[arsivci]

Hello, while configuring my Kodi subtitle features, I noticed that the KSR does not read the multichoiced selections. Under settings/subtitle menu, we could choose more than one language for which the subtitle could be downloaded. When pressed enter located on each language , the selection turns into green, if I am not mistaken. Yet the screen reader does not notify us if the selection is added or removed. As another problem, in some addons, before you enter there seems to be license agreement dialogues that we need to accept in order to login into the addon. Never the less the Kodi Screen Reader does not notice the Agree/Disagree controlls.

I've got a fix for multi-select dialogs that will be in the next update. Please PM me the name of an addon with an agree/disagree control.

Hi again, I noticed a small issue that Kodi Screen Reader could not reach all the available subtitle languages under download subtitle menu in video osd menu. İt tells the amount of languages but only the first option/language could be clicked. I write this feed-back of course after setting my default tv and movie subtitle service and after arranging my subtitle languages.

I can see what's going on. I'll fix it. You can actually select all of the languages, but if all of the options are variations on the same language, it won't read them because it is only using the language name to detect if the selected item has changed. I just need to have it read and compare the other info.

Thanks for your consideration and PM got sent.

[arsivci]

Hi again, I think that I have to report another problem. When I open my kodi, I heare that addon failure/error sound, possibly an addon I installed is incompatible or it is another Kodi error, yet I cannot get what it is, because the error is not red by the reader. I use the latest version 14.0 Helix on Windo 7 64bits machine.

[arsivci]

The specific addon that you are using has caused me nothing but troubles since Kodi alpha began. I quite using it because the developer refuses to really fix anything, Even now after the official release of Kodi, the only thing that gets it working it all is a bandaid others have slapped on it. Even with those band aids don't fully fix it and I have found they cause me other issues in general usage. Things like high memory usage, Script errors, crashing and such. Soon as I removed it and went on to other things I quite having those issues.

With that said, Same Windows 7 x64 and Kodi and I'm not seeing the issue you are, With the exception of the already known issues TTS and Kodi have been working on all my systems flawlessly. However a log would probably help us figure out how to help you with your issue.

Ruuk.
That is ok, I got my answer with regard to that addon from Ruuk.

---

The problem I describe is different. That is, the Kodi Screen Reader does not read the error notifications in case of an addon failure either because of compatibility or due to another reason.

[arsivci]

Do you have the Speak Background Progress Updates setting enabled? Wonder if that might read the error notifications? Im not sure as I have never actually enabled the setting myself.

Yes, it is enabled. But it still does not read addon failure errors or other kind of error notifications. I hear the sound and understand that something goes wrong, but do not know what it is, since it does not read outloud.

[ruuk]

Hi again, I think that I have to report another problem. When I open my kodi, I heare that addon failure/error sound, possibly an addon I installed is incompatible or it is another Kodi error, yet I cannot get what it is, because the error is not red by the reader. I use the latest version 14.0 Helix on Windo 7 64bits machine.

Error notifications are not something I can have the addon read currently, but it may be possible for me to come up with a workaround. I'll have to look into it when I get a chance.

[eckythump]

Hi all!

Been a while, glad to see this is still quite active and things are moving forward. Congrats on becoming a member of team Kodi, ruuk. I'm glad all your hard work and talent has been officially recognised and rewarded.

I managed to kill my old, updates-disabled, ain't-broke-don't-fix-it raspberry pi and now that I'm trying to use the new version, I've noticed a couple of issues.

My perl speech server no longer works with it. I'm not particularly bothered by this, as it's easy enough to run the python one, too, but probably worth either fixing it, or removing support for it entirely

With the minimal testing I've done, the python one seems to work fine for streaming wav's back to the client, but I can't get the "speak on server" feature working, either from XBMC itself, or attempting to manually interact with the server using python.

ruuk: Can you give me a crash course on how the client communicates with your speech server? Is it just a case of POST'ing engine, voice, rate, text, pitch and volume to either /wav or /say ? Which, if any, of those variables is optional? Using a python shell, I can get the server to throw a wav at me if I post to /wav, but /say does nothing, which no speech eminating from my XP machine that's running the server. The logfile shows:

Code:

speech.server: Mode: WAVOUT
speech.server: [127.0.0.1] SAPI - SAY: one two three
speech.server: Mode: WAVOUT
speech.server: [127.0.0.1] SAPI - WAV: one two three

I'm guessing the mode being WAVOUT for SAY isn't supposed to happen.

Also, does the python speech server save settings at all, or will it happily handle multiple unique clients with unique preferences? I'm hoping the latter, but I'm a little overwhelmed by the code at present as I've not been paying attention and it's been a few months.

Hope everyone is well. Thanks to everyone who's helping, testing and contributing. I'm going to try and be around a bit more again and help out where possible.

[ruuk]

My perl speech server no longer works with it. I'm not particularly bothered by this, as it's easy enough to run the python one, too, but probably worth either fixing it, or removing support for it entirely

I'm fine either way. If you want me to fix the support I'll look into what's wrong.

With the minimal testing I've done, the python one seems to work fine for streaming wav's back to the client, but I can't get the "speak on server" feature working, either from XBMC itself, or attempting to manually interact with the server using python.

You found a bug. I have fixed it and updated the version on the downloads page.

ruuk: Can you give me a crash course on how the client communicates with your speech server? Is it just a case of POST'ing engine, voice, rate, text, pitch and volume to either /wav or /say ? Which, if any, of those variables is optional? Using a python shell, I can get the server to throw a wav at me if I post to /wav, but /say does nothing, which no speech eminating from my XP machine that's running the server. The logfile shows:

Code:

speech.server: Mode: WAVOUT
speech.server: [127.0.0.1] SAPI - SAY: one two three
speech.server: Mode: WAVOUT
speech.server: [127.0.0.1] SAPI - WAV: one two three

I'm guessing the mode being WAVOUT for SAY isn't supposed to happen.

Also, does the python speech server save settings at all, or will it happily handle multiple unique clients with unique preferences? I'm hoping the latter, but I'm a little overwhelmed by the code at present as I've not been paying attention and it's been a few months.

You can use /wav or /say without any arguments, and it will either auto-select a backend or use the backend last used during the current session. You can have multiple unique clients, though I haven't done much testing with this. If you have different clients use a different engine, it will switch the backend each time the other client makes a request, I don't think this will cause any issues, but I'm sure it will perform better if they all use the same backend.

The problems you had with /say are likely due to the bug I fixed.

As for the 'Mode: WAVOUT' being displayed when using /say, I think it is in fact using WAVOUT and playing the wav on the server. I can't remember if this was intentional or not as I haven't worked on the server in a while. Most likely it was just easier to use the same mode for /wav and /say and I never got around to adding direct output. I guess that's something for the future

[ruuk]

I really like having the speech server available. I mean even if you just left it alone for someone else to fix or something or as a hidden feature or such. It kinda keeps the door open for ANY system to run Kodi. I have experimented with it and even used it for a bit in the past and had things in mind to play with it in the future. I hope you don't decide to remove it.

Don't worry. I was talking about support for eckythump's perl server, not my python server

[eckythump]

Hi,

I've tried the speech.server again, and sadly now it doesn't work for me at all. When I post to /say I get nothing, and I'm not even seeing anything logged, and when I post to /wav I get a 500 internal server error.

I'm using XP and python 2.7. Is there anything I should check? The server has definitely worked in the past (though I never used the speak-on-server option).

In the meantime, if you can have a looksee and try to ascertain why the perl server no longer works, that'd be good. Also, the speed setting is ghosted out when the perl server is selected in the last few newish releases I've tried. Be good if you can fix that, too, if it's still an issue.

Thanks again for all your work. I've just made a small donation to the project.

Do you have a TODO list? If you're able to post it somewhere, I'd like to try and do my bit to clear some of the entries from it for you. I may also be able to enlist the help of others who wouldn't have much interest in the project, but would probably enjoy the challenges it offers.

Thanks again!

[eckythump]

Haven't had a chance to do any debugging on the python speech server, but I looked into the issues with the perl server.

The issue related to the perl server not replying to invalid URL's with a 404, but instead it was sending a helper form with a 200 response code. I've made it now return a 404 and commited this change to git, so now it works again.

One issue I did find, though, is that the settings allowed me to set the speed to > 10. The perl server accepts speed rates between -10 and 10 and if you go outside of that, it's not very forgiving.

If you're able to add in a constraint to that range when perlServer is True, that'd probably be a good idea.

[ruuk]

First off, sorry for taking so long to reply. I've been extraordinarily busy as of late. I also have the bad tendency when I'm busy of waiting to post a reply until I have more time when if I can't reply quickly, and then forgetting to do it later.

Hi,

I've tried the speech.server again, and sadly now it doesn't work for me at all. When I post to /say I get nothing, and I'm not even seeing anything logged, and when I post to /wav I get a 500 internal server error.

I'm using XP and python 2.7. Is there anything I should check? The server has definitely worked in the past (though I never used the speak-on-server option).

In the meantime, if you can have a looksee and try to ascertain why the perl server no longer works, that'd be good. Also, the speed setting is ghosted out when the perl server is selected in the last few newish releases I've tried. Be good if you can fix that, too, if it's still an issue.

I'll have to give it a go on my XP VM.

Thanks again for all your work. I've just made a small donation to the project.

Thank you, that's very kind of you

Do you have a TODO list? If you're able to post it somewhere, I'd like to try and do my bit to clear some of the entries from it for you. I may also be able to enlist the help of others who wouldn't have much interest in the project, but would probably enjoy the challenges it offers.

The changelog has a TODO and there is also a file in the addon todo.txt which mostly has notes on Traker1001's suggestions. Otherwise, I've just been finding or hearing about what isn't speaking and fixing those as I find them.

Haven't had a chance to do any debugging on the python speech server, but I looked into the issues with the perl server.

The issue related to the perl server not replying to invalid URL's with a 404, but instead it was sending a helper form with a 200 response code. I've made it now return a 404 and commited this change to git, so now it works again.

One issue I did find, though, is that the settings allowed me to set the speed to > 10. The perl server accepts speed rates between -10 and 10 and if you go outside of that, it's not very forgiving.

If you're able to add in a constraint to that range when perlServer is True, that'd probably be a good idea.

I'll do that when I get a chance.

[ruuk]

Just a quick note on for the broken keyboard dialog. I've got a pull request for a fix and mkortstiege helped me clean it up

I'm just awaiting other eyes to take a look at it to see if it's fine to merge into master.

This fix also lays some groundwork for accessing the labels of some other controls that the addon currently can't.

[byron27]

What do you guys think'll happen if I got either a Raspberry Pi 2 or a Fire TV or Fire Stick TV? Would the TTS work out of the box without an external server in any of these cases? If not, what set-top box would be best for built in speech? I really don't wanna have to run a server just to generate speech files. Thanks!

[ruuk]

The addon will not work "out of the box" with any Android devices yet. It should work with OpenElec installed on a Raspberry Pi 2. I have one on order, so I'll be able to report directly on this as soon as I get it. The original Raspberry Pi was disappointing as a speech enabled Kodi device, but I have high hopes that the Pi 2 will be the go to low cost Kodi device for speech. Some of the team members already have one and initial reports seem encouraging.

[byron27]

I appreciate the quick reply, I'll be anxious to hear your findings! Thank you!