Better Voices

[GoogleCodeExporter]

I have installed gespeaker. But sound quality is very bad. I am a window user, 
using NaturalReader. On linux, the voices are too bad.

http://www.youtube.com/watch?v=gcIBjAsc5iQ

See this video, that is a related video of gspeaker video.. see the robot.. 
gspeaker voice is like that.. anything near natural.

I am trying to switch to ubuntu. Since i use TTS a lot, i need to find a tts 
that have good understandable voice.

I found 

https://www.cepstral.com/demos/

They have linux voices. If there is decent voice in linux, i can get it as i 
use tts everyday.

It can be used with gspeaker ?

Also if gspeaker can read from selected text, it will be very useful.

Original issue reported on code.google.com by flashweb@gmail.com on 8 Jun 2010 at 9:28

[GoogleCodeExporter]

what voice are you actually using in the gespeaker interface?
gespeaker can use both espeak and mbrola voices.

about cepstral voices that's a closed project with its own voices that cannot 
be used with espeak/mbrola, so it's actually impossible to include them in 
gespeaker.
I'm actually writing external plugin support to be able to include external 
speeching apps.

Original comment by mufl...@vbsimple.net on 8 Jun 2010 at 10:59

[GoogleCodeExporter]

Thanks for the reply. Nice to know gespeaker will support external apps.

I installed following voice from ubuntu 10.04 repo.

mbrola-us1
mbrola-us2
english-mb-en1

mbrola-us1 is a male voice even though it says female.

I downloaded us1 from

http://tcts.fpms.ac.be/synthesis/mbrola/mbrcopybin.html

It is better than default other voices. Stil not natural voice.

Now i have to select mbrola-us1 voice every time i start gespeaker. Is there 
anyway i can set my prefared voice as default ?

I tried to delete some files from espeaker voice directory, but that won't 
remove the listing from gespeaker langiage listing. It will be great if we can 
select language. There is no use to have languages we can understand.

Original comment by flashweb@gmail.com on 11 Jun 2010 at 9:23

[GoogleCodeExporter]

I see gespeaker save previous settings. About previous mbrola-us1 comment, it 
may be my mistake, i see both files are of same size.

Now i have a question, on web site

http://tcts.fpms.ac.be/synthesis/

There is demo of 

American English (us1 female voice, 198K Wav file)
http://tcts.fpms.ac.be/synthesis/mbrola/demo/us1.wav

It sound more natural, but when i use the same sound with gespeaker i am not 
getting same quality sound. Any idea why this happens ?

Original comment by flashweb@gmail.com on 11 Jun 2010 at 10:18

[GoogleCodeExporter]

you can only set the proper advanced settings for the mbrola voice.
there's no other choice, neither in mbrola nor in gespeaker.

Original comment by mufl...@vbsimple.net on 12 Jun 2010 at 11:00

• Added labels: Milestone-Later, Priority-Low
• Removed labels: Priority-Medium

[GoogleCodeExporter]

Original comment by mufl...@vbsimple.net on 12 Jun 2010 at 11:00

• Added labels: Type-Enhancement
• Removed labels: Type-Defect