Release - XBMC TTS (Text To Speech) [Alpha]

  Thread Rating:
  • 6 Votes - 4.5 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Post Reply
ruuk Offline
Team-Kodi Member
Posts: 1,587
Joined: Aug 2005
Reputation: 185
Location: Bremerton, WA USA
Post: #16
Thanks for taking the time to look at the code Smile

Are you talking about actually adding new methods to the monitor class?

(2014-06-06 18:42)Martijn Wrote:  Implementation would be something like this https://github.com/xbmc/xbmc/pull/4766/files

Or are you only talking about triggering an onNotification()?

So insert something similar to this:

Code:
CAnnouncementManager::Get().Announce(GUI, "xbmc", "AddonUpdatesAvailable");
Probably with a different flag than GUI

After this:
(2014-06-06 18:42)Martijn Wrote:  and i think relevant file to add it to would be
https://github.com/xbmc/xbmc/blob/master...y.cpp#L319 (some where after this line of code. next line does the GUI popup).

[Image: sig.jpg]
find quote
Traker1001 Offline
Senior Member
Posts: 134
Joined: Jun 2013
Reputation: 3
Post: #17
does openelec have espeak installed or even Mpg123? Also correct me if I am wrong, But Openelec is still using PulseAudio which doesn't play well. I got Openelec working for a while after alot of messing around, but not well. I ended up going with RaspBMC and installing Alsa and Espeak and it works very well.

@Edit, On second thought, I recall Openelec did work for me on default with Espeak, Again not very well because of the Pulseaudio. But I never got it working constantly with anything else.

for your Openelec do you have your background sounds turned on?
(This post was last modified: 2014-06-07 05:27 by Traker1001.)
find quote
RiChaDo Online
Junior Member
Posts: 48
Joined: Nov 2010
Reputation: 0
Post: #18
no i dont have background sounds turned on..ill give it a try

[Image: widget]
[Image: widget]
find quote
ruuk Offline
Team-Kodi Member
Posts: 1,587
Joined: Aug 2005
Reputation: 185
Location: Bremerton, WA USA
Post: #19
(2014-06-06 15:37)RiChaDo Wrote:  I got openelec but after installing this addon what todo next? because It's not doing anything
In the addon settings, are you able to select eSpeak under 'Default TTS Engine'?

[Image: sig.jpg]
find quote
ruuk Offline
Team-Kodi Member
Posts: 1,587
Joined: Aug 2005
Reputation: 185
Location: Bremerton, WA USA
Information  v0.0.70
Post: #20
Added a new version to my repository: 0.0.70.

Get it or the repository from the Downloads Page.

Changes:
  • Now speaks when video or music database scan has started or finished
  • Improved speaking of window names including speaking the addon name for addon windows
  • Added code to load snd-bcm2835 module for other Raspberry Pi distributions
  • Will now check for and load snd-bcm2835 module for all players/engines if needed
  • Fix for sound icons failing due to file paths with non-ascii characters

[Image: sig.jpg]
find quote
ruuk Offline
Team-Kodi Member
Posts: 1,587
Joined: Aug 2005
Reputation: 185
Location: Bremerton, WA USA
Information  v0.0.71
Post: #21
Added a new version to my repository: 0.0.71.

Get it or the repository from the Downloads Page.

Changes:
  • Now shows the busy dialog when waiting for the engines, voices or players list to display (useful on the Raspberry Pi, which can be slow to display)
  • Fixed a bug where changing from a player that could pipe to one that could not would cause speech to stop

[Image: sig.jpg]
(This post was last modified: 2014-06-09 01:35 by ruuk.)
find quote
ruuk Offline
Team-Kodi Member
Posts: 1,587
Joined: Aug 2005
Reputation: 185
Location: Bremerton, WA USA
Post: #22
FYI: The fix I submitted for textboxes was merged into xbmc (with help from jmarshall) and should go into the next Gotham release. This will get OK, Yes/No and progress dialogs speaking for Confluence and other skins using the new dialog format, not to mention all the other textboxes that will now be accessible.

[Image: sig.jpg]
(This post was last modified: 2014-06-11 00:13 by ruuk.)
find quote
ruuk Offline
Team-Kodi Member
Posts: 1,587
Joined: Aug 2005
Reputation: 185
Location: Bremerton, WA USA
Information  v0.0.72
Post: #23
Added a new version to my repository: 0.0.72.

Get it or the repository from the Downloads Page.

Changes:
  • Update Comtypes to dev version to remove NumPy dependency
  • SAPI can now speak with XBMC audio which is now the default
  • Added SAPI setting 'Play Speech Via XBMC'

Windows users: You will want to disable the 'Play Speech Via XBMC' option if you don't have an esoteric audio setup. I added this and changed the default because on some setups when XBMC is running it takes over the audio. Playing audio through XBMC is the safe choice, so it is the default, but it is not the best choice.

[Image: sig.jpg]
find quote
Traker1001 Offline
Senior Member
Posts: 134
Joined: Jun 2013
Reputation: 3
Post: #24
(2014-06-11 02:06)ruuk Wrote:  Update Comtypes to dev version to remove NumPy dependency

the update for Comtypes to remove NumPy dependency, Is that just for the addon, Or did you do that to the http server as well?

Man your awesome. I'm currently setting this up for a couple friends, and they are loving it.
(This post was last modified: 2014-06-11 16:28 by Traker1001.)
find quote
ruuk Offline
Team-Kodi Member
Posts: 1,587
Joined: Aug 2005
Reputation: 185
Location: Bremerton, WA USA
Post: #25
(2014-06-11 15:44)Traker1001 Wrote:  
(2014-06-11 02:06)ruuk Wrote:  Update Comtypes to dev version to remove NumPy dependency
the update for Comtypes to remove NumPy dependency, Is that just for the addon, Or did you do that to the http server as well?
It will be as soon as I update the server.
(2014-06-11 15:44)Traker1001 Wrote:  Man your awesome. I'm currently setting this up for a couple friends, and they are loving it.
Thanks. That's good to hear Smile

[Image: sig.jpg]
find quote
ruuk Offline
Team-Kodi Member
Posts: 1,587
Joined: Aug 2005
Reputation: 185
Location: Bremerton, WA USA
Information  speech.server v0.0.7
Post: #26
The speech server has been updated on the Downloads Page.

It includes the updated comtypes which no longer depends on numpy, although I imagine numpy may possibly improve performance.
It also fixes an issue with the server crashing on exit due to a problem when closing the SAPI backend.

[Image: sig.jpg]
find quote
ruuk Offline
Team-Kodi Member
Posts: 1,587
Joined: Aug 2005
Reputation: 185
Location: Bremerton, WA USA
Information  v0.0.73
Post: #27
Added a new version to my repository: 0.0.73.

Get it or the repository from the Downloads Page.

Changes:
  • JAWS and NVDA will now fallback to SAPI if not running
  • Added ability to speak text in the text viewer window using the future xbmc ability to get textbox text directly
  • Fix for some multiline texts not having formatting codes removed before being spoken

[Image: sig.jpg]
find quote
Traker1001 Offline
Senior Member
Posts: 134
Joined: Jun 2013
Reputation: 3
Post: #28
Well, I coulda sworn I posted something the other day. But I guess I never hit submit or something.

Anyways, In regards to

Quote:JAWS and NVDA will now fallback to SAPI if not running

I'm confused, I thought it already did this?
Quote:Added ability to speak text in the text viewer window using the future xbmc ability to get textbox text directly
Is this something that will be available in helix?
Quote:Fix for some multiline texts not having formatting codes removed before being spoken
Big Grin Actually recognized this happening, but didn't realize it was a bug in speech, thought it was just mis-formatting.

With the Speech Volume up and down, It'd be awesome if it had a key commands so that I can assign it to remote control key.

Also with the Speak extra information setting, I know you added a touch of a delay between main line item and the extra info, however it would be nice if that were customizable setting, So I.E. I could specify Read the main item but wait 20 second or 10 seconds or such then read the extra info based on how I scroll though things. Amusingly, some of the sighted people I know have who have generally no use for tts have actually liked that feature. Smile

Also I have really taken a liking to XbmcUbuntu setup, and am planning on purchasing Cepstral for it. I know you have made improvements, but was wondering how well Cepstral is working with the add-on, Mainly responsiveness wise. Of course last time I messed with it being on the Pi, which I have now retired to other duties.

Side note, We are loving some of the latest improvements, Percentages for video window, Window names, Video database speaking and such. Just when I think the addon couldn't possibly get any better you add something really neat to the mix. Big Grin
(This post was last modified: 2014-06-16 07:44 by Traker1001.)
find quote
ruuk Offline
Team-Kodi Member
Posts: 1,587
Joined: Aug 2005
Reputation: 185
Location: Bremerton, WA USA
Post: #29
(2014-06-16 07:43)Traker1001 Wrote:  Well, I coulda sworn I posted something the other day. But I guess I never hit submit or something.

Anyways, In regards to

Quote:JAWS and NVDA will now fallback to SAPI if not running

I'm confused, I thought it already did this.
I added the framework for falling back, but had only implemented it for the server backend.
(2014-06-16 07:43)Traker1001 Wrote:  
Quote:Added ability to speak text in the text viewer window using the future xbmc ability to get textbox text directly
Is this something that will be available in helix?
Should be in the Gotham 13.2.
(2014-06-16 07:43)Traker1001 Wrote:  
Quote:Fix for some multiline texts not having formatting codes removed before being spoken
Big Grin Actually recognized this happening, but didn't realize it was a bug in speech, thought it was just mis-formatting.

With the Speech Volume up and down, It'd be awesome if it had a key commands so that I can assign it to remote control key.

Also with the Speak extra information setting, I know you added a touch of a delay between main line item and the extra info, however it would be nice if that were customizable setting, So I.E. I could specify Read the main item but wait 20 second or 10 seconds or such then read the extra info based on how I scroll though things. Amusingly, some of the sighted people I know have who have generally no use for tts have actually liked that feature. Smile
Will do.
(2014-06-16 07:43)Traker1001 Wrote:  Also I have really taken a liking to XbmcUbuntu setup, and am planning on purchasing Cepstral for it. I know you have made improvements, but was wondering how well Cepstral is working with the add-on, Mainly responsiveness wise. Of course last time I messed with it being on the Pi, which I have now retired to other duties.
It's hard to say as I only have the demo version. It's slower than some but probably comparable to Windows 8 and sapi. You can try the demo voices if you want a rough idea. If you purchase it, I'd love to hear how it performs with the actual version.
(2014-06-16 07:43)Traker1001 Wrote:  Side note, We are loving some of the latest improvements, Percentages for video window, Window names, Video database speaking and such. Just when I think the addon couldn't possibly get any better you add something really neat to the mix. Big Grin
I'm glad the things I'm adding are actually useful Smile

[Image: sig.jpg]
find quote
ruuk Offline
Team-Kodi Member
Posts: 1,587
Joined: Aug 2005
Reputation: 185
Location: Bremerton, WA USA
Information  v0.0.75
Post: #30
Added a new version to my repository: 0.0.74.

Get it or the repository from the Downloads Page.

Changes:
  • Fix for window xml parser failing on skins with non-standard folder paths
  • Now speaks new versions for all addon updates
  • Added the sub-setting 'Delay (seconds)' to the 'Speak Extra Item Information Automatically' setting
  • Now says 'sub-setting' before speaking a sub-setting's name
  • Now speaks the category description as extra info on the settings category selection window
  • Added the ability to change backend speech volume with ctrl-plus and ctrl-minus

I'll probably work on the ability to customize keymap keys next. For now if you want to assign the the volume up/down to a different key/button, you can edit the addon keymap and change the key to something else. If you want to set it to a remote button you will have to copy the relevant lines to a remote tag instead of within the keyboard tag.

[Image: sig.jpg]
find quote
Post Reply