Hungarian recognition and TTS using Google

[Kalle]

I‘m absolute with James - you will have more problems than fun.
The wearable solution (Revolabs or similar) is the best.


Kalle

[krysn95]

Yes, probably you are right, so I can accept this.

However the basic configuration (VC with Chrome Google Speech TCP plugin) works very well, so as I see, I just need to find the right microphone solution.

(for example something like this: https://www.amazon.com/CM-1000USB-Conference-Microphone-ECS-3-5CM-Microphones/dp/B01MXY82LD)

According to your experience and opinion, how do you see..this configuration could be enough reliable and sensitiv?

thanks guys!

[jitterjames]

I have no idea if these microphones are any good but I think it's unlikely that you will be able to chain them between multiple rooms.  You would need to run a very long 3.5mm cable to daisy chain them and the quality loss on a cable like that will probably be very high.  Plus you will need to make holes in the walls etc.

[krysn95]

Guys, does anybody tried to run the VC on Raspberry?
2gb ram is enough to run continuesly the VC and the Chrome page under windos 10?
(according to my laptop this setup using 1,5gb with win 7)

(Im thinking to build a small raspberry assistant)

thanks!

[Kalle]

I have read about that it is possible to install Windows 10 on a Pi3+ or better Pi4. It will run on a Pi, but it will take 10 seconds and more to open only a browser window or context menu.
So I guess VC and speech recognition will not really works.


And here the answer from the pi forum


https://www.raspberrypi.org/forums/viewtopic.php?t=151859

[krysn95]

thanks again..you know everything...than I will find another way...

[Kalle]

By the way, there are some tiny Windows machines available which you can use for such a project.
Example: Zotac Pico

https://www.zotac.com/de/product/mini_pcs/zbox-pro-pi335-pico

https://www.zotac.com/de/product/mini_pcs/pi330-1

[krysn95]

Yes the zotac is great just a bit too expensive..im looking for something under ~100$.
The VC working only full win 7/8/10, isnt it?

Do you have any other cheap mini pc suggestion what have lo consumtion and can run VC?

Thanks Kalle

[Kalle]

There are a lot of mini PC on Amazon or ebay availabe.

I have used VC on a Mini PC with Intel Atom Z8350 for a long time. The power consumption of these devices is often below 10 watts, which is a very good value. The only annoying thing about Win 10 are the constant updates.
But I recommend at least 4GB RAM and 64GB eMMC.
Here a example on Amazon: https://www.amazon.de/ACEPC-T11-Pro-Prozessor-Anschlüssen/dp/B07Q9JRM5F/ref=psdc_427954031_t2_B07H884WQT


Changed the link with a better one 

[krysn95]

Hi guys!

I can not say how excited I am.
On the weekend I tried to build up a simple demo setup and actually it seems work very well.

The setup:
Now im using the Google speech recognition TCP plugin together with the GTTS plugin (both on Hungarian language)

For the hands free always listening part, I made a small pc what is running win 10 and keep open the Chrome page with the Google speech, on the right IP port.


So, to step forward, please let me ask some help again:

1)
After 10-15 command, the VC just stop working and freez. After arestart works again and I know that the trial version has 40 command limit, but there is no any error message, just stop working.
The pro version will solve this?

2)
Now Im using one central VC what have an own mic and another mini pc what is just "listenning" with another own mic, throug the IP port. This configuration can work together with 4-5 "listening" pc also?

3)
How can I identify wich lsitenning mini pc sent the command to the VC?
Should I use another VC on the "listening" pc and sending the command somehowe to the center VC?
(Or there is another more simpple way also?)

Guys..thanks a lot!
Im really gratefull!
 

[jitterjames]

1)
After 10-15 command, the VC just stop working and freez. After arestart works again and I know that the trial version has 40 command limit, but there is no any error message, just stop working.
The pro version will solve this?

If you have used all your commands, you should normally get a message telling you so.  It should not "stop working".  If VoxCommando freezing this is not expected behaviour.  I would need to do some tests but probably buying a license is not going to solve this one.  Unfortunately "stop working" could mean a lot of different things.  Is VC frozen?  Is VC responding to normal voice commands?  In other words, is it just the Google speech that is not working or are other parts of the program not working.

[jitterjames]

2)
Now Im using one central VC what have an own mic and another mini pc what is just "listenning" with another own mic, throug the IP port. This configuration can work together with 4-5 "listening" pc also?

You already asked us this one and the answer was "yes".

[jitterjames]

3)
How can I identify wich lsitenning mini pc sent the command to the VC?
Should I use another VC on the "listening" pc and sending the command somehowe to the center VC?
(Or there is another more simpple way also?)

You already asked this one too.  As I said before, there is currently no way to know which PC sent recognized speech from the Google Web page to VC.

Yes, if you run a complete installation of VoxCommando on another machine then it's possible to relay the commands to another PC running VC along with information about which PC is sending the command.  Then you have to figure out how to use that information.

[jitterjames]

Also, I'm glad to hear you are having fun!.

[krysn95]

If you have used all your commands, you should normally get a message telling you so.  It should not "stop working".  If VoxCommando freezing this is not expected behaviour.  I would need to do some tests but probably buying a license is not going to solve this one.  Unfortunately "stop working" could mean a lot of different things.  Is VC frozen?  Is VC responding to normal voice commands?  In other words, is it just the Google speech that is not working or are other parts of the program not working.

The Google Speech working (the Chrome page still recognizing the speech), just the VC frozen after ~20 command and not react to anything.
The situation is the same on the Win7 and Win 10 also. I can not even close or restart the VC just with task manager or ith complet windows restart. Am I doing something wrong? Maybe the ram is not enough (4gb on win7; 2gb on win 10)?