrs4** » Mon Jun 23, 2003 6:13 pm

Thanks. I'll give it a try when I get the computer back from the repair shop. It was freezing just about every time I ran gcc

and when I checked the CPU temp via the bios, it said "0 C" after running for a whole week, so I think I'm going to be RMA'ing some part(s) or other.

dylix · Post by **dylix** » Wed Jun 25, 2003 4:55 am

haha.. i added (below) to my local.start, haha

Code: Select all

echo "System security activated to high. Please stand away from the computer" > /dev/speech

my g/f tripped out about it.

ElzWhere · Post by **ElzWhere** » Wed Jun 25, 2003 7:03 am

Hell yeah this is tyte! I had it reading my dmesg when i first got it installed and she thought someone else was in the house and tripped out too.

When i first saw this topic i had to get this installed. I had trouble on the laptop...so i thought i'd try it out on the server, and sure nuff it worked perfectly. I had to get the female voice on it though.

the reason why i had trouble on the speech tools compile with festival on the laptop was due to the day before i emerged -uP on accident...it Pruned my whole system so festival couldnt compile because of the gcc version issue. It was a bitch to fix with all the patches found on the festival site but well worth it. Thanks for
the post Tobi.!

There is no chance for the mentally dead but there is a chance for the physically dead....so thats my job! To resurrect the physically dead, by diggin up dem graves!

[/code]

wrs4 · Post by **wrs4** » Wed Jun 25, 2003 11:46 am

Hmm...now you have me thinking about the possibilities of speechd, some x10 motion detectors, and audio inputs from my computer (in the bedroom) to my receiver in the front room. A la

Intruder detected, please step away from the merchandise. Detonation will commence in 10 seconds. 10, 9, 8, ...

JM_ · Post by **JM_** » Wed Jun 25, 2003 4:27 pm

Hello. I'm trying to compile speechd, but festival 1.4.2-r3 fails. I also tried the 1.4.3 (latest) downloaded from the official website and it fails too. Has anyone got this to work with gcc 3.3 and glibc 2.3.2-r2? (actually I have glibc 2.3.2 with 20030616 branch update)
This is what I get:

Making in directory ./siod ...
making dependencies -- siodeditline.c el_complete.c editline.c el_sys_unix.c slib.cc slib_core.cc slib_doc.cc slib_file.cc slib_format.cc slib_list.cc slib_math.cc slib_sys.cc slib_server.cc slib_str.cc slib_xtr.cc slib_repl.cc siod_fringe.cc siod_server.cc io.cc trace.cc EST_SiodServer.cc siod.cc siod_est.cc
gcc -c -fno-implicit-templates -O3 -Wall -Wno-non-template-friend -DSUPPORT_EDITLINE -I../include slib.cc
In file included from /usr/lib/gcc-lib/i686-pc-linux-gnu/3.3/include/g++-v3/backward/iostream.h:31,
from ../include/EST_iostream.h:53,
from ../include/EST_String.h:50,
from ../include/siod.h:17,
from slib.cc:88:
/usr/lib/gcc-lib/i686-pc-linux-gnu/3.3/include/g++-v3/backward/backward_warning.h:32:2: warning: #warning This file includes at least one deprecated or antiquated header. Please consider using one of the 32 headers found in section 17.4.1.2 of the C++ standard. Examples include substituting the <X> header for the <X.h> header for C++ includes, or <sstream> instead of the deprecated header <strstream.h>. To disable this warning use -Wno-deprecated.
In file included from ../include/EST_String.h:50,
from ../include/siod.h:17,
from slib.cc:88:
../include/EST_iostream.h:54:26: strstream.h: No such file or directory
In file included from ../include/EST_TList.h:50,
from ../include/EST_string_aux.h:43,
from ../include/siod.h:18,
from slib.cc:88:
../include/EST_TIterator.h:206: warning: `EST_TStructIterator<Container,
IPointer, Entry>::Iter' is implicitly a typename
../include/EST_TIterator.h:206: warning: implicit typename is deprecated,
please see the documentation for details
../include/EST_TIterator.h:206: warning: `EST_TStructIterator<Container,
IPointer, Entry>::Iter' is implicitly a typename
../include/EST_TIterator.h:206: warning: implicit typename is deprecated,
please see the documentation for details
../include/EST_TIterator.h:233: warning: `EST_TRwIterator<Container, IPointer,
Entry>::Iter' is implicitly a typename
../include/EST_TIterator.h:233: warning: implicit typename is deprecated,
please see the documentation for details
../include/EST_TIterator.h:233: warning: `EST_TRwIterator<Container, IPointer,
Entry>::Iter' is implicitly a typename
../include/EST_TIterator.h:233: warning: implicit typename is deprecated,
please see the documentation for details
../include/EST_TIterator.h:280: warning: `EST_TRwStructIterator<Container,
IPointer, Entry>::Iter' is implicitly a typename
../include/EST_TIterator.h:280: warning: implicit typename is deprecated,
please see the documentation for details
../include/EST_TIterator.h:280: warning: `EST_TRwStructIterator<Container,
IPointer, Entry>::Iter' is implicitly a typename
../include/EST_TIterator.h:280: warning: implicit typename is deprecated,
please see the documentation for details
make[1]: *** [slib.o] Error 1
make: *** [siod] Error 2

!!! ERROR: media-sound/festival-1.4.2-r3 failed.
!!! Function src_compile, Line 69, Exitcode 2
!!! (no error message)

JM_ · Post by **JM_** » Wed Jun 25, 2003 5:35 pm

Ok... sorry, my mistake. I forgot to check bugs.gentoo.org.

Schwin97 · Post by **Schwin97** » Wed Jun 25, 2003 5:56 pm

Not sure about this but...

Since this is just a device... with permissions, couldn't you have your server ssh into your workstation (or any computer with this installed) to notify you of problems, intrusions, etc? Just for those who may have the server in a different room/building...

Just a thought

Schwin97

schlehmil · Post by **schlehmil** » Sun Jun 29, 2003 7:50 am

Hello,

at first I would like to say thank you for that fine tutorial.
but there is one problem with using the mbrola voise set by default.
I've created .festivalrc with following content:

(set! voice default 'voice_us1_mbrola)

But everytime I restart the computer and init-Script "speechd" (found in this thread" is load the default male voice is loaded. When I restart the init-Script by "/etc/init.d/speechd restart" everything is working fine.
Does anybody has the same problem?[/quote]

smouge · Post by **smouge** » Sun Jun 29, 2003 9:55 pm

Great post,

I have just changed the local.start and local.stop a little. The meaning is to do an extra check if speechd is running besides the existence of /dev/speech.

The first time I tried to use these scripts my computer hang because I had not added speechd to the start-up scripts. With this extra check this will be prevented if for some reason speechd had not been started, or stopped before shutting down the system.

/etc/conf.d/local.start

Code: Select all

# Copyright 1999-2002 Gentoo Technologies, Inc.
# Distributed under the terms of the GNU General Public License, v2 or later
# $Header: /home/cvsroot/gentoo-src/rc-scripts/etc/conf.d/local.start,v 1.2 2002/05/12 21:48:18 azarah Exp $

# This is a good place to load any misc.
# programs on startup ( 1>&2 )


if [ -e "/dev/speech" ] ; then 
   pid=`ps ax | grep "/usr/bin/perl -w /usr/bin/speechd" | grep -v grep | awk '{print $1;}'`
   if [ "X${pid}" != "X" ] ; then
      date="$(date +%H)" 
      if [ "$date" -ge "6" -a "$date" -le "12" ] ; then 
         say "Good morning" 
      else 
         if [ "$date" -gt "12" -a "$date" -le "18" ] ; then 
            say "Good afternoon" 
         else 
            if [ "$date" -gt "18" -a "$date" -le "24" ] ; then 
               say "Good evening" 
            else 
               if [ "$date" -gt "0" -a "$date" -lt "6" ] ; then 
                  say "Good night or is it morning already" 
               fi 
            fi 
         fi 
      fi 
      sleep 1
      say "Gentoo system $HOSTNAME is ready" && sleep 4 && saytime 
   else
      echo 'Error .. speechd not loaded' 
   fi
else 
      echo 'Error .. speechd not loaded' 
fi

/etc/conf.d/local.stop

Code: Select all

# Copyright 1999-2002 Gentoo Technologies, Inc.
# Distributed under the terms of the GNU General Public License, v2 or later
# $Header: /home/cvsroot/gentoo-src/rc-scripts/etc/conf.d/local.stop,v 1.2 2002/05/12 21:48:18 azarah Exp $

# This is a good place to unload any misc.
# programs you started above.
# For example, if you are using OSS and have
# "/usr/local/bin/soundon" above, put
# "/usr/local/bin/soundoff" here.

RUNLV=`runlevel | cut -d ' ' -f 2` 
if [ -e "/dev/speech" ] ; then 
   pid=`ps ax | grep "/usr/bin/perl -w /usr/bin/speechd" | grep -v grep | awk '{print $1;}'`
   if [ "X${pid}" != "X" ] ; then
      if [ "$RUNLV" == "0" ] ; then 
         say "$HOSTNAME is shutting down" && sleep 3 
      elif [ "$RUNLV" == "6" ] ; then 
         say "$HOSTNAME is rebooting now" && sleep 3 
      else 
         echo "Found invalid runlevel" 
      fi 
   else 
      echo 'Error .. speechd not loaded' 
   fi
else 
   echo 'Error .. speechd not loaded' 
fi

ErnieS · Post by **ErnieS** » Tue Jul 08, 2003 2:00 am

Has anyone found the fix for this?

Code: Select all

# /etc/init.d/speechd stop
 * Stopping speechd...                                                    [ ok ]
# festival --server
socket: bind failed

MaGuS · Post by **MaGuS** » Tue Jul 08, 2003 2:26 pm

It's relay cool, but it would be nice to get the german languages out of my pc.

Anyone knows a solution?

nephros wrote:thanks, that thing is way cool.

God, this reminds me of old Amiga times!

Now, anyone know how to set up other languages.

I downloaded the German deX packages, and unzipped them to /usr/lib/mbrola/voices/german/de{1..5},
edited voices.scm to include "de1" on top of the list
Code: Select all
(defvar default-voice-priority-list
  '(de1
    en1_mbrola
    ...
and edited festivalrc to:
(set! voice_default 'voice_de1)

but that does not work...

Chris Finch · Post by **Chris Finch** » Wed Jul 09, 2003 12:22 pm

Sharing my experience:

I've had some troubles with the init scripts posted in this thread (btw. mind that in one of them two closing brackets are missing, in case you stumbled over that). So what I did was plainly adding

Code: Select all

speechd
echo "Welcome to gentoo linux." >/dev/speech
echo "Today is $(date +"%A, %B %-d. The time is %-M minutes after %-I")." >/dev/speech

to /etc/conf.d/local.start and

Code: Select all

if [ -e "/dev/speech" ] ; then
   pid=`ps ax | grep "/usr/bin/perl -w /usr/bin/speechd" | grep -v grep | awk '{print $1;}'`
   if [ "X${pid}" != "X" ] ; then
        echo "Shutting dow the computer at $(date +"%-M minutes after %-I.")" >/dev/speech && sleep 6
   fi
fi

to local.stop. The checking (taken from one of the foregoing posts) is advisable in order to avoid a hang on shutdown in case that speechd was killed for some reason. I don't kill the demons, they're going to get killed, anyway. Is there any disadvantage of doing things in this way?

Two hints: If you have a problem with speechd but none with festival, have a look at http://www.speechio.org. There is a faq entry that helped me. The reason for my problems was that I had /etc/hosts set up wrong. On the same site you find a little perl script called "speechbiff" that you can call from procmail and that reads sender and subject of new mail on arrival to you. Was very easy to set up and is a fancy gadget to have.

akiross · Post by **akiross** » Sun Jul 13, 2003 12:42 pm

Hey, It's wonderfoul!

but, i dont like the voice... is there a way to have a voice like a robot?
a mechanized voice a mean...

thnx very much

bye

nalin · Post by **nalin** » Sat Jul 26, 2003 12:10 am

Just a question here, and probably one that is not directly due to speechd, but...

I have a desktop where speechd works fine, I have never noticed a long lag whatsoever, i echo, it talks

I just set this up on my notebook, and sometimes (~50% of the time) it seems to buffer output, so that I echo and speech occurs anywhere from immediately to say a minute down the road. If it occurs a minute down the road and I have echoed multiple things, it says them all at once (in sequence, not concurrently). If I do something else that causes sound it plays as normal and does not suffer this buffering effect, so I dont believe it is a module unloading or arts shutting down, etc.

My question here is whether anyone else has noticed this, and is there something I can do to avoid this lag.

The desktop is using oss drivers for a sblive card, the laptops alsa drivers for an intel ac97 codec card.

niltonvolpato · Post by **niltonvolpato** » Sat Jul 26, 2003 3:20 am

nalin wrote: I have a desktop where speechd works fine, I have never noticed a long lag whatsoever, i echo, it talks

I have the same problem. It happen also with xmms, or mpg123. There is a BIG delay till the real sound plays. I almost trying to suicide.

If someone had the same problem or know how to solve it, please help!

Thanks,
-- Nilton

Oxidative · Post by **Oxidative** » Sat Jul 26, 2003 12:25 pm

I have had the same problem. I fixed it by replacing my /usr/bin/say script with

Code: Select all

echo $1 | festival --tts

Now when you use the say script, it's much faster!

Good luck

nalin · Post by **nalin** » Sat Jul 26, 2003 6:58 pm

nalin wrote:Just a question here, and probably one that is not directly due to speechd, but...

I have a desktop where speechd works fine, I have never noticed a long lag whatsoever, i echo, it talks

I just set this up on my notebook, and sometimes (~50% of the time) it seems to buffer output, so that I echo and speech occurs anywhere from immediately to say a minute down the road. If it occurs a minute down the road and I have echoed multiple things, it says them all at once (in sequence, not concurrently). If I do something else that causes sound it plays as normal and does not suffer this buffering effect, so I dont believe it is a module unloading or arts shutting down, etc.

My question here is whether anyone else has noticed this, and is there something I can do to avoid this lag.

The desktop is using oss drivers for a sblive card, the laptops alsa drivers for an intel ac97 codec card.

I figured out my problem:
My soundblaster desktop system has hardware sound mixing, arts takes up a single channel, and the other channels remain open for things such as /dev/speech without explicitly piping them through artsdsp
My ac97 laptop lacks hardware mixing and thus has a single sound channel which arts monopolizes for 60 seconds after any sound has played. [when kde or gnome or any wm that has a program such as esd, artsd, etc which handles sound output and is running] /dev/speech explicitly waits for the channel to become available, in my case it takes 60 sec for arts to timeout and unload...hence the delay

I *could* fix this to some extent by changing my settings so that arts unloads after say 5 seconds, that way, as long as I am not doing anything which explicitly hogs the channel such as listening to mp3's, the speech will be delayed minimally.

I would be more interested in piping the /dev/speech output through artsdsp, but cannot seem to get this to work properly (I bought the soundblaster in my desktop soley to avoid piping through artsdsp and associated software mixing problems, and the ease of use and avioded hassle was worth every penny). Speechd has a esd command line option as well as a compilation flag, but nothing yets exists for arts, and im not sure it would work without significant modification (it is implemented by calling "esdsp festival --server" rather then "festival --server", I have tried killing all festival and speechd stuff and runing "artsdsp festival --server," and then "echo text | festival -tts", but it does not pipe through arts, it still waits for /dev/dsp to becomee available.

Anyone with a workaround or suggestions would be appreciated.

p.s. Just a thought but is ln -s /usr/kde/3.1/bin/artsdsp /usr/bin/esdplay a viable option for creating a fake esd interface (and vice versa if your using gnome) when programs support exactly one of the sound systems?

EDIT
echo 'Hello from Festival'| artsdsp festival --tts
it works...working on making it pipe through /dev/speech
END EDIT

OneOfOne · Post by **OneOfOne** » Mon Jul 28, 2003 3:08 am

Gio wrote:In bugzilla #22053 I updated festival and speech-tools to compile on gcc-3.2.3. Also did some seperation and clean up - comments welcome.

** update ** compiles on gcc 3.3 too.

wish you were a girl

THANK YOU

peace

viperlin · Post by **viperlin** » Mon Jul 28, 2003 5:51 am

BradB wrote:OK, now for an interesting use of /dev/speech - where abouts would be fun places to symlink it to?
ie ln -s /dev/null with /dev/speech - (urg I'd think)
or could you hook it up so that kernel logs got spoken out?

Probably annoying, but maybe fun

Brad

i tryed "echo /var/log/syslog > /dev/speech"

when i hit CTRL-C it didn't stop, it's still going.

I HATE YOU, it's 6:54am now.....the log is up to reading out 1:23am logs.....
i'm going to surf aroud top to find it and kill it.

panserg · Post by **panserg** » Mon Jul 28, 2003 4:49 pm

Does anyone know any good place to download some new voices? I, personally, would like to find the voice of "John Frink", the comuter geek from Simpsons.

glamdringlfo · Post by **glamdringlfo** » Fri Aug 01, 2003 8:14 pm

Hey, that's awesome.

I wrote up some short scripts so that when gkrellm notices my cpu temp passes a certain threshhold, the computer warns the user with audible alerts. I set the threshold low to test it, and it freaked my roommate out.

Thanks for explaining how it works. The different voices bit, too.

Stay cool.

Skal!
Glamdringlfo

iplayfast · Post by **iplayfast** » Mon Aug 04, 2003 7:24 am

Is there some way to make festival go to it's own audio channel? The offical way of making it louder doesn't work, and it's a mutter against my music.

herring · Post by **herring** » Tue Aug 05, 2003 12:37 am

carrett wrote:hey, this is awesome. when i do: echo "Hello world" > /dev/speech
the system just sits there, no sound comes out and i have to ctl-C to stop the command. however, echo "Hello world" | festival --tts -
works perfectly, anybody know how to make the device work?

Hi !
Had this problem myself.
Examining /var/log/speechd (EDIT: and festival --server), I found festival was rejecting my localhost client. Reason was, localhost was not named localhost, but gentoo

However, making a /root/.festivalrc
containing

Code: Select all

(set! server_access_list '("gentoo"))

festival happily accepted my localhostclient, and finally piping to /dev/speech worked as expected

EDIT: This was probably a /etc/hosts issue (see Chris Finch post above)
http://www.speechio.org/speechio.faq.html

error26 · Post by **error26** » Mon Aug 11, 2003 4:45 pm

Hm, when i write to /dev/speech on a kde terminal it qeuques until i logg out, but w/ fluxbox it works.
Do you have any idea why this could be?

gmoney · Post by **gmoney** » Thu Aug 21, 2003 7:12 pm

Anyone know how to get festival to bind just to localhost (127.0.0.1) instead of *? I would like to use it but leaving ait running as root and waiting for connections from the whole internet isn't an option. Firewalling off ports doesn't give me the warm fuzzy feeling that forcing it to bind to localhost would......

[SOUND] speechd lets your Computer speak

HaHa

festival --server bind failed

Re: This is cool,

Same problem

Re: festival and speech-tools update in bugzilla

this all works good but...

on KDE it queqes

Security?