Wednesday, 3rd June 2009
I've rewritten the remote control signal decoding software to handle multiple protocols. As well as SIRCS, it now supports RC-5, NEC, JVC and two Panasonic codes (one "old" 11-bit code and one "new" 48-bit code). There's not much in the way of screenshots at the moment, other than a debug window that gets filled when keys are pressed:
NecCommand Address=24, Command=87, Extended=False, Repeat=1
RC5Command Address=20, Command=53, Repeat=True, Repeat=1
NewPanasonicCommand OEM Device 1=2, OEM Device 2=32, Device=144, Sub-Device=0, Command=10, Repeat=1
RC5Command Address=8, Command=35, Repeat=True, Repeat=1
OldPanasonicCommand Address=0, Command=20, Repeat=1
SircsCommand Address=2362, Command=121, Length=20, Repeat=1
SircsCommand Address=7002, Command=84, Length=20, Repeat=1
NecCommand Address=64, Command=146, Extended=False, Repeat=1
NecCommand Address=81, Command=8, Extended=False, Repeat=1
JvcCommand Address=3, Command=23, Repeat=1
Gripping stuff, I'm sure you'll agree.
The C# source code for this can be downloaded here.
A keyring remote control (courtesy of Poundland) has highlighted one possible issue in handling repeating buttons. Rather than target any particular device, it will try and brute-force a response. For example, here's the result of pressing the power button once in one particular mode:
SircsCommand Address=1, Command=21, Length=12, Repeat=2
OldPanasonicCommand Address=0, Command=32, Repeat=1
NecCommand Address=32, Command=11, Extended=False, Repeat=1
NewPanasonicCommand OEM Device 1=2, OEM Device 2=32, Device=128, Sub-Device=0, Command=61, Repeat=1
That's four different protocols from one button. I suppose some sort of mapping from protocol-specific code to a string (so those five commands would be translated into five "power" strings) and comparing the time between signals to turn the input into something meaningful may help, but that would require an enormous database of known codes.