Did anybody test Elechouse's voice recognition modules ? They are sold as extensions for Arduino with software, but since "they reveal" the serial communications protocol, it should be possible to use them with a prog written in GC Basic. Does anybody test them ? There are 2 modules, SimpleVR (they say "user independent", but only in English ... and Chinese : I found the utility, but not at Elechouse) and the Voice Recognition Module V3, which is speaker dependent and needs preliminary training (7 words are enough for me). Or does anybody know another "good" module, i.e. which could be used by kids too (for my train layout), without frustrating them after maximum 5 minutes ? Some modules on ALIxy ... seem very interesting, but it is impossible to find a good technical description or programming utility (especially for the "user independent" modules), even if You contact the manufacturer : no response ! And some other "well known" modules are v-e-r-y s-l-o-w (more than 0.75 sec. ) : look at the demos on YouTube !
Thanks for any answer ...
Last edit: Bertrand BAROTH 2020-08-17
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Now I tested VeeaR's Easy VR3+ ; forget the Speaker independent mode, it is all the way too slow and the software to "customize" the vocabulary is VERY expensive. But the speaker dependent mode seems interesting, especially if it used in "fast" mode, at least fast enough to recognize 6 keywords used to trigger the 6 commands associated with the 6 Digitasts on the control panel of my railroad layout. The only issue with this device is that one must wait for at least 40 (or better 50 for safety) ms between acknowledgement and sending the next command ; and, when the device notifies that it HAS recognized something (some time after saying the word), You must wait again 50 ms before sending a space to get the index of the command recognized. So I am afraid that it would not be adequate for a RC vehicle (my other hobby) but for the railroad layout (see attachment, picture about 3 months ago) it is all the way fast enough ... And I will try using a text LCD display, to display the various informations, especially for training the voice "on site" : I think such a feature will be of interest for the grandchildren of my friends ... :)
About Cortana : I don't wanna put my computer beside the layout ; and my OS is still Windows 7 pro x64, not 10 ; but I must say that vocals.exe (which creates keyboard shortcuts, using the built-in speech recognition of Windows) works fine, and could be used together with an interface program to send the commands via serial. Presently I use it for my RC-tanks in 1/24, I built an USB-interface for the transmitters using CodeMercs's IOWarrior and wrote an interface prog in Delphi 7.
Concerning Alexa (and other online speech recognition systems, like Google or Siri) : I wonder if they are not directly connected with Echelon (or Prism nowadays), and if it is not possible to hack them and listen to everything in the house (a similar hack already exists for the webcams) ... :(
Last edit: Bertrand BAROTH 2020-09-15
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am not familiar with speech recognition except my old android phone that can search for spoken words.
I searched rpi speech recognition and got lots of results. raspberry pi might suit you.
Some speech recognition is internet based but some is stand alone but needs memory.
Google's speech to text api looks interesting but not free.
good luck.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
My final notice : Veear's Eays VR3 is in fact the best (and almost the only) voice recognition system which works "nearly" in offline SD mode, but sometimes it confuses some words : for example, in french, the word "itinéraire" was recognized once as "départ" (?), it's rare but it happens ! So the system is not reliable enough for demonstrations in front of youngsters, used to their Cortana or Siri ... They would only say : "Your system is ... crap ! " Finally I will keep my system with the digitasts only ... MikroElektronika's Speak Up is good, but it doesn't allow training "in situ", it needs to be connected to a PC.
:(
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
don't beam me up now, I'm having a
I looked up speech recognition for rasperry pi,as it would need a powerful processor
and it looks complicated with the various programs.Open source was first look.
There's loads. I got a few rpis but don't do python so later.
Speech recognitiopn is not "on a plate" ie not trivial.
could you train a dog ,under the table ,to push buttons faster than using a micro?
cost you a few biscuits
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Finally I found an "old" Easy VR3 (not Plus) at a bargain price at RobotShop together with the license to create custom vocabularies (rather expensive if bought alone), it seems that it was the last in stock ; the only differences between the "new" and "old" version seem to be a fast speaker dependent mode and more memory in the Plus version, no speed difference in speaker independent mode. So I bought it, created appropriate vocabularies for my layout (in English, French and German, according to the awaited visitors ! ), each containing 7 expressions, and tested it on my STK200 ; it works fine, fast enough for a railroad layout (not for a RC vehicle real time control, in this case I think that the Plus version in SD fast mode should be used), and so I will "embed" it into my console in a few months (I am presently working on the landscape).
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Did anybody test Elechouse's voice recognition modules ? They are sold as extensions for Arduino with software, but since "they reveal" the serial communications protocol, it should be possible to use them with a prog written in GC Basic. Does anybody test them ? There are 2 modules, SimpleVR (they say "user independent", but only in English ... and Chinese : I found the utility, but not at Elechouse) and the Voice Recognition Module V3, which is speaker dependent and needs preliminary training (7 words are enough for me). Or does anybody know another "good" module, i.e. which could be used by kids too (for my train layout), without frustrating them after maximum 5 minutes ? Some modules on ALIxy ... seem very interesting, but it is impossible to find a good technical description or programming utility (especially for the "user independent" modules), even if You contact the manufacturer : no response ! And some other "well known" modules are v-e-r-y s-l-o-w (more than 0.75 sec. ) : look at the demos on YouTube !
Thanks for any answer ...
Last edit: Bertrand BAROTH 2020-08-17
Now I tested VeeaR's Easy VR3+ ; forget the Speaker independent mode, it is all the way too slow and the software to "customize" the vocabulary is VERY expensive. But the speaker dependent mode seems interesting, especially if it used in "fast" mode, at least fast enough to recognize 6 keywords used to trigger the 6 commands associated with the 6 Digitasts on the control panel of my railroad layout. The only issue with this device is that one must wait for at least 40 (or better 50 for safety) ms between acknowledgement and sending the next command ; and, when the device notifies that it HAS recognized something (some time after saying the word), You must wait again 50 ms before sending a space to get the index of the command recognized. So I am afraid that it would not be adequate for a RC vehicle (my other hobby) but for the railroad layout (see attachment, picture about 3 months ago) it is all the way fast enough ... And I will try using a text LCD display, to display the various informations, especially for training the voice "on site" : I think such a feature will be of interest for the grandchildren of my friends ... :)
Last edit: Bertrand BAROTH 2020-09-14
Ever thought of alexa or win 10 cortana?
About Cortana : I don't wanna put my computer beside the layout ; and my OS is still Windows 7 pro x64, not 10 ; but I must say that vocals.exe (which creates keyboard shortcuts, using the built-in speech recognition of Windows) works fine, and could be used together with an interface program to send the commands via serial. Presently I use it for my RC-tanks in 1/24, I built an USB-interface for the transmitters using CodeMercs's IOWarrior and wrote an interface prog in Delphi 7.
Concerning Alexa (and other online speech recognition systems, like Google or Siri) : I wonder if they are not directly connected with Echelon (or Prism nowadays), and if it is not possible to hack them and listen to everything in the house (a similar hack already exists for the webcams) ... :(
Last edit: Bertrand BAROTH 2020-09-15
I am not familiar with speech recognition except my old android phone that can search for spoken words.
I searched rpi speech recognition and got lots of results. raspberry pi might suit you.
Some speech recognition is internet based but some is stand alone but needs memory.
Google's speech to text api looks interesting but not free.
good luck.
My final notice : Veear's Eays VR3 is in fact the best (and almost the only) voice recognition system which works "nearly" in offline SD mode, but sometimes it confuses some words : for example, in french, the word "itinéraire" was recognized once as "départ" (?), it's rare but it happens ! So the system is not reliable enough for demonstrations in front of youngsters, used to their Cortana or Siri ... They would only say : "Your system is ... crap ! " Finally I will keep my system with the digitasts only ... MikroElektronika's Speak Up is good, but it doesn't allow training "in situ", it needs to be connected to a PC.
:(
Computer, "Tea-Earl Grey hot"
yeah
And Your replicator will deliver "cat food number 26" instead ...
You see, we have the same "sources" !
:)
don't beam me up now, I'm having a
I looked up speech recognition for rasperry pi,as it would need a powerful processor
and it looks complicated with the various programs.Open source was first look.
There's loads. I got a few rpis but don't do python so later.
Speech recognitiopn is not "on a plate" ie not trivial.
could you train a dog ,under the table ,to push buttons faster than using a micro?
cost you a few biscuits
Finally I found an "old" Easy VR3 (not Plus) at a bargain price at RobotShop together with the license to create custom vocabularies (rather expensive if bought alone), it seems that it was the last in stock ; the only differences between the "new" and "old" version seem to be a fast speaker dependent mode and more memory in the Plus version, no speed difference in speaker independent mode. So I bought it, created appropriate vocabularies for my layout (in English, French and German, according to the awaited visitors ! ), each containing 7 expressions, and tested it on my STK200 ; it works fine, fast enough for a railroad layout (not for a RC vehicle real time control, in this case I think that the Plus version in SD fast mode should be used), and so I will "embed" it into my console in a few months (I am presently working on the landscape).