Active Topics
-
Porting apps to Leste (32)
to Maemo 7 / Leste by Arno_11 - 6 days, 1 hr ago - more...
|
2012-06-20
, 16:42
|
|
Posts: 958 |
Thanked: 3,426 times |
Joined on Apr 2012
|
#242
|
Originally Posted by biketool
Thank you! I'll see what I can do to get this integrated.
Misiak has posted his progress on getting your headsets working for use with Saera in the thread linked below. Visiting, contributing any dbus technical help, testing, and even thanking those working will help bring this important feature to Saera. http://talk.maemo.org/showthread.php?t=84827&page=2
Originally Posted by ivgalvez
I did not; nobody's replied to me about that yet.taixzo: did you finally got permission to upload to extras-devel?
Originally Posted by marmistrz
That would be great!taixzo: if you still don't have the permission, I may upload saera from my garage account. And you'll upload stuff from your account as soon as you get permission.
@All regarding German translations:
I'm not fluent in German so I can't say which is the best, so I'll just take the most recent. I realize that times and numbers will need a bit of work to sound right.
Anyone interested in having Saera *understand* a different language: The first step would be translating corpus.txt to your language, from which I should be able to build a voice model.
__________________
Saera: A Siri clone for the N900, N9(50) and Jolla
Follow me on Twitter: twitter.com/taixzo
Saera: A Siri clone for the N900, N9(50) and Jolla
Follow me on Twitter: twitter.com/taixzo
| The Following 8 Users Say Thank You to taixzo For This Useful Post: | ||
|
|
2012-06-20
, 16:47
|
|
Posts: 293 |
Thanked: 163 times |
Joined on Jan 2012
@ beijing-islamabad
|
#243
|
anyone interested in translation of URDU ,pakistani language ? I guess that would be great if saera could do that only if some one manages to make an urdu translation ;p
taixzo we are waiting for a great great saera running flawlessly on our N900s . thanks buddy
please do consider my previous proposals too .
taixzo we are waiting for a great great saera running flawlessly on our N900s . thanks buddy
please do consider my previous proposals too .
|
|
2012-06-20
, 17:18
|
|
Posts: 804 |
Thanked: 1,598 times |
Joined on Feb 2010
@ Gdynia, Poland
|
#244
|
Originally Posted by taixzo
Give me a bit more time, I will come up with some clean Python code for you to use.
Thank you! I'll see what I can do to get this integrated.
|
|
2012-06-20
, 17:42
|
|
Posts: 3,328 |
Thanked: 4,476 times |
Joined on May 2011
@ Poland
|
#245
|
Originally Posted by taixzo
So send me the tar.gz source package and I'll upload it.
Thank you! I'll see what I can do to get this integrated.
I did not; nobody's replied to me about that yet.
That would be great!
@All regarding German translations:
I'm not fluent in German so I can't say which is the best, so I'll just take the most recent. I realize that times and numbers will need a bit of work to sound right.
Anyone interested in having Saera *understand* a different language: The first step would be translating corpus.txt to your language, from which I should be able to build a voice model.
My e-mail marmistrz at linux dot pl
| The Following 4 Users Say Thank You to marmistrz For This Useful Post: | ||
 |
|
2012-06-20
, 20:51
|
|
Posts: 7,074 |
Thanked: 9,069 times |
Joined on Oct 2009
@ Moon! It's not the East or the West side... it's the Dark Side
|
#247
|
Any plans to add a commando to open terminal and become root?
Possible to trigger any other application?
Possible to trigger any other application?
__________________
Do something for the climate today! Anything!
I don't trust poeple without a Nokia n900...
Do something for the climate today! Anything!
I don't trust poeple without a Nokia n900...
|
|
2012-06-20
, 20:55
|
|
Posts: 958 |
Thanked: 3,426 times |
Joined on Apr 2012
|
#248
|
Has anyone successfully tried to train the data-set in order to achieve a better recognition accuracy ?
Originally Posted by Dave999
I plan to allow for modules to be added. I might add that natively, but I don't want Saera to have a dependency on rootsh/sudser/whatever.
Any planned to add a commando to open terminal and become root?
__________________
Saera: A Siri clone for the N900, N9(50) and Jolla
Follow me on Twitter: twitter.com/taixzo
Saera: A Siri clone for the N900, N9(50) and Jolla
Follow me on Twitter: twitter.com/taixzo
| The Following 4 Users Say Thank You to taixzo For This Useful Post: | ||
|
|
2012-06-20
, 21:50
|
|
Posts: 7 |
Thanked: 17 times |
Joined on Jun 2012
|
#249
|
Hi taixzo,
that german translation is absolutely correct except two minor spelling errors. It should be "Wusstest Du, dass das N900 Mac OSX verwenden kann?" instead of "Wußtest Du, das das N900 Mac OSX verwenden kann?" but this does not influence espeak's pronunciation.
Let me tell, I also started developing a simple voice-control app for my N900 a few days before you started this thread. I don't even try to go towards AI. Currently I am pretty unhappy regarding the recognition accuracy and I am working on that part. One of the problems is, that pocketsphinx will always give you the best match it found, even when you said something (like "blahblahblah") that is not in your corpus/dictionary. pocketsphinx internally calculates a score but when using it with gstreamer there is no way (I know about) to get this value. I am a bit happy that I was able to patch the gst-plugin (gstpocketshinx) and to cross compile it for the N900. Since something like "blahblahblah" gives a significant lower score than a word which is in the corpus/dictionary this kind of "final results" can be sorted out.
This is a nice first step, but still I am trying to improve the accuracy. If anyone has managed to improve the recognition accuracy I would be really glad to get some help.
Greets
Last edited by myra; 2012-06-20 at 21:51. Reason: spelling error
that german translation is absolutely correct except two minor spelling errors. It should be "Wusstest Du, dass das N900 Mac OSX verwenden kann?" instead of "Wußtest Du, das das N900 Mac OSX verwenden kann?" but this does not influence espeak's pronunciation.
Let me tell, I also started developing a simple voice-control app for my N900 a few days before you started this thread. I don't even try to go towards AI. Currently I am pretty unhappy regarding the recognition accuracy and I am working on that part. One of the problems is, that pocketsphinx will always give you the best match it found, even when you said something (like "blahblahblah") that is not in your corpus/dictionary. pocketsphinx internally calculates a score but when using it with gstreamer there is no way (I know about) to get this value. I am a bit happy that I was able to patch the gst-plugin (gstpocketshinx) and to cross compile it for the N900. Since something like "blahblahblah" gives a significant lower score than a word which is in the corpus/dictionary this kind of "final results" can be sorted out.
This is a nice first step, but still I am trying to improve the accuracy. If anyone has managed to improve the recognition accuracy I would be really glad to get some help.
Greets
Last edited by myra; 2012-06-20 at 21:51. Reason: spelling error
|
|
2012-06-20
, 22:39
|
|
Posts: 958 |
Thanked: 3,426 times |
Joined on Apr 2012
|
#250
|
Originally Posted by myra
One thing that I hope for is runtime model/dictionary switching, to allow full dictation when sending e.g. a text. Hopefully I can use some gstreamer trickery (like using a filesink at the end of the pipeline), and re-try to understand what was said (with a larger dictionary/model) if the score is low...I hope there is some way to extract the score, if not I may have to dive into the plugin code and try to find a way to extract it.
One of the problems is, that pocketsphinx will always give you the best match it found, even when you said something (like "blahblahblah") that is not in your corpus/dictionary. pocketsphinx internally calculates a score but when using it with gstreamer there is no way (I know about) to get this value. I am a bit happy that I was able to patch the gst-plugin (gstpocketshinx) and to cross compile it for the N900. Since something like "blahblahblah" gives a significant lower score than a word which is in the corpus/dictionary this kind of "final results" can be sorted out.
This is a nice first step, but still I am trying to improve the accuracy. If anyone has managed to improve the recognition accuracy I would be really glad to get some help.
Greets
__________________
Saera: A Siri clone for the N900, N9(50) and Jolla
Follow me on Twitter: twitter.com/taixzo
Saera: A Siri clone for the N900, N9(50) and Jolla
Follow me on Twitter: twitter.com/taixzo
 |
| Tags |
| saera, speech-to-text |
«
Previous Thread
|
Next Thread
»
|
All times are GMT. The time now is 13:01.
If you want to support my work, you can donate by PayPal or Flattr
Projects no longer actively developed: here