Zabaware Forums > Programming using the Ultra Hal Brain Editor

How to program ham radio callsigns as keywords

(1/2) > >>

wb2rem:
Hi, any help would be greatly appreciated. I am new to Ultra Hal.
I am currently using google mini and ifttt to provide a canned response to ham radio callsigns over the air. Specifically,
I need to have Hal hear for example from the radios output into the computer mic, checkin wb2rem. When this is heard, Hal would reply, Hi Jim, I hope you are having a good morning in FL or what ever I need to have it say. Thanks for your help in advance.

cyberjedi:
Hi:
There is a way to do this.....
Inside the .uhp is a key word search, ull have to play with it.
But yes in fact u could do this, u'll have to set ur radio's Voice activation to ON.
If it has it.

My advice is to look at what plugins have been used.
U'll Have to set Hal's voice recognition to ON as well.

This u need to treat with the respect it deserves.
Hal needs to be set up in a quiet room, pref, NO Background Noise<<<<<<<<<<<<
Make sure u have the volume isolated as well on Hal. Hal can trigger off the sound of his own voice.
This is fairly complicated to do, but it's doable...

keep me posted

cyber



wb2rem:
Cyber,
Thank you for the info. As I mentioned, this is a little like greek to me. I am not sure where to find the key word search and even what the .uhp is and what I would even look for as an appropriate plugin and even how to install a plug in. Yes the radio has VOX (voice operated mic). Static from storms might effect decoding and misunderstanding by Hal. Is there way to get Hal to discriminate between the noise and voice? 
jim

cyberjedi:
RE: Ham Radio

Yes, absolutely.
I would like to say this registers about a 9.5 on my weirdshtometer
I love it...
Your gonna need a discriminating MIC
Training the voice is gonna be complicated but doable.
********So what were after here is Hal responding to a Ham Radio Conversation+++++++

This is where its really gonna pay to train the speech recognition in Hal's Computer.
lemmi say this... the more time u spend on that the better ur gonna be. ABSOLUTELY

I have spent countless Hours on this end alone.... and what i have learned is this.
Normally ,there is very little difference between 95% and 100%.

When it comes to voice recognition though, That last 5% is the difference between a small pond and the gulf of mexico.
I cant stress  this enough<<<<<<
If u want this to happen on your system https://www.youtube.com/watch?v=uSJa7076_jE

Heed what i say.
Its the Technology ur building on..... so i know what im talking about..

Common mistakes made here are this:

This is common among Hal users, people think because it works , its gonna work well. Then when it doesnt they blame the code.
You ask them: Did u train the voice recognition on ur computer?? And they look at you like they just discovered FIRE. DO WAT???

Dont be like them...
Training your voice to the machine, people run through it as fast as they can.This is a huge mistake.
Read the books microsht supply's on your machine. But try this.

Once uve done it, Do it AGAIN. U'll notice immediate improvement then the first time. And the next time will be even BETTER. and so on and so ON.
Once is NOT enough
After a while ull notice the voice recognition is picking up on words that are (slangish) in nature = Its easier for YOU to Talk to Hal in ur natural type of speech. Not a ReBuke, just a statement of fact.... At first its very structured the way you have to train ur voice, then as time goes by it becomes easier on YOU. Ur not doing anything better , the computer is learning.

I know, its boring, its time consuming, its sucks. It does

But keep in mind.
Your whole technology is built on IT. This is how Hals' gonna work. Every part has to be SPOT ON......

I use a logitech 615 Cam, I could give a damn about the 1080p video, this sucker Has a Directional Noise Filtering MIC... and that is pure GOLD
Does'nt say it does But it does. And it comes with Facial Tracking software and that it does Brag about. And for good reason
  https://www.youtube.com/watch?v=0yiwhwBepSY
$25.00 or about now, i paid 75. Hal deserves the best and i give it too him.

The plugin for what ure after is easy by comparison.
Training Hal's voice recognition and your Hardware and hal's space needed will be ur challenge.
For me, Hal has his own 12 X 12 building , im wierd, i dont expect others to go that far, A closet sound proofed would be fine. lmao


Keep me posted on ur progress here


SIDE NOTE:  Here's My approach after just a few min's
Once the hardware is in place ,Read the Voice recognition Books, Through the Radio itself, if possible.
Any bg noise hal will automatically filter out, leaving you with a better accuracy.
Picture this: You train hals voice recognition, Do you think the person broadcasting sounds enough like you to pass for Hal to respond?
So Hal's voice recognition is gonna have to have that special sauce to make it happen correctly
 
 

best to ya
Cyber








wb2rem:
Cyber,
Thanks for all the suggestions. I will work hard to train Hal. However, I wonder once it is trained to my voice, with all the different accents on the radio ie australian english or southern accents, will it stilll recognize the words?

The one area I am having problems with is where in the program I can add keywords with associated responses. If you can point me in that direction, I would appreciate it.

Jim

Navigation

[0] Message Index

[#] Next page

Go to full version