Major update ALPHA 2.3.0.*

[jitterjames]

I normally put all new releases on the main download page of the VoxCommando website.

In this case I have been working on so many different changes in the core program that I feel it would be best to release an Alpha version here before putting it somewhere that may confuse new users.

The beta releases are now back on the regular download page:

https://voxcommando.com/home/downloads/

The information below is still important:

Added 2 new plugins:  GTTS and AccuW  (more on these later).  Both these plugins require the user to set up their own credentials to use the required cloud services.  At the moment it's free but there are extra steps.

Changes to the core program:

64bit.  VC was always compiled as a 32bit program for compatibility reasons.  The new version of VoxCommando has 4 executibles to choose from instead of two.

VoxCommando.exe (Will run on 64 bit machines as 64 bit application but will still run on 32 bit machines - does anyone still have one?)
VoxCommando32.exe (Will run as 32bit.  This is similar to previous versions of VC)
VoxCommandoSP.exe  (As above but the SP version of the speech engine - 64bit)
VoxCommandoSP32.exe (As above but the SP version of the speech engine -32bit)

I would recommend using the 64bit versions unless you run into a compatibility issue.  For example, some TTS voices will only run on 32bit and some others will only run on 64bit.  Most TTS voices *should* run on either.

Other core changes:
Action parameters are now multi-line.  I had to create a home-grown autocomplete for variables to work on multi-line text boxes but the good news is that we can use autocomplete anywhere in the parameter text box now, not just in a blank text box.

Compiled now using .NET 4.5.2 instead of 4.5.1.  This should not be noticeable in any way.  Eventually we will move to 4.6.1 but I did not want to rock the boat too much more than I already have.

Regex Tool.  Changed how line endings are evaluated in the tool window (\r\n instead of \n) so that it will match the way actions work.

RoboB plugin(RoboBrowser) : Minor but important bug fixes.

Please use this version and report any issues.  If you are updating an existing configuration please be sure to back up your whole folder first!

[jitterjames]

Using the GTTS plugin

This plugin was suggested by Kalle.  It uses the Google cloud text-to-speech service, which has very high quality voices in many languages.

https://cloud.google.com/text-to-speech/

You must sign up as a developer and go through the steps necessary to download your own authorisation json file.

You'll need to do this *before* trying to enable the plugin in VC, or you'll get an error message.

Instructions here: https://cloud.google.com/text-to-speech/docs/quickstart-client-libraries

You only need to follow steps 1 to 4 to obtain the .json file.
* It is not necessary to modify your environment variables.
* It is not necessary to install the SDK.

Once you do this you must rename the json file provided by Google to GTTSAuth.json and save it in your plugin folder: ...VCroot\plugins\GTTS

I believe you are limited to 1 million characters per month which should be enough for anyone.  All TTS audio will be cached to the plugins\GTTS\cache folder as mp3 files, so if you use the same TTS text with the same voice, pitch, and speed then it will re-use the mp3 file instead of generating a new request.  This should save your quota, reduce your bandwidth, and be faster but could eventually result in a lot of storage space occupied by the cache.

[jitterjames]

Using the AccuW plugin

In order to use the AccuWeather plugin (AccuW) you will need to obtain your own API key.

Follow instructions here to get your free key: https://developer.accuweather.com/

Enable the AccuW plugin and enter the key in the plugin settings, and save settings.

You will be limited to 50 calls per day.  The plugin will cache information and only update when it is stale but there are different types of calls (Current, forecast, hourly) so be careful not to exceed your daily limit when experimenting.  If you exceed your limit I think you will simply be denied access for that day.  It won't cost you anything as far as I know.

[jitterjames]

If someone would like to create a video walk-through showing how to sign up for Google services and obtain the .json file I would appreciate it.  Otherwise the instructions provided by Google should be fairly straightforward to follow.

[Kalle]

absolute great work 

[PegLegTV]

this looks awesome, I can't wait until I have some free time to really test it out, but right now when I try to enable the AccuWeather plugin I get a system error.

I enable the plugin and select "Save and Close" then click ok on the popup, then I see this window immediately after



I click continue and then select a full restart and I see an error in the history window when VC restarts

Error:

Code: [Select]

Error loading plugin: C:\Users\pegle\Desktop\New\Plugins\AccuW\AccuW.dll
I've tried this on my full VC setup and on a fresh install to see if anything would change but I get the same thing with both setups

looking at the log file it makes a mention of:

Code: [Select]

Could not find file 'C:\Users\pegle\Desktop\New\Plugins\AccuW\varinfo.json'.
Log file is from the fresh install, thanks for the great work

[PegLegTV]

not trying to nit pick the alpha release but I noticed that the action boxes are cut-off when they are not in focus


the top action box is only full because I clicked in the text field, but if I focus another action the top one becomes cut-off and the action I click becomes full sized

[jitterjames]

this looks awesome, I can't wait until I have some free time to really test it out, but right now when I try to enable the AccuWeather plugin I get a system error.

Sorry about that. Looks like we are missing this attached file.  Please put the attached file in your AccuW folder for now and I think it will load OK.

[PegLegTV]

That solved the AccuW plugin issue Thanks

[PegLegTV]

hey I'm loving the GTTS and the AccuWeather plugins, great work, like always ,

here's a bug report / request for the two new plugins

would it be possible to add in something similar to TTS.SpeakSync or Sound.PlayWav's wait option, apparently I have several commands that use TTS.SpeakSync, I was going to switch all my TTS.Speak actions over to GTTS.Speak to really put it to the test but apparently I use TTS.SpeakSync for a lot more commands then I thought, and those commands rely on the pause while its talking.

Looking at the AccuWeather plugin the AccuW.GetForcast says "Format 0-3 What format string to use" when hovering over the format box, but when I try to use "3" the output is "expecting integer from 0 to 3" looking at the AccuWeather plugin under the forecast tab it is showing options "1,2,3", so the integer options aren't lining up properly.

also when first setting up the AccuWeather plugin Current tab when I clicked Test current weather format string the variable [C.Wind.Speed.U] doesn't get filled, looking at the list of variables it should be [C.Wind.Speed].

I already hit my limit for my AccuWeather plugin API , totally worth it when testing out a new plugin.

keeps up the great work!
 

[nime5ter]

Thanks for all your efforts with alpha-testing, PegLegTV. Historically this phase was my job, but I've been too busy at work. Glad James has other good folks who're willing and able to contribute!

And yes, wrt speaksync, James mentioned to me last week that it was going to be a challenge to find a good solution for that; rest assured it's on the list.

[jitterjames]

The problem with GTTS speaksync is that we are downloading .mp3 files and then playing them with a windows mediaplayer object and it does not support waiting for the media to finish.  I might be able to find a solution but for the moment you should assume that it won't be available in the near future.  I am about to go on vacation so no new development is going to happen until September.

As far as the AccuW plugin goes.  Please ignore any variables that might have been in the plugin setting fields and use the variable picker.  The variables that are there were from early phases of testing.

Thanks.

[jitterjames]

I am about to go on vacation so no new development is going to happen until September.

OK, maybe I lied a bit.  Today I did some programming when I should have been packing...

Please try the new download link at the start of this thread.  2.302.

[PegLegTV]

hey great work on the new Gtts actions!  they work great from the testing I've done so far, I think I'm going to change the voice for my theater experience commands just to make it feel a little bit more like a movie theater and being able to change the voice and pitch will make it a lot easier. keep up the great work and enjoy your vacation.

[Kalle]

Now I have tried both new versions of the plugins and they works great.

The GTTS.SpeakSync action is fantastic!

Again, absolute great work 

See you in a few days 

[Kalle]

I noticed an "bug" in the AccuWeather plugin - the location is only temporary saved.
When VoxCommando is restarted, the field with the location in the plugin is empty and causes an authorization error 401.
I have solved it temporarily by first executing the"AccuW.Change" action with the location code followed by the weather query.

[nime5ter]

oops.  I'm not sure what happened there.  I will take a look at this when I get home at the end of August.

[keithj69]

Just found the alpha release.  I'll start playing around with it and report any issues if I find any.   Thanks for the continued development.

[jitterjames]

oops.  I'm not sure what happened there.  I will take a look at this when I get home at the end of August.

OK.  I think I've fixed this now.  The new version is linked to at the top of this thread, but the version of VC is still 2.3.0.2 because only the plugin was updated.  The new version of the plugin will be 2.0.0.7

[PegLegTV]

Hey I've been using the new version for a little over a week now as my main VC And it's working great, I've set Gtts as my tts for everything, but I haven't had time test out AccuWeather more.

I've noticed that  | (pipe) doesn't work with Gtts, so you can't use random phrases like the tts plugin does,

Would it be possible to add in the ability to amplify the MP3 that is played back when using Gtts? with tts I was using a command that made the tts volume seem normal at all levels ex. if watching a video that had a low volume level it would turn down the tts volume so it didn't sound like it was screaming at you and vice versa.

also would It be possible to store the mp3 file name created by the gtts plugin in some type of variable, this could be used to send the mp3 to another device for playback

keep up the great work! 

[jitterjames]

Hey Pegleg.

Nice to know someone is actually using the new version.

I noticed this too and had started to work on that and some other additions.  So today I finished up what I was doing.   Please try out 2.3.0.3 at the top of this thread.

The volume was at 100% so it cannot be "amplified" but now you can set it to anything from 1% to 100%.

You can also set the output device etc.

This version is also much easier on the RAM as the last one was not always disposing of the sound player object properly after playing the TTS audio.

I've added info on the speak actions which returns the relative path of the audio file.  You can access this using the {LastInfo} variable.  Note that it's relative to the root VC folder so if you are passing it to another program you'll need to add to the beginning to create the full path.

[PegLegTV]

Awesome work, apparently great minds think alike  , I'll have to find some time in the near future to take a look, I'm working on learning my new CNC   

[Kalle]

Hehe, I used this version too 


That is awesome work James

[Aniv_D]

Thanks for the update!

more question. is it possible to implement 2 new actions.

Marker and Jump to marker.

which made it possible to move to a specified location within the command.
only I do not know whether it will be actual for other users and whether it is real to implement such.

[jitterjames]

I have considered this before and I think it is too difficult to implement properly.

I think you can achieve the same effect using events to trigger other commands, or logic blocks and loops.

Or if you need to to something more advanced then I think it makes sense to use python.

[jitterjames]

Thread split and continued here:

Topic: Hungarian recognition and TTS using Google
https://voxcommando.com/forum/index.php?topic=2941.msg24831#msg24831

[Telorast]

I noticed the AudioEndpoint plugin is 32b only. Will it ever be 64b?

[jitterjames]

Sorry for the delay in answering this but I had to do a bit of research first, since I have not touched that plugin since 2016!

Anything that relies on a 32bit dll will only work in a 32bit application.  The AudioEndpoint plugin uses coreaudio.dll which appears to be 32bit only.

The DirextX plugin is another example of a plugin that relies on a 32bit dll.

I will look to see if there is another option instead of coreaudio.dll that works in both 32bit and 64bit programs.  If there is I will need to rewrite the plugin.  Not exactly from scratch but all the important code will need to be replaced to use a different api.

But I should point out that there is really no advantage to using the 64bit version of VoxCommando instead of the 32 bit version of VoxCommando, at least in terms of performance.  I  have included both options of VC only for the sake of compatibility.  There may be something that only works in 64bit and in that case you might need to use VC 64bit but if everything works for you in the 32bit version, I recommend that you just keep using that instead of the 64bit.  You won't be missing anything.

So far the only thing I'm aware of that requires a 64bit version of VC is certain TTS voices seem to work only in 64 bit applications.  There might be other things like 3rd party plugins that only work in 64bit but if there are I am not aware of them.

[Aniv_D]

Hi jitterjames !!!
I don't know, but it may have been fixed, but still...
I noticed that in 2.3.0.8 in the Russian version there are some errors in the algorithms.
1. Not locked these fields in the algorithm. I think this may confuse users and they will enter data in the wrong fields
2. Some operators contain an extra # character, and the operator "(A)Contains(B)" is also translated and also has an extra #character)

[jitterjames]

Wow.  I did not know about this.  Big problem.

Thanks for letting me know! 

[Aniv_D]

Hello, James.
I'm back to you with new bugs found in 2.3.0.8 (probably in earlier versions, I didn't check)
The problem is this.
1. if you run a command to play audio in the PY plugin and the file name starts with the following letters (a, b,f,n,r,t,v,z - does not work) and numbers (1-7, 0 does not work), the code does not work correctly, as shown in screenshot 1.
2. if the file name starts with a capital letter A or B or C, it works correctly, but with other D-Z capital letters it returns an error message, as shown in screenshot 2.

P.S. if these letters and numbers are used in other places (for example, in the middle or end of the name ), then everything works fine.

[jitterjames]

I don't think this is atually a bug.

I believe this is just the way that Python deals with strings.  Backslash '\' is a special symbol that is used to encode special things.  For example \t is interpreted as a TAB.

Try using double backslashes whenever you want to represent a single backslash.

Example:

Code: [Select]

"D:\\Temp\\Test1.wav"

[Aniv_D]

damn, right. the way it is. I forgot about '\\'.
Sorry to bother you and thank you for your response.

[jitterjames]

No problem.  I prefer to be told about potential bugs, even if I just end up helping someone, which I am also happy to do.

I can see why you made this mistake.  It is very confusing that in some cases the single backslash works normally.  To me this is a sort of bug, or at least a design flaw in Python itself.

[jitterjames]

Hi jitterjames !!!
I don't know, but it may have been fixed, but still...
I noticed that in 2.3.0.8 in the Russian version there are some errors in the algorithms.
1. Not locked these fields in the algorithm. I think this may confuse users and they will enter data in the wrong fields
2. Some operators contain an extra # character, and the operator "(A)Contains(B)" is also translated and also has an extra #character)

1) is a bug caused because the combobox values for the "IF" operator should not be translated at all.

2) The number signs are used to show things that were computer translated (by a web service like Bing or Google translate).  Uusally, if you see anything that starts with a '#' it means that a human needs to check that the translation is correct.  Once someone tells me that it is correct I can remove the '#'.  If anyone wants to help me with this I can send them a list of these computer translations to be approved or corrected if wrong.

[Aniv_D]

1) is a bug caused because the combobox values for the "IF" operator should not be translated at all.

2) The number signs are used to show things that were computer translated (by a web service like Bing or Google translate).  Uusally, if you see anything that starts with a '#' it means that a human needs to check that the translation is correct.  Once someone tells me that it is correct I can remove the '#'.  If anyone wants to help me with this I can send them a list of these computer translations to be approved or corrected if wrong.

I could try to do that. (true, some things I don't quite understand, such as in the screenshot and therefore don't know how to translate them)
P.S. could you show an example of using what is in the screenshot

[jitterjames]

2 phrase payloads are usually used for advanced music payloads.  These are payloads that contain two phrases separated by && which map to a single value.

For example you want to be able to say:

Play [All Along the Watchtower] by [Jimi Hendrix].
or
Play [All Along the Watchtower] by [Bob Dylan].

You might do this if you have multiple versions of the same song by different artists.  I never use this personally but it was requested by users a long time ago.  It can be very hard on memory if you have a large library of songs.  Still it could be useful for some things.

The setting you are asking about allows you to choose the word that connects the two phrases together.  In other words, what word to use for "by".  This would allow us to use different languages.  In French you might want to use "par" as the connector, as in:

Jouer la chanson [Ain't That a Shame] par [John Lennon]

I talked a bit about this feature when it was first released: https://voxcommando.com/forum/index.php?topic=488.msg3585#msg3585

[jitterjames]

GenjXML in the MediaMonkey plugin automatically generates special payload xml files such as "MMSongByArtistShort.xml" whenever the same song is performed by multiple artists.  Here is an example of the xml for the payloads. You can see in the phrase that the song title and the artist are separated by &&.  VC recognizes this as a multi-phrase payload when reading the payload XML files.

Code: [Select]

    <payload>
        <value>6463</value>
        <phrase>96 Tears&amp;&amp;The Stranglers</phrase>
        <subsetmatching>true</subsetmatching>
    </payload>
    <payload>
        <value>6922</value>
        <phrase>96 Tears&amp;&amp;Question Mark &amp; the Mysterians</phrase>
        <subsetmatching>true</subsetmatching>
    </payload>

[Aniv_D]

Thank you, it's clearer now.

1) is a bug caused because the combobox values for the "IF" operator should not be translated at all.

2) The number signs are used to show things that were computer translated (by a web service like Bing or Google translate).  Uusally, if you see anything that starts with a '#' it means that a human needs to check that the translation is correct.  Once someone tells me that it is correct I can remove the '#'.  If anyone wants to help me with this I can send them a list of these computer translations to be approved or corrected if wrong.

If you still need a person to check, then I'm ready. And then I really want to fix the Russian version.

[jitterjames]

Great.  Are you able to edit an Excel sheet?

[Aniv_D]

if you are asking if I have Microsoft Excel or its equivalent, then yes there is.

[jitterjames]

I noticed the AudioEndpoint plugin is 32b only. Will it ever be 64b?

We just released a new version of VoxCommando.

The audioendpoint plugin has been removed, but there are now core actions under Sound.* which allow you to set default inputs and outputs, and they should work in both 32 and 64 bit.