Version 0.86

[jitterjames]

OK.  I have uploaded the patch to 0.860 on the first post of this thread (topic)
http://voxcommando.com/forum/index.php?topic=339.msg1545#msg1545

This is a patch to update your existing VC.

- Make a copy of your existing VC folder
- Unzip the contents of the attached file into the new folder.

 I have changed a lot of internal code.  There could some major bugs so keep your original folder in a safe place!

[painy]

I've got a problem with my mediamonkey edition (witch i patched to 0,860).
I use two instances of voxcommando, one for mediamonkey and one for XBMC and the PC
I get the following error only with the mediamonkey program (see image). The 'mediamonkey play artist ....' command and others don't work. Other commands work without a prob

Also, about learning mode, how does the computer know when it made an error? because if you say something and it interprets the command wrong and executes an event, how does voxcommando know it made an error?

One final remark: I added TTS.speak to commands like set rating and set volume because i was never sure if VC heard me saying numbers correctly, but thats just my personal preference

grtzi

[jitterjames]

yes I'm having a similar problem.  Try moving the command "play hits by" to another group and see if it fixes it.

I am looking for a solution.

Also, about learning mode, how does the computer know when it made an error? because if you say something and it interprets the command wrong and executes an event, how does voxcommando know it made an error?

It doesn't know if it makes a mistake.  That's why it is a good idea to turn learning off, and only turn it on when you have a quiet room and are speaking carefully.

In older versions of VC it was on all the time.

[Kalle]

yes I'm having a similar problem.  Try moving the command "play hits by" to another group and see if it fixes it.

I am looking for a solution.

Also, about learning mode, how does the computer know when it made an error? because if you say something and it interprets the command wrong and executes an event, how does voxcommando know it made an error?

It doesn't know if it makes a mistake.  That's why it is a good idea to turn learning off, and only turn it on when you have a quiet room and are speaking carefully.

In older versions of VC it was on all the time.

James,

simlar problem, but with your tip is fixed. Great work in the Command-Tree-Editor, yes, that is it what i have meaning.

Kalle

[painy]

That fixed it indeed.

Could you elaborate quickly how the learning mode works?

Thx, painy

[jitterjames]

I am hoping that this .exe will fix the problem you had and allow you to put "play hits" back into the same group as "play artist".

Attachments removed... there is a newer version

[Kalle]

I am hoping that this .exe will fix the problem you had and allow you to put "play hits" back into the same group as "play artist".

i tested tomorrow, thanks and good night (in Germany).

Kalle

[Wanilton]

I've got a problem with my xbmc  edition (witch i patched to 0,860).
Problem with this group of commands - XBMC Smartplay lists

[jitterjames]

Wanilton:

I'm sorry to hear it (but not surprised!). Can you please be more specific about the problem?  

Did you try the updated .exe ?

I can't really test smart playlists just from the file "XBMC LISTAS INTELIGENTES.xml".

SmartPlaylists depend on the commands (which you uploaded) but also on your library data (genxml) and also on the files in your smarts folder.

Please describe the problem in more detail, and if you want you can send me the whole VC folder.

[Wanilton]

Hello James,

Download my directory vox, exe is updated with the last version
http://github.com/downloads/Wanilton/VoxCommando/VoxCommando.zip

[jitterjames]

That fixed it indeed.

Could you elaborate quickly how the learning mode works?

Thx, painy

VC uses the Windows built in speech engine.  By default, the windows engine will always try to learn your accent as you use it.  This is good if you are the only person using it, and you are in a quiet environment, but this is not always the case.

In VC if the "learning mode" box is checked, then whenever VC understands a command, it should improve its accurace for your accent, especially for that command or similar commands.  If it doesn't understand you, then nothing happens.  If the box is not check, then it does not learn.

If you have a strong accent, then you should always do the windows training whenever you first create a new speech profile.  This is a more powerful form of learning, because it is only using one sentence at a time, and it also does some important things for new profiles, like determine if you are a man or a woman.

Once you have the profile well trained for you voice, it is best to leave learning turned off, especially if you leave your microphone on all the time.  Certain repeating sounds (dogs barking etc.) could fool the engine and if it thinks that it understood a command it could start to learn the dog bark instead of your accent!

This is not all 100 science or fact, because Microsoft does not make this information easy to find.  Some of it is rumour, some of it is based on what "seems" to happen.  I do have a access to a sapi command that lets me change the setting for learning mode to "off".  I do not know exactly what it does.

Is there something in particular that I can explain more clearly?

[Kalle]

I am hoping that this .exe will fix the problem you had and allow you to put "play hits" back into the same group as "play artist".

it works fine

[Wanilton]

James, solved my problem, I found that uninstalling the XBMC, and re-install, I lost a payload that wore, what was causing the error, but now it is okay again.
Anyway, my commands are as many long, I'll redo it, based on its default configuration file XBMC because there everything is running very fast, including the boot, whereas my current one, even without my mistakes this very slow to load.
 

[painy]

That fixed it indeed.

Could you elaborate quickly how the learning mode works?

Thx, painy

VC uses the Windows built in speech engine.  By default, the windows engine will always try to learn your accent as you use it.  This is good if you are the only person using it, and you are in a quiet environment, but this is not always the case.

In VC if the "learning mode" box is checked, then whenever VC understands a command, it should improve its accurace for your accent, especially for that command or similar commands.  If it doesn't understand you, then nothing happens.  If the box is not check, then it does not learn.

If you have a strong accent, then you should always do the windows training whenever you first create a new speech profile.  This is a more powerful form of learning, because it is only using one sentence at a time, and it also does some important things for new profiles, like determine if you are a man or a woman.

Once you have the profile well trained for you voice, it is best to leave learning turned off, especially if you leave your microphone on all the time.  Certain repeating sounds (dogs barking etc.) could fool the engine and if it thinks that it understood a command it could start to learn the dog bark instead of your accent!

This is not all 100 science or fact, because Microsoft does not make this information easy to find.  Some of it is rumour, some of it is based on what "seems" to happen.  I do have a access to a sapi command that lets me change the setting for learning mode to "off".  I do not know exactly what it does.

Is there something in particular that I can explain more clearly?

No, thats ok, i understand enough now
Cheers!

[Kalle]

That fixed it indeed.

Could you elaborate quickly how the learning mode works?

Thx, painy

VC uses the Windows built in speech engine.  By default, the windows engine will always try to learn your accent as you use it.  This is good if you are the only person using it, and you are in a quiet environment, but this is not always the case.

In VC if the "learning mode" box is checked, then whenever VC understands a command, it should improve its accurace for your accent, especially for that command or similar commands.  If it doesn't understand you, then nothing happens.  If the box is not check, then it does not learn.

If you have a strong accent, then you should always do the windows training whenever you first create a new speech profile.  This is a more powerful form of learning, because it is only using one sentence at a time, and it also does some important things for new profiles, like determine if you are a man or a woman.

Once you have the profile well trained for you voice, it is best to leave learning turned off, especially if you leave your microphone on all the time.  Certain repeating sounds (dogs barking etc.) could fool the engine and if it thinks that it understood a command it could start to learn the dog bark instead of your accent!

This is not all 100 science or fact, because Microsoft does not make this information easy to find.  Some of it is rumour, some of it is based on what "seems" to happen.  I do have a access to a sapi command that lets me change the setting for learning mode to "off".  I do not know exactly what it does.

Is there something in particular that I can explain more clearly?

No, thats ok, i understand enough now
Cheers!

painy, you can find your training files here: C:\Users\"you"\AppData\Local\Microsoft\Speech\Files

Greetings Kalle