[stuCONNERS]

one more thing before i look into this, would i be ok with getting a usb microphone thats supported by linux. Do i need anymore hardware?

[Kempniu]

one more thing before i look into this, would i be ok with getting a usb microphone thats supported by linux.

The microphone you're going to use has to work with the audio software you're going to use (ALSA or PulseAudio, depending on the version of pocketsphinx installed). In other words, if the microphone works with other utilities in your system, you should be fine.

Do i need anymore hardware?

No. Hardware-wise, you don't need anything more than a microphone.

[Kempniu]

Name migration to kodivc has been completed. I have both good and bad news. The bad news is I decided to stop creating PPA packages for anything before Ubuntu 14.04 Trusty Tahr (so XBMCbuntu Gotham is still fine, but not XBMCbuntu Frodo). I wanted to finally break free from the outdated dhuggins CMU Sphinx PPA for Ubuntu 10.04. Also, Launchpad will soon disable building packages for Lucid (as its end-of-support date is nearing), which would have complicated things even more, so I decided that this is the right thing to do. The good news is that for those of you that still want to use kodivc on older releases (but hey, you can't even download XBMCbuntu Frodo ISO from the official Kodi website anymore...), you can either follow the old installation instructions to install from the xbmcvc PPA (I left it intact, but no more updates will be released there) or install the new version the hard way, i.e. compiling from Git source. Note that older XBMC versions are still supported by the software itself. Changes described above only apply to the PPA.

The only major difference between the old version and the current one is that the unlock command has been changed from XBMC to KODI. Packages are ready for Ubuntu 14.04 (Trusty Tahr), 14.10 (Utopic Unicorn) and 15.04 (Vivid Vervet). Updated installation instructions can be found, as usually, on the project's homepage. If you have any questions, issues, suggestions or complaints, you're welcome to post them here or on GitHub.

[stuCONNERS]

would this microphone work ok?
http://www.amazon.co.uk/Kinobo-Microphon...B00NSOWWIS

I cant seem to find any that work with ALSA. Also is it possible to say a word and have load a button mapped on keymap.xml

eg Say "Codec" and it simulates "o" being pressed which is codec info. Sorry if it sounds like im going on but im quiet interested in this.

[Kempniu]

would this microphone work ok?
http://www.amazon.co.uk/Kinobo-Microphon...B00NSOWWIS

I cant seem to find any that work with ALSA.

I'm going to reiterate what I've written in my previous response: if it works with ALSA or PulseAudio, you'll be fine. I have no idea if this particular microphone is any good, though. Google and/or ALSA/PulseAudio mailing lists might.

Also is it possible to say a word and have load a button mapped on keymap.xml

eg Say "Codec" and it simulates "o" being pressed which is codec info. Sorry if it sounds like im going on but im quiet interested in this.

The example you provided can be achieved using the Input.ShowCodec JSON-RPC method, so that's no biggie, I can add it if you want. For other needs, there is the Input.ExecuteAction method which can do quite a lot, but probably not everything imaginable.

[ysilvela]

very interesting proyect.

[Hedda]

Would be very cool if the search function for this addon could be enhanced by using Web Speech API via plugin modules for Zypr, Google Speech, or similar online cloud services for advanced advanced speech recognition powered by by third-parties?

https://dvcs.w3.org/hg/speech-api/raw-fi...chapi.html

Being able to leverage powerful third-party online cloud services for advanced speech recognition instead of having to rely on offline speech recognition would take this to the next level?

Maybe then it would be possible to enable enhanced keyword searches instead of having to spell each letter just to do an searches as suggested here

http://forum.kodi.tv/showthread.php?tid=199486

Many remote controls have microfone. Built-in microfone. So why we dont use them with XBMC? Example for my suggest:
image



This screen maybe work speech to text? ı think this is usefull feature. Some think this is how work.

Microfone--> xbmc --> Google speech to text api --> Text ->> Text in textbox the end.

Bump this now that both Amazon Fire TV and Google Nexus Player (Android TV) have good proof-of-concept for these.

This of course relate both directly and indirectly to this other old thread about voice recognition and control

http://forum.kodi.tv/showthread.php?tid=9280

Now days there are also ways of accessing third-party open APIs online such as Google Speech and Apple HomeKit / Siri.

There is also Zypr API which aggregates many of those third-party APIs for voice commnads https://www.zypr.net

Again a proof-of-concept is Amazon Fire TV which have a very good "Voice Search" feature which shows how voice command adds accessibility to the search function in a media player

Don't get me wrong, offline speech recognition for basic voice commands still have its place for simple control usage, but it can not compete when doing advanced keyword searching.

Here is an article which lists some pros and cons of offline verses online speech recognition http://cmusphinx.sourceforge.net/2015/02...d-smarttv/

[Kempniu]

Thank you for your input. These are undoubtedly interesting ideas. Yet kodivc is so tightly coupled with CMU Sphinx that it doesn't even implement continuous listening on its own, instead relying on pocketsphinx to do it. Which means that in order for any kind of hot-switchable voice recognition plugins to work with kodivc one would first need to either:

• implement a continuous listening algorithm which would tell speech and silence apart (and one that would work with both ALSA and Pulseaudio) and then hand captured utterances over to external speech recognition APIs,
  
• get a good understanding of how pocketsphinx works so that it can still be used for silence detection etc. while being able to process captured utterances independently.
  

The second option seems more viable. I might look into it, but please don't expect any immediate results as the spare time I can nowadays spend daily on experimenting with a microphone is very limited (minutes rather than hours).

[RockerC]

Not sure if this is still being developed but if it is then this project should probably be renamed to meet the official trademark policy?

http://kodi.wiki/view/Official:Trademark_Policy

You should not use "Kodi" or "XBMC" in the product name, see this section:

http://kodi.wiki/view/Official:Trademark...d_Software

Related Software

XBMC Foundation products are designed to be extended, and we recognize that community members writing add-ons and supplemental software need some way to identify the XBMC Foundation product to which their add-on/software pertain. Our main concern about add-ons and related software is that consumers not be confused as to whether they are official (meaning approved by XBMC Foundation) or not. To address that concern, we request that add-ons and supplemental software names not include, in whole or in part, the words "XBMC Foundation" or "Kodi" in a way that suggests a connection between XBMC Foundation and the add-on or software (e.g., "Frobnicator for Kodi," would be acceptable, but "Kodi Frobnicator" would not).

[Kempniu]

Thank you for your remark. While I'm still subscribed to this thread, the program itself is not really actively developed (10 commits in the past 2 years, none of which extends functionality in any way) and, unfortunately, changing its name is quite a bit of work which I don't have the time to do, at least not immediately... I also don't have any ideas for an alternative name - the trademark policy leaves developers with virtually no way of associating their software with Kodi name-wise (I'm not going to ship a Linux binary called voice-commands-for-kodi, that's an abomination). Suggestions?

[[email protected]]

I tried to follow the instructions and use the git clone and make method, however I get Package pocketsphinx and sphinxbase was not found in the pkg-config path. I have installed pocketsphinx and sphinxbase and running grep shows they are both present. Any ideas on how I can use this method?

any help is appreciated.

[Kempniu]

You didn't mention which distribution you're using, but the most likely cause of your problems is that you haven't installed the development packages. On Ubuntu, you'll need:

• libsphinxbase-dev
  
• libpocketsphinx-dev
  
• libcurl4-openssl-dev
  

[[email protected]]

Thanks a million. that did the trick. I am running ubuntu 14.04. I did have to install pocketsphinx-hmm-en-hub4wsj separately for kodivc to run.

Is it possible to add new commands such as search or find? I have looked at the directory structure and as nearly as I can tell there are files under the model folder which seem to hold the phonetic spelling of words. Do you have a quick how-to on this?

[Kempniu]

Is it possible to add new commands such as search or find?

What is it specifically that you would like kodivc to do when it hears such a command?

I have looked at the directory structure and as nearly as I can tell there are files under the model folder which seem to hold the phonetic spelling of words. Do you have a quick how-to on this?

Sure, you can add words in there, but you'd still have to hook them up in the program itself, otherwise you'll just get Unknown action messages for these new commands. Just let me know what you would like to be added and we'll go from there.

[[email protected]]

Ultimately I would like to be able to search my movie library. So perhaps issue a voice command to open up a search window and then go to spelling mode, spell out the movie and after hearing search, a search string would be passed to kodi and the results returned. Is this possible?