speech-dispatcher rate and volume settings for espeak 1.25

speech-dispatcher rate and volume settings for espeak 1.25

[William Hubbs]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

All,

I am having difficulty figuring out how to set the speech-dispatcher
settings for  GenericVolume and GenericRate for espeak 1.25.

The rate isn't as big of a deal because the mapping is pretty close --
the default for speech-dispatcher maps to 160 which is 10 wpm slower
than the espeak default.

The volume, however, is way off.  Espeak's volume range is 0 to 200 with
a default of 100, and speech-dispatcher is from -100 to 100 with a
default of 0.  I have attempted "GenericVolumeAdd 100" and
"GenericVolumeMultiply 100", but that did not seem to work.  Does anyone
have any other suggestions?

Thanks much,

- -- 
William Hubbs
gentoo accessibility team lead
williamh@gentoo.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)

iD8DBQFGYxWGblQW9DDEZTgRArpOAJ48XH6y5QuZAcSNaYrp7E5fxzyfEQCgkezm
sTncUaTwf8maqR61zvm4pF8=
=W6tt
-----END PGP SIGNATURE-----

Re: speech-dispatcher rate and volume settings for espeak 1.25

[Hermann]

Hi William,
I have the same settings, and my synth is very loud.
Perhaps a silly question, but did you restart Speech-dispatcher after 
changing the settings?
You can try adjusting the system volume by using aumix; is it installed on 
your machine?
If so, try:
aumix -v100 -w100
If you are satisfied with the volume type:
aumix -S
Hermann

Re: speech-dispatcher rate and volume settings for espeak 1.25

[William Hubbs]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sun, Jun 03, 2007 at 10:09:59PM +0200, Hermann wrote:
> Hi William,
> I have the same settings, and my synth is very loud.
> Perhaps a silly question, but did you restart Speech-dispatcher after 
> changing the settings?
Hi Herman,

here is my situation right now:  The GenericVolumeAdd and
GenericVolumeMultiply settings in /etc/speech-dispatcher/modules/espeak-generic.conf are both 100.
Also, I have turned debugging on in espeak-generic.conf so I can see
what is going on with the module.

If I type:
espeak "This is a test."
then
spd-say "This is a test."
the second time is much louder than the first, and should not be since
the default volume is supposed to be the same.

Looking further into this, I find from /var/log/speech-dispatcher.log:

 Sun Jun  3 15:21:25 2007 [765657]: child: synth command = |echo "This is a test." | espeak -w /tmp/espeak.wav -v en -s 160 -a 200 -p 50 --stdin && aplay /tmp/espeak.wav|

 Notice the "-a 200" on the espeak command line above.
 Speech-dispatcher is setting the espeak volume to maximum, which is not
 what I would expect with those settings.

Thanks,

- -- 
William Hubbs
gentoo accessibility team lead
williamh@gentoo.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)

iD8DBQFGYyW2blQW9DDEZTgRAq/xAKC4WNrn+0zMK1Wz7KgxwHCv+Zq5LACeK6Xg
CNCKGuxEEy9V+TYUndJS/nU=
=Nkcd
-----END PGP SIGNATURE-----

Re: speech-dispatcher rate and volume settings for espeak 1.25

[Hermann]

OK, what result do you expect? Is the synth too loud or too low?
If too loud, you can decrease the GenericAdd and GenericMultiply values - let's 
say to50.
If it's to low, you can change the default volume in speechd.conf to a 
value between 0 an 100.
Or you can try to change the values in the GenericExecute string in 
espeak-generic.conf. Instead of $VOLUME you can try to insert a specific 
value.
I guess you have to experiment a bit. I figured out my favorite values this 
way.
Hermann

Re: speech-dispatcher rate and volume settings for espeak 1.25

[William Hubbs]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sun, Jun 03, 2007 at 11:38:49PM +0200, Hermann wrote:
> OK, what result do you expect? Is the synth too loud or too low?

The speech from

espeak "This is a test."
and
spd-say "This is a test."

should have the same volume.  I was just wondering if anyone on the list
had figured out what the proper  "genericVolumeAdd" and
"genericVolumeMultiply" values are to make that happen.

> Or you can try to change the values in the GenericExecute string in 
> espeak-generic.conf. Instead of $VOLUME you can try to insert a specific 
> value.

I really don't want to do this, because if I can get the values right, I
want to set it up that way for the gentoo distribution.

Thanks,

- -- 
William Hubbs
gentoo accessibility team lead
williamh@gentoo.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)

iD8DBQFGY0dGblQW9DDEZTgRAtZNAKCPtSRpYCy/v6Sg7pGAfUGra9/fCACdH5Nz
MG8nKChxqWU8nWi7hc37jfs=
=5LKF
-----END PGP SIGNATURE-----

Re: speech-dispatcher rate and volume settings for espeak 1.25

[Hynek Hanke]

On Sun, 2007-06-03 at 15:33 -0500, William Hubbs wrote:
> If I type:
> espeak "This is a test."
> then
> spd-say "This is a test."
> the second time is much louder than the first, and should not be since
> the default volume is supposed to be the same.
>  Sun Jun  3 15:21:25 2007 [765657]: child: synth command = |echo "This is a test." | espeak -w /tmp/espeak.wav -v en -s 160 -a 200 -p 50 --stdin && aplay /tmp/espeak.wav|
>  Notice the "-a 200" on the espeak command line above.
>  Speech-dispatcher is setting the espeak volume to maximum, which is not
>  what I would expect with those settings.

I think your eSpeak settings are correct. Please, what is the
DefaultVolume setting in your speechd.conf? I´d expect it is 100 (max),
which explains your situation (100 gets mapped to 200).

It is not unreasonable to have maximum volume set by default in
speechd.conf. Since we miss a proper audio system on GNU/Linux,
with the simple audio output through ALSA in Speech Dispatcher, setting
DefaultVolume to 0 would mean that for most speech synthesizers
the default volume will be very quiet by default comparing to the rest
of your system :( So there is this setting to 100 so that users
don´t have a problem right from the start. You can allways change it
to what you like.

I´ll change the default espeak-generic.conf parameters for the new
eSpeak range in the next version of speechd.

With regards,
Hynek Hanke

Re: speech-dispatcher rate and volume settings for espeak 1.25

[Jonathan Duddington]

On 04 Jun, Hynek Hanke <hanke@brailcom.org> wrote:
>
> I´ll change the default espeak-generic.conf parameters for the new
> eSpeak range in the next version of speechd.

For eSpeak, a volume/amplitude value of 100 is "full volume".

However, eSpeak can use greater values than this to increase the
loudness.  If you increase the value above 100, eSpeak will apply audio
compression to prevent waveform clipping.  So the voice sounds louder,
but there will be some slight distortion, although this isn't too bad
at values below 200.

You can try using the command-line espeak command to show this effect,
and hear which is the best value for you.
   espeak -a200 "hello, this is a test"

The SSML speech markup language defines volume of 100 to be both the
maximum and also the default.  This seems strange.

eSpeak allows volume greater than the default.  A human can speak
(shout) at a greater volume than his default, and that may include some
"distortion".

Re: speech-dispatcher rate and volume settings for espeak 1.25

[William Hubbs]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Mon, Jun 04, 2007 at 01:36:49PM +0200, Hynek Hanke wrote:
> On Sun, 2007-06-03 at 15:33 -0500, William Hubbs wrote:
> > If I type:
> > espeak "This is a test."
> > then
> > spd-say "This is a test."
> > the second time is much louder than the first, and should not be since
> > the default volume is supposed to be the same.
> >  Sun Jun  3 15:21:25 2007 [765657]: child: synth command = |echo "This is a test." | espeak -w /tmp/espeak.wav -v en -s 160 -a 200 -p 50 --stdin && aplay /tmp/espeak.wav|
> >  Notice the "-a 200" on the espeak command line above.
> >  Speech-dispatcher is setting the espeak volume to maximum, which is not
> >  what I would expect with those settings.
> 
> I think your eSpeak settings are correct. Please, what is the
> DefaultVolume setting in your speechd.conf? I?d expect it is 100 (max),
> which explains your situation (100 gets mapped to 200).

Hello Hynek,

Yes, that is the setting in speechd.conf.  I'll take a look at this
again.  Thanks much for the information.

- --
William Hubbs
gentoo accessibility team lead
williamh@gentoo.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)

iD8DBQFGZBGsblQW9DDEZTgRAqREAKCmy5d9g6MqzrSEGgbhV6BPIZNyaQCfcgit
P7pq7UthBD4fa5F28h9ZJnY=
=VQGq
-----END PGP SIGNATURE-----