Jump to content
C4 Forums | Control4

Multi-Tap Morse Code


bebster

Recommended Posts

Do you think it is possible to use a keypad and multi-tap it to generate Morse Code? I'm thinking that either composer code or a special purpose driver could interpret the multi-tap timings and create a text string.

That string could then be sent as a text message to a cell phone for example, or be used as a simple command to execute additional actions in programming.

For example,

.- or "A" would mean execute lighting scene A

.---- or "1" would mean play playlist #1 on the room's speakers

... --- ... or "SOS" coming from any keypad would be an instruction to flash all lights and send the "panic" command to the security system.

"@1.pick up more beer" would send text message "pick up more beer" to cell phone 1

Morse Code translator: http://www.onlineconversion.com/morse_code.htm

Link to comment
Share on other sites


This seems possible but I doubt it would work very well in the real world. Too much margin for error.

I think you would have problems with timing sequences from both the system and user input. You would have to determine the length of a press for a dot (i.e. a short press) and the length of a press for a dash (i.e. long press) as well as the length of time for a pause between dots and dashes for each character and a longer pause for the separator between words.

I think it would be easier to use a button press count to fire an event. Example .----- (1) is 6 different button presses. Why not just have one press of a button to play that playlist.

Here is

"@1.pick up more beer"

.---- .-.-.- .--. .. -.-. -.- ..- .--. -- --- .-. . -... . . .-.

That is a lot of button presses. I would just call the cell phone and tell them to pick up more beer.

I think this is a neat idea so by all means give it a try.

Link to comment
Share on other sites

For the amount of time it'll take to teach your wife, children, and houseguests morse code, you could probably get a side job that would make enough to just buy some 6-button keypads, or even touchpanels...

Seriously, the biggest issue with trying something like this is that there's no 'button was held X milliseconds' type functionality built-in. You could do it with a delay in-between taps, but the problem with that is the 'in-flight' time and processing time on the controller may be wildly varying, depending on the ZigBee traffic and business of the controller at the time you're trying to 'tap in'.

This idea is wildly imaginative, extremely impractical, and nearly impossible to implement.

My favorite kind of idea.

:)

To the person who suggested that Control4 build-in some voice recognition into each 2/3 button keypad: How much would you want to pay for such a device. Adding voice recognition would make the keypad pretty much as expensive as a touchpanel, which kind of defeats the point, since I can control the system much faster with a touchpanel.

RyanE

Edited to add: Even if you were to do morse code, it would be difficult to parse the characters into real commands, since you would have to track what character you've received, one at a time, and match them to commands. You could never have commands that are partial words of other commands. A *lot* more complexity there than you might think on first blush.

Link to comment
Share on other sites

^No, just the mic component integrated into the KP. Send the signalover the ACbus to a local controller housing the guts/driver. The cost is no different than your typical sub-system found on a mobile phone hands-free kit.

Link to comment
Share on other sites

OK, so let me get this straight:

1) Add a mic (which includes plastic tooling, too!)

2) Add components to digitize the mic data

3) Add networking components (i.e. not ZigBee, since latency would kill an audio app)

4) Add a speech recognition engine on the Control4 side.

5) Add all the integration code so a speech recognition engine would work in Control4's environment

6) ???

7) Profit !!!1!1!111!!!

And keep the cost of the keypad the same.

Hmm. Yeah, I'll get right on that.

Seriously, open-air microphone speech recognition is pretty hard to get right in the first place, and even if it were to happen, it likely would not happen on something like a keypad, but on something like a touchpanel, where at least there's already hardware, even if there's no software support ATM.

RyanE

Link to comment
Share on other sites

I've got that already, Alan. It's in my engine_compression_robot_ai.c4i DriverWorks driver.

It actually calculates the compression ratios for *any* engine produced in the last 200 years, including the Wankel, stirling heat engines, and the James Watt steam engine.

It was still easier to write than open mic voice recognition.

:)

RyanE

Link to comment
Share on other sites

OK, so let me get this straight:

1) Add a mic (which includes plastic tooling, too!)

2) Add components to digitize the mic data

3) Add networking components (i.e. not ZigBee, since latency would kill an audio app)

4) Add a speech recognition engine on the Control4 side.

5) Add all the integration code so a speech recognition engine would work in Control4's environment

6) ???

7) Profit !!!1!1!111!!!

And keep the cost of the keypad the same.

Hmm. Yeah, I'll get right on that.

Seriously, open-air microphone speech recognition is pretty hard to get right in the first place, and even if it were to happen, it likely would not happen on something like a keypad, but on something like a touchpanel, where at least there's already hardware, even if there's no software support ATM.

RyanE

Your right, go with the Morse code thing instead...

Link to comment
Share on other sites

While I agree that speech recognition is not a practical solution, it is not necessarily as complicated as Ryan listed. If you do simple phrase recognition rather than free speech, then a low end processor can handle the load. Any of the current controllers would probably handle it, and the HC-1000 would be able to handle many streams at once. You would need a decent quality mic, but it wouldn't really be all that expensive. The big issue (which Ryan pointed out as well) is connecting the mic to the controller to do the speech recognition. Having to rip up the wall and add in new wires kind of defeats the purpose, and the potential for signal loss is rather large. So then we're back to morse code.

Link to comment
Share on other sites

I'm not even saying that 'free phrase' is difficult, even though it is more difficult than 'command and control', which is what you propose.

I'm saying that 'free air' microphones pick up a *lot* of ambient noise, and is much more difficult to make work than a dedicated (i.e. headphone with mic / phone system) microphone is.

RyanE

Link to comment
Share on other sites

I have a hard time believing speech recognition would be THAT hard considering I just bought an app for my blackberry that pretty accurately writes text messages for me

It is not so much that speech recognition is hard to do as much as it is the cost of adding speech recognition to a C4 dimmer/switch and 2/3/6 button keypads.

Example: It could cost ten of thousands of dollars to have the molds used to make the button on just a dimmer changed to add little holes for the required microphone.

When all the required modifications are made to add speech recognition to the dimmers and switches and keypads are done the cost of a dimmer will go up. Would you want to pay (as an example) $50.00 to $100.00 more for each dimmer for speech recognition?

After you paid extra money for the dimmers you would need to either add Ethernet wire to each dimmer location (costly) or install a GOOD wireless network (also costly) to transport the encoded voice commands back the the C4 controller.

In summary.

Adding speech recognition to a blackberry which already has a microphone, A/D converter, fast enough CPU and memory and is designed to process speech is fairly easy since it is mostly software changes.

Adding speech recognition to a light switch which has non of the required hardware is that hard and that costly.

One other thing to said

that pretty accurately writes text messages for me

"pretty accurately" won't cut it to control a home automation system.

If speech recognition was that easy and that accurate and that cheap we would see many more devices that use speech recognition.

Link to comment
Share on other sites

I wasn't thinking of putting it in the ligth switch they aren't generally centrally located in the room and would be difficult to speak to, I was thinking in the controllers. Just wait till the next development cycle and add a mic in with an array mic mounted somewhere. If the homeowner has the controllers centrally located then they could make a mic module specifically for it.

About the "pretty good" speach recognition thats with free form text they could start with command lists, although those are admittedly less upscale they wuld do the job very accuretely.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.