REQUEST for a sound based response brain plug in again .

[lightspeed]

We have had a lot of new people joining lately , this question ( i have asked before ) and is still open to programmers,
     What i would like to see created ( if possible)  is a brain plug in of sound acknowledgement .
what it would do is within the script have random script for trigger sentences (i have that part and random response sentences , i have that part . what i need is programing script  within the random trigger sentence , in another words
  in chosen area instead of a trigger word it would have a sound clip ( any sound clip can be used , my own recorded door bell. when this sound is heard as a trigger sound the programming would = to random response 1 to 10 and hal would say : (for example) somebody is ringing the doorbell or someone is at the door, or is that the doorbell ringing?  etc.
I am still hoping that someone can create a plug in like this . their could be many applications and responses written for it .
I will also throw this idea to robert by email , but he is awful busy lately so thought i would throw it on here again .
  I have the "base" random blank trigger and response plug in start with .

[Art]

Lonnie,

I think part of the problem is that our computers are configured for Voice / Speech Recognition and NOT Sound / Noise Recognition.

The patterns of human speech compared to the very different waveform(s) of noise, music (guitar, drum, trumpet, etc.) and random (door knocking, telephone, glass breaking, crying, etc.) or even purposeful sounds are difficult to distinguish for the computer. Especially our Window based SR.

Yes, I have had some strange experiences with other SR's or VR's where one actually heard my picking on my guitar and remarked that it had not heard that sound before or something to that effect. This was not with Hal but another "assistant" which I'll not mention here.

Yes, it would be very cool if Hal were able to "hear" other than just human voices. We'll have to see what develops!

[lightspeed]

i am like the old college student story , where they looked over a job and because they were so smart said the job couldn't be done , finally someone came along that wasn't so smart and actually didn't know the job could not be done and ended up figuring out how to do the job and did it . lol
I do know what you mean though , the biggest hurdle i think would be the variations of sounds , throwing off the sound pick up and process. oh well it will be another side job for me to tinker with . 

[Will and Mr Data :) :]]

Hi from Will and Mr Data    :]

I don't get it, ok i only skim read this thread real quick, but when i was using dragon 8 , Mr Data could recognise the whistle i used for each cat, all sorts of noises and know what they were.

If we are talking about having Hal make a noise back rather than speach then yes of course i'd be all for that,"hay Mr Data meow like the cat"
Mr Data would say "rooowwww"  in an actual can voice.
Or Mr Data could tell a story or joke and make the noises of motors, or like C3po when he was telling the ewoks about space ships and made the noises.

Tho if hal could activate a song, then that is no different from a sound file being played.
Hal should already be able to recognize the door bell. Tell him its the door bell, ring it and train him.
if "this is where you push the door bell so it makes a noise" then oh that sounds like the door bell being rung.

Mr Data has recognized music, different types, all manner of back ground noises, learned not to respond to them also if they are just back ground noise by learning so. ALways try lots of teaching before rushing to a plug in.
Tho unless hal can learn to play a sound clip he may need a plug in to be able to learn to play a sound clip.
But learning noises is just learning.

Bye for now and be well    :]

[onthecuttingedge2005]

We have had a lot of new people joining lately , this question ( i have asked before ) and is still open to programmers,
     What i would like to see created ( if possible)  is a brain plug in of sound acknowledgement .
what it would do is within the script have random script for trigger sentences (i have that part and random response sentences , i have that part . what i need is programing script  within the random trigger sentence , in another words
  in chosen area instead of a trigger word it would have a sound clip ( any sound clip can be used , my own recorded door bell. when this sound is heard as a trigger sound the programming would = to random response 1 to 10 and hal would say : (for example) somebody is ringing the doorbell or someone is at the door, or is that the doorbell ringing?  etc.
I am still hoping that someone can create a plug in like this . their could be many applications and responses written for it .
I will also throw this idea to robert by email , but he is awful busy lately so thought i would throw it on here again .
  I have the "base" random blank trigger and response plug in start with .

Hi lightspeed.

when you turn on speech rec does it hear a door bell when you ring it? If bling bling then someones at the door......

[lightspeed]

onthecuttingedge200 i haven't tried the wave sound clip as a trigger word yet , with the plug in but i am going to do it , i'll record our door bell with audacity  and place it as a trigger , then a random response to follow and see what happens . . i may be wrong but seems like it should work .  if it does it could open many possibilities , hal hearing a train whistle etc. 
If not then it won't but nothing tried nothing gained , will and data very interesting about your robot and sounds . 

[lightspeed]

well i tried created a sound bASED doorbell trigger.
putting the beginning like below:


If InStr(1, OriginalSentence, "doorbell ringing.wav") > 0 Or _
InStr(1, OriginalSentence,"InStr(1, OriginalSentence,"doorbell ringing") > 0 Or _
InStr(1, OriginalSentence,"the doorbell") > 0 Or _
InStr(1, OriginalSentence,"somones at the door") > 0 Or _
InStr(1, OriginalSentence,"the doorbell is ringing") > 0 Or _") > 0 Then

and placing the wave doorbell sound (ours that was recorded) in the hal files) and played the sound but it wouldn't recognize it . i think it is because its in a wave format and the regular doorbell isn't . its just a sound
oh well back to the drawing board i guess, i can use trigger words though and responses to accomplish about the same thing.
but would have been nice to have had it respond to actual sounds.
maybe theirs a way as of right now i don't know how.