Alright, So I have gone ahead and requested a xbmc wiki account. Once I get in, Ill be posting the information I have compiled from the thread there. What do you think about me starting a new thread in the Addons area with the information as well.
I finally got my pi running halfway seemingly decently. I found small bug in the Openelec on the Pi. If the Nav sounds are turned off in the skin settings menu, Speech doesn't work. And Play speech directly via eSpeak seems to kill speech altogether no matter what I do. It doesn't throw any errors at all in the log though, speech just stops, Log shows like it should be speaking.
Something else I have found that I can replicate.
Goto Video Addons, Get More, Scroll down to "Arrêt sur Images" and speech has crashed with an error in the log.
Code:
04:35:30 T:4080 ERROR: Exception in thread TTSThread: SAPI:
Traceback (most recent call last): File "C:\Program Files\XBMC\system\python\Lib\threading.py", line 808, in __bootstrap_inner self.run()
File "C:\Program Files\XBMC\system\python\Lib\threading.py", line 761, in run self.__target(*self.__args, **self.__kwargs)
File "C:\Users\Admin\AppData\Roaming\XBMC\addons\service.xbmc.tts\lib\backends\base.py", line 214, in _handleQueue self.threadedSay(text)
File "C:\Users\Admin\AppData\Roaming\XBMC\addons\service.xbmc.tts\lib\backends\sapi.py", line 44, in threadedSay ssml = self.baseSSML.format(saxutils.escape(text))
UnicodeEncodeError: 'ascii' codec can't encode character u'\xea' in position 3: ordinal not in range(128)
Interestingly Enough this doesn't crash with using NVDA as speech backend on windows, Only Sapi directly and On the PI with Openelec and Espeak it crashes.
I have not been able to get the Python Speech server running as of yet to give that a try with that.
Probably a useless bit of info here, however.
I can tell you Sending "Arrêt sur Images" to SAPI via Say command from the following site does work.
http://krolik.net/post/Say-exe-a-simple-...ndows.aspx
(2014-05-12, 03:50)Traker1001 Wrote: [ -> ]However, I'm getting crashing (TTS stays started, But no more speech, Have to disable and Re-enable to get it speaking again). Haven't had chance to browse through a large log, However, I gleened this with every crash.
Quote:UnicodeEncodeError: 'ascii' codec can't encode character u'\xf3' in position 18: ordinal not in range(128)
I'll have to fix this. Somewhere encoding is going wrong.
(2014-05-12, 03:50)Traker1001 Wrote: [ -> ]And I also saw this error a few time, however, doesn't seem to cause it to crash.
Quote:20:32:50 T:4900 ERROR: Control 61 in window 10147 has been asked to focus, but it can't
You can safely ignore theses messages. They are usually caused because a list is created empty and can't get focus because it doesn't have items yet. It doesn't really affect much.
(2014-05-12, 03:50)Traker1001 Wrote: [ -> ]Does the "Enable debug logging" Create a separate log somewhere, Or just more detailed log in the xbmc.log?
More detailed log in xbmc.log.
(2014-05-12, 05:32)Traker1001 Wrote: [ -> ]Couple things that would be cool in the PVR Guide. Is if it read the date of the selection, even better if the reading was limited to day/month, Even cooler if it recognized the date of current selection and only read the date if it didn't = current date. And if it read the at least the Time and date even if extra info item wasn't selected.
Another neat item, would be ability to enable/disable the automatic read extra info via keystroke. I find there are some areas where it is handy and others where it isn't.
And one other thought would be some kind of acknowledgment when toggle is hit, Beep, speech saying item on or off, something of that sort.
Added to my todo list.
(2014-05-13, 03:09)Traker1001 Wrote: [ -> ]Alright, So I have gone ahead and requested a xbmc wiki account. Once I get in, Ill be posting the information I have compiled from the thread there. What do you think about me starting a new thread in the Addons area with the information as well.
I'll create a release thread for the addon soon, I want to be able to edit the initial post, then you can use the second post to post the info.
(2014-05-13, 12:40)Traker1001 Wrote: [ -> ]I finally got my pi running halfway seemingly decently. I found small bug in the Openelec on the Pi. If the Nav sounds are turned off in the skin settings menu, Speech doesn't work. And Play speech directly via eSpeak seems to kill speech altogether no matter what I do. It doesn't throw any errors at all in the log though, speech just stops, Log shows like it should be speaking.
That's not really a bug. By default with many speech engines the addon plays the speech using nav sounds functions. On the Pi with OpenElec, that is the only option. For the Pi there is no alsa enabled in the normal builds.
(2014-05-13, 03:09)Traker1001 Wrote: [ -> ]Something else I have found that I can replicate.
Goto Video Addons, Get More, Scroll down to "Arrêt sur Images" and speech has crashed with an error in the log.
Code:
04:35:30 T:4080 ERROR: Exception in thread TTSThread: SAPI:
Traceback (most recent call last): File "C:\Program Files\XBMC\system\python\Lib\threading.py", line 808, in __bootstrap_inner self.run()
File "C:\Program Files\XBMC\system\python\Lib\threading.py", line 761, in run self.__target(*self.__args, **self.__kwargs)
File "C:\Users\Admin\AppData\Roaming\XBMC\addons\service.xbmc.tts\lib\backends\base.py", line 214, in _handleQueue self.threadedSay(text)
File "C:\Users\Admin\AppData\Roaming\XBMC\addons\service.xbmc.tts\lib\backends\sapi.py", line 44, in threadedSay ssml = self.baseSSML.format(saxutils.escape(text))
UnicodeEncodeError: 'ascii' codec can't encode character u'\xea' in position 3: ordinal not in range(128)
Interestingly Enough this doesn't crash with using NVDA as speech backend on windows, Only Sapi directly and On the PI with Openelec and Espeak it crashes.
Probably a useless bit of info here, however.
I can tell you Sending "Arrêt sur Images" to SAPI via Say command from the following site does work.
http://krolik.net/post/Say-exe-a-simple-...ndows.aspx
Yeah as before this is some encoding issues I need to fix. Thanks for the log snippet though.
ruuk Wrote:I'll create a release thread for the addon soon, I want to be able to edit the initial post, then you can use the second post to post the info.
That's not really a bug. By default with many speech engines the addon plays the speech using nav sounds functions. On the Pi with OpenElec, that is the only option. For the Pi there is no alsa enabled in the normal builds.
Yeah as before this is some encoding issues I need to fix. Thanks for the log snippet though.
Awesome, Will do.
Other than something minor like that, is there any other benefits to going to a build with Alsa?
Do you want me to post any other log snippets for that error?
Your Awesome!
(2014-05-13, 20:53)Traker1001 Wrote: [ -> ]Other than something minor like that, is there any other benefits to going to a build with Alsa?
Well, alsa is new (I believe) in OpenElec for all platforms. For some reason they disable it for the Pi. I don't really know much about what's going on with this.
Currently, playing through XBMC seems to be broken. I contributed code to XBMC to keep it from caching wavs when using this method, but it still seems to consume memory with each wav played. I don't know what the issue is and haven't had time to look into it yet.
So basically, if you can avoid using XBMC to play the speech (ie via ALSA) then that would be better. Right now extended usage will eventually fill memory on the Pi.
Basically, the Pi is becoming a big Pain In The Ass for me
I have to keep doing special things just for the Pi, which is fine, but annoying
(2014-05-13, 20:53)Traker1001 Wrote: [ -> ]Do you want me to post any other log snippets for that error?
Not yet. You've shown me the direction to look.
ruuk Wrote:Currently, playing through XBMC seems to be broken. I contributed code to XBMC to keep it from caching wavs when using this method, but it still seems to consume memory with each wav played. I don't know what the issue is and haven't had time to look into it yet.
So basically, if you can avoid using XBMC to play the speech (ie via ALSA) then that would be better. Right now extended usage will eventually fill memory on the Pi.
Basically, the Pi is becoming a big Pain In The Ass for me
I have to keep doing special things just for the Pi, which is fine, but annoying
[quote='Traker1001' pid='1708839' dateline='1400007231']
Do you want me to post any other log snippets for that error?
Honestly, The Pi is cool and all, But I really don't see all that much benefit to it. It has been fun to play with and a nice challenge to setup for myself that likes a challenge, However, I am finding even the Celeron Nuc runs significantly better. When I get little more money, Ill probably be buying another one. Even Price wise a Celeron Nuc with 4gb memory and SSD HDD can be built for less than $200. The Pi w/ all fixins ran me right at $100. So Yeah Double the price, Half the work, quadruple the performance.
And I imagine that XBMC on the droid platform, just from general perspective, is better than both. I mean you can buy a dual core droid with hardware acceleration for less than $100 these days, can't you?
Then there is things like the what you run it on the Ouya, which read pretty cool. I realize its there is no TTS addon working for the droid, and likely wont be for a while, But I am just speaking from like a general perspective.
Personally, I am very happy and content to have a wonderfully working windows platform, and from experience, I can tell you it works excellently when one is not hosing with various messing around with xml files and other strange addons to see how far down the rabbit hole really goes, lol.
I hope to someday be able to contribute to this addon code wise some day, Im a ways into these online python lessons, So I decided to look at some of the code for the your addon and was like
, I still got much to learn, lol.
Hi together
and firstly let my thank you all for the great work on making xbmc accessible.
I am using xbmc on two raspberry pie under raspbmc.
Is there any chance to get tts working on this plattform or do I have to switch to another operating system for my pi?
And is there any step by step guide how I could set up tts on the pi? I read a lot of this thread but really am not sure what to do and how to do it.
I tried to install espeak and the addon a few months ago but my pi crashed - since that I just read what you where developing and was waiting until it sounds a bit more stable. But now I really would like to give it another try so hope to get some help to set it up.
thanks and keep on the great work.
cheers,
Martin
Hello, mtsonline:
There are few way to install it, Best I have found for the PI is as such:
Download Ruuk's repo at
http://ruuks-repo.googlecode.com/files/r...-1.0.0.zip
Place the zip somewhere on your network that you XBMC Pi can browse to.
Goto: System - Setting - AddOns
Select: Install from Zip
Browse to the location on the network you saved "ruuk.addon.repository-1.0.0.zip" and Select it.
Now you have installed Ruuk's Repo
Now Goto System - Settings - AddOns - Get Addons - Ruuks Addon Repo - Services - XBMC TTS, And Select it.
The Addon should install and enable itself immeadatily.
As far as testing goes, I have never installed RaspMC on my Pi, So I don't know how well, badly or if it works, Be sure to let us know.
(2014-05-13, 23:40)Traker1001 Wrote: [ -> ]Honestly, The Pi is cool and all, But I really don't see all that much benefit to it. It has been fun to play with and a nice challenge to setup for myself that likes a challenge, However, I am finding even the Celeron Nuc runs significantly better. When I get little more money, Ill probably be buying another one. Even Price wise a Celeron Nuc with 4gb memory and SSD HDD can be built for less than $200. The Pi w/ all fixins ran me right at $100. So Yeah Double the price, Half the work, quadruple the performance.
And I imagine that XBMC on the droid platform, just from general perspective, is better than both. I mean you can buy a dual core droid with hardware acceleration for less than $100 these days, can't you? Then there is things like the what you run it on the Ouya, which read pretty cool. I realize its there is no TTS addon working for the droid, and likely wont be for a while, But I am just speaking from like a general perspective.
That's how I feel.
(2014-05-13, 23:40)Traker1001 Wrote: [ -> ]Personally, I am very happy and content to have a wonderfully working windows platform, and from experience, I can tell you it works excellently when one is not hosing with various messing around with xml files and other strange addons to see how far down the rabbit hole really goes, lol.
I hope to someday be able to contribute to this addon code wise some day, Im a ways into these online python lessons, So I decided to look at some of the code for the your addon and was like , I still got much to learn, lol.
Since I never took any Python lessons of any kind, maybe my code is not the best example to look at
I like to think my lack of structured learning helps me think "outside the box". Of course if said box is in a dark dirty basement under a pile of trash...
(2014-05-14, 13:01)mtsonline Wrote: [ -> ]I am using xbmc on two raspberry pie under raspbmc.
Is there any chance to get tts working on this plattform or do I have to switch to another operating system for my pi?
In addition to Traker1001's instructions, I wanted to mention that on RaspBMC it would probably be beneficial to enable alas and use that for sound output, rather than the default which hijacks the XBMC sound effects interface to play speech.
I was able to get alsa working on RaspBMC with the following commands:
Code:
sudo apt-get install alsa-utils
sudo modprobe snd_bcm2835
(2014-05-14, 13:01)mtsonline Wrote: [ -> ]I tried to install espeak and the addon a few months ago but my pi crashed
There was a bug with XBMC on the Pi that was causing crashes when playing larger wavs through the SFX interface. This was likely the cause of crashes in your earlier attempts. This has been fixed in recent versions, and if you can get ALSA working it's no longer an issue.
I have been having an issue and I'm not sure how to document this other than to just say what's been happening. I haven't been able to reliably reproduce this issue and I couldn't say what if anything is causing it. Sometimes when using SAPI as my speech engine the speech will randomly stop working and restarting XBMC is the only way to get it back. Other times certain things won't read that SHOULD, like context menus, or the F6 menu. Does anyone else have this issue. I'm on Windows 7 64-bit and I believe I'm still running Frodo. I've been scared to upgrade to Gotham. Should I do that or stick with Frodo?
(2014-05-15, 01:17)byron27 Wrote: [ -> ]I have been having an issue and I'm not sure how to document this other than to just say what's been happening. I haven't been able to reliably reproduce this issue and I couldn't say what if anything is causing it. Sometimes when using SAPI as my speech engine the speech will randomly stop working and restarting XBMC is the only way to get it back. Other times certain things won't read that SHOULD, like context menus, or the F6 menu. Does anyone else have this issue. I'm on Windows 7 64-bit and I believe I'm still running Frodo. I've been scared to upgrade to Gotham. Should I do that or stick with Frodo?
There is a bug with non-ascii text and SAPI (and most of the backends) that I have just fixed and it will be in the next release. This is possibly what is causing the issue, as what you described is what happens with this bug. If the next release does not fix it, I will probably need a log to see what's up.
Gotham should be safe, but OK, YES/NO and progress dialogs will no longer speak their text depending on the skin. I've submitted some code for XBMC that will fix this plus make it possible to speak a lot of things that weren't accessible before. I'm working with one of the developers to get it in, and maybe I can get it into Gotham.
The dialogs didn't stop speaking because of a change in the XBMC code but because some of the skins changed to using textboxes in the dialogs, which have never been accessible to the addon.
Do you know if Confluence is still working properly?