Vox Commando - Voice Commands

[jitterjames]

phrase dont accept ├ñ,├Â,├╝,├ƒ (many german words use this)

does this version fix the problem?

[matt_jackson]

my maaan - I need more from u @ my work
fast, faster, James!

├ñ,├╝,├Â,├ƒ work with new editon.

computer generated (Text to Speech).

yes this is nice, I have make some songs with this plugin (text 2 speech)

@ phrase after comma space?
I don't understand what you mean

I mean if u write more phrases sometimes it seperate after comma with an space somtimes without space
example:
"select, okay, enter, go"
"select,okay,enter,go"

make this a difference?

under phrase "abc" changeable? german, english?
I don't understand what you mean.

I think u call it payloads?!? "abc" is the icon under the phrase and after the icon are payloads?
Here I write german, too?

best confidence was 72%,intersection 50% it is okay or work on articulation?
I don't know about this. I usually get >90% when I use a good microphone in a relatively quiet environment. Is this using german words and the german recognizer set it windows control panel? You also need to set up your microphone correctly and adjust the volume. Turning down the input volume can sometimes help.

kk microphone adjust down was helpfully but better than 80% never managed


One wish again in ur editor: pls make one button 4 "expand all"


What kind of icon are u need from me? Only icons or a new skin?

[jitterjames]

lol, playing is always faster than real work!

here's your "expand all" version.

I mean if u write more phrases sometimes it seperate after comma with an space somtimes without space
example:
"select, okay, enter, go"
"select,okay,enter,go"

make this a difference?

No, I don't think it matters. Your two examples should work the same as each other.

I think u call it payloads?!? "abc" is the icon under the phrase and after the icon are payloads?
Here I write german, too?

yes you can write german here. Some payloads won't work any more if you change them though. Example, the command goto. Only those payloads will make sense to the WMC plugin. It is possible to create payloads in a separate xml file so that each payload can have its own phrases, but I am still working on this.

solution: if you want to be able to say "Ttime to watch movies" instead of saying "goto movie library"
right now this goto command works with payloads. the command is "goto" and the payload is "MovieLibrary".
you can create a new command called "goto&&MovieLibrary" and then put whatever phrase you want, like "Time to watch movies" or something in German. && needs to match the delimiter setting found in options, under UDP settings. && is the default.

I don't *need* any icons, but if you had some ideas to replace my Vox icons, that might be nice. It's hard because they need to be easy to read. That is why I use the red, green, and yellow background for my listen/ignore icons. If you had ideas for the Phrase, Payload List, Payload Range icons etc. that might be nice. You don't have to though, only if you get a good idea and want to try it.

[jitterjames]

Matt: I am working on a way to scan the media center to get the list of Artists. Can you please unzip the attached to a folder and run it. When you click the button, it should fill the lists with the names of your artists, albums and songs.

Please let me know if it works. It should be relatively quick to fill the list when you click the button.

J

[matt_jackson]

Matt: I am working on a way to scan the media center to get the list of Artists. Can you please unzip the attached to a folder and run it. When you click the button, it should fill the lists with the names of your artists, albums and songs.

Wow This is my most favorite wish.

kk first I set my phone clock on Quebec time; we have 6h delay

One point 2 interferes me is the low hit rate, I think the reason for this problem is my BT HS. I buy later 4 testing an cable HS but this is not the solution cuze I wanna use this voice command 2 control my home technology and a constant cable is absolutely useless there, also a state microphone cuze if I run 2 the mic I can also hit the monitor.
I see some slick computer BT HS, I must test more...

* I hear many hiphop music and its very funny cuze VoxCommand respond with some words of artist. My idea was make a adjustable "gate" 4 the microphone. Idk if it possible, if it useable...one idea...

* I use an multitouchscreen monitor and WMC 7 supported him. There are some nice animations about kinetics and scroll fx. Can u make a control like "scroll right/left/up/down" up to the point where I say "stop" ?

That is why I use the red, green, and yellow background for my listen/ignore icons. If you had ideas for the Phrase, Payload List, Payload Range icons etc. that might be nice.

Yes thats was a impressive idea, simple and easy 2 understand.
kk I think design stuff can we put @ last.

Please let me know if it works. It should be relatively quick to fill the list when you click the button.

I test it later @ home - now @ work

MaTT


post scriptum: sry if I write every wish/idea down, dont understand this as an order! If it easy 2 make and meaningful, fine - is it shit, forget it

[jitterjames]

One point 2 interferes me is the low hit rate, I think the reason for this problem is my BT HS. I buy later 4 testing an cable HS but this is not the solution cuze I wanna use this voice command 2 control my home technology and a constant cable is absolutely useless there, also a state microphone cuze if I run 2 the mic I can also hit the monitor.
I see some slick computer BT HS, I must test more...

that explains it. Bluetooth is no good. There is no such thing as a bluetooth headset that works well with voice recognition. Bluetooth frequency range and compression is not suited for recognition. If you create a small number of commands that sound very different from each other, then you can just manage with bluetooth. In my opinion, it is not worth even trying.
All of the solutions that I have found, are either very expensive, or use a large headset.
2 headsets that work, but are big are
logitech clearchat wirelss: http://www.logitech.com/en-us/webcam_co ... vices/4226
and the plantronics 995: http://www.amazon.com/Plantronics-Audio ... =pd_cp_e_1

I have the plantronics and it works well. I often wear it around my neck with the microphone sticking up but it's not ideal. I was thinking about taking it apart and trying to make it into a small headset...

if you don't mind spending more there are some options here: http://www.emicrophones.com/c-17-wirele ... tooth.aspx

apparently the samson's are good. I tried the cs-55 and it was totally useless.

there are some hybrid bluetooth microphones that are "supposed" to work, but they don't work when in bluetooth mode, they have some other mode that needs to be active, then you can still switch to bluetooth to use it with your phone.

I have used a $0.01 microphone with my ipod and it is much better than any bluetooth headset on the planet!

I don't really understand about the scrolling, but if it is in WMC I am limited to the commands listed here: http://vmccontroller.codeplex.com/wikip ... umentation

maybe we should switch to using email, and dropbox to transfer files? It would be easier for me...

[jitterjames]

I test it later @ home - now @ work

does it work? I tried it on 2 win7 machines and 1 vista and it seems good.

[matt_jackson]

I test it later @ home - now @ work

does it work? I tried it on 2 win7 machines and 1 vista and it seems good.

I┬┤m back now and first I do is test it
I think u waiting 4 = It work very well! Fast & fine

More later, cya!
MaTT


edit 1 : expand/collapse button dont fixed..if u transforme editor window...
edit 2 : after sennheiser Hs buy & use confidence between 72-85%
edit 3 : ur payload example work, yeah ur make me happy, step by step 2 the perfect world
edit 4 : Pm out with mail contact

[andyss]

Hey James, great progress on the app, I'm testing it all the time. We discussed BT headsets a while back and I noticed you recently testing out a workaround on your ipod, but I wanted to mention what I've been using. I use an iphone app called PocketAudio from Senstic. It installs a sound driver into Windows 7 and the iphone app connects to it over wifi essentially making your iphone a wireless mic. It actually works really well and I leave the driver as the default in 7 so that I can use it whenever without worrying if it is the default or not. It works pretty good at a distance as well. Check it out if you get the chance.

As for VC, I was wondering if you could help me out with a few things:

I'm trying to get Google search working, ie: say "Google Search for something" and a webpage opens up with "for something" already searched by Google (obviously). Did you get this working? I've noticed VC also returns a message to EG with numbered selections, how do you use this to make a selection if VC didn't get the correct search term(s)?

How did you program the "Read Weather" functionality as you did in one of your videos?

I also noticed you doing some stuff in the iTunes plugin thread but it seems a bit outdated. With your iTunes version of VC do you think you can make available the commands you used to control iTunes or maybe do something with that plugin? Or even incorporate that into a newer, integrated version of VC?

As always thanks for the great program!

Andy

[jitterjames]

Hey Andyss, glad to hear you are still with us. Thanks for the heads up on pocketaudio. It sounds like exactly what (well one of the many things) I've been looking for? Do you know if it compresses the audio, or would it be equivalent to a wav? Is the reliability good?

Stay tuned cause I've got a few things in the works. I'm planning on building in a few things that will work without eventGhost.
-media center, then maybe xbmc, then maybe itunes. The itunes plugin for EG died and I doubt it will ever be resurrected, at least not by me.
-much more power through custom xml files that contain payloads, that can be reused, do partial matching, and have multiple phrase aliases for each payload. I actually have this working, but no user interface for it yet.
-built in wake on lan
-built in ability to launch applications without eventGhost
-ability to detect and use wav files as input. Ok you've seen this, but I haven't got the interface there yet for other users to select what folder to monitor etc.
-I've moved the command tree editor to its own window and I'm doing the same thing with the options. I'm then breaking the options up into tabs for differnt programs and features like xbmc, wol etc. this should streamline the main interface a lot leaving just the stuff you need to see.
-Eventually I plan to add more custom controls to each command branch and command so you can set priority (i.e. if you commands matches two different commands use the one with higher priority), the ability to select partial matching for individual phrases etc.

I think I've got 1000 other ideas kicking around back there... Needless to say some of this stuff won't be happening for a while, but I'm excited about ALL of it.

Ok, for the google searches:

here's the xml for vox:

Code: Select all

        <command name="Google">
            <phrase>search web for,google search</phrase>
            <payloadDictation />
        </command>

So if you say "search web for george washington" it should send an event to EG called Google with payload "george washington"

then all you need is the EG macro:

Code: Select all

<?xml version="1.0" encoding="UTF-8" ?>
<EventGhost Version="1462">
    <Macro Name="Voice Google Search" Expanded="True">
        <Event Name="Broadcast.Google" />      
        <Action>
            System.Execute(u'C:\\Users\\james\\AppData\\Local\\Google\\Chrome\\Application\\chrome.exe', u'"http://google.com/search?sourceid=navclient&q={eg.event.payload}"', 0, False, 2, u'')
        </Action>
    </Macro>
</EventGhost>

You'll need to adjust the path to point to your browser's .exe
I'm using payload dictation, but if you want more accuracy for bookmarks or something you could use a payload list.

For the options, you just say "option 2" and it will send whatever option 2 was as the command. To display the possible "alternate" options in EG

Code: Select all

<?xml version="1.0" encoding="UTF-8" ?>
<EventGhost Version="1462">
    <Macro Name="Display VoxCommando Message" Expanded="True">
        <Event Name="Message" />       
        <Action>
            EventGhost.PythonScript(u'fileencoding = "iso-8859-1"\neg.globals.message = eg.event.payload.decode(fileencoding)\n#print u"before:",eg.globals.message\n')
        </Action>
        <Action>
            EventGhost.ShowOSD(u'{eg.globals.message}', u'0;-37;0;0;0;700;0;0;0;0;3;2;1;34;Arial', (255, 255, 255), (0, 0, 0), 6, (0, 60), 0, 6.0, True)
        </Action>
    </Macro>
</EventGhost>

part of the default setup for vox contains the options command, but in case you deleted it it looks like this:

Code: Select all

        <command name="Option">
            <phrase>Option,Select Option</phrase>
            <payloadRange>1,6</payloadRange>
        </command>

How did you program the "Read Weather" functionality as you did in one of your videos?

Let's save that one for a rainy day. It's a hidden feature in vox that only works for my location... Another thing to adapt one day for general use. Keeping track of all the user settings is actually one of the biggest pains in my ass. Probably because I've never done anything like this before, and I'm making it all up as I go along. There is a plugin available for eventghost, and I invite you to figure it out and send the results to the TTS plugin. I don't like that particular plugin myself, but hopefully it'll work for you. The other option is to write some python code to scrape xml info from your favourite weather page. If you know how to program, and just a bit of python, it's actually pretty easy, though of course, time consuming.

I also noticed you doing some stuff in the iTunes plugin thread but it seems a bit outdated. With your iTunes version of VC do you think you can make available the commands you used to control iTunes or maybe do something with that plugin? Or even incorporate that into a newer, integrated version of VC?

I kind of covered this already, but ya, maybe one day... iTunes sucks though... I'm more interested in having a button in vox to scan your mediamonkey library for music data.

If you get a chance, please make me a video of your Vox in action!

Feel free to contact me again if any of this is not clear. My eyes are starting to blur a bit...

JitterJ

[andyss]

Thanks for the heads up on pocketaudio. It sounds like exactly what (well one of the many things) I've been looking for? Do you know if it compresses the audio, or would it be equivalent to a wav? Is the reliability good?

I'm not sure if it compresses the audio or not but I usually get reliability in the 90's. The vendor website has no info on format/compression but it seems to stream realtime (with a small network delay ) like a wireless mic.

iTunes sucks though

Yes I know!

Thanks for the hand holding with regards to the other stuff!

Andy

[jitterjames]

hey I just tried the app out and it works great. A little delay that takes getting used to but very good. Certainly better than what I was doing with record pad. I am pleased to see that I can sleep my ipod and it still works, though I'd imagine it'll kill the battery pretty quick nonetheless.

My problem is that I can't get it through my firewall (weird). I've added exceptions, (actually they were added automatically, but I added more to all the files in the installed directory), and no go. If I turn off my firewall it works fine. I've sent sentic an email but I was wondering if you had any experience with this.

So this $2.00 app with the stock ipod headphones, or the 1 cent microphone work a million times better than that stupid cs-55 that I bought for over $200. Of course I returned it but it was a royal pain in the arse with all the paperwork, phone calls shipping etc. What a joke. Anyway yay!

Do you have an actual iPhone, or an iPod? Now I'm thinking I'd like to find an adapter. the ipod uses a weird single headphone jack that houses the wires for the earphones, mic, and play buttons. It would be nice if I could put a cheap mic on a wire and clip it to my collar... I'm going to go look on amazon now...

thanks again for the great find.

[andyss]

My problem is that I can't get it through my firewall (weird). I've added exceptions, (actually they were added automatically, but I added more to all the files in the installed directory), and no go. If I turn off my firewall it works fine. I've sent sentic an email but I was wondering if you had any experience with this.

Are you using it outside your local LAN? I thought it only worked locally, never tried it outside of that. Maybe port forwarding is the issue?

Do you have an actual iPhone, or an iPod?

I have an iPhone 3GS so the mic dongle thing doesn't apply.

Andy

{Edit}
Did you install the Apple Bonjour service? Not sure if it needs that or not... can't remember.

[jitterjames]

ok thanks. no I'm using it on my LAN and I installed bonjour, even though I had it already from something else.


I'll see what they have to say.

[matt_jackson]

If you get a chance, please make me a video of your Vox in action!

Is this 4 andyss only or 4 all?

I was just 2 make a video as soon as ur unbelievable artist/song/album searching machine with VC work.

MaTT