JP1 Remotes

The Robman · Posted: Wed Jul 29, 2009 11:37 am Post subject:

It looks good, though you're missing the leadout time (which should be about -77810 uS). It took me a moment to figure out what the "7:3,0:3,1:1" part was all about, I think I'd prefer to see the "+930 -930" leadin represented differently.

I'm not confident that I know how to write IRP, but I think it should look something like this:

Telefonica {56.3k,310,msb}<1,-3|-1,1,-2|-2,1,-1|-3,1>(3,-3,1,D:8,S:8,F:8,C:8,-251)+
_________________
Rob
www.hifi-remote.com
Please don't PM me with remote questions, post them in the forums so all the experts can help!

mathdon · Posted: Thu Jul 30, 2009 3:21 am Post subject:

You're absolutely right, of course, Rob. I was trying to unscramble this protocol in my mind, but the protocol won. Scrambled brains! Laughing

What I was trying to do was indicate that the pulse trains for the bit pairs were not as unsystematic as they first seem, but I encoded the wrong pulse times. It was intended to come out as

Telefonica {56.3k,310,msb}<-1|1>(3,-3,1,<8:4|4:4|2:4|1:4>(D:8,S:8,F:8,C:8,^77m))+

It would seem even more systematic if one could write

Telefonica {56.3k,310}<-1|1>(3,-3,1,<1:4|2:4|4:4|8:4>{msb}(D:8,S:8,F:8,C:8,^77m))+

as the encoding of the bit pairs makes more sense in lsb form while the encoding of the data in those bit pairs is in msb form. But I've never seen anything John has written looking quite like either of these, so I await his judgement.

While I'm seeking John's view, could I also ask about the notation IRP uses for lead-outs. I've written ^77m there, which I think means "upwards of 77 milliseconds" but I'm not really sure of the meaning of the ^. There are also lead-outs given as "???". Does that mean "unknown" or does it have some quantitative meaning?
______________
Graham

johnsfine · Posted: Thu Jul 30, 2009 7:04 am Post subject:

Changing the bit definitions mid stream, even using the outer to define the inner, is fine.

I don't like adding {msb} mid stream. There is a format for reversing an individual field (but I forget what it is). In your case, if the important data is msb why do you need to start with lsb?

In irp notation, the ^ for lead out time means the lead out makes up the difference such that the total back to the last open paren adds up to that time. If you want a simple, fixed duration lead out, that is just - instead of ^.

There is no ??? in irp notation. There is just a person who got lazy when writing many of the irp strings and didn't bother to compute some values and wrote ??? instead.

The Robman · Posted: Thu Jul 30, 2009 7:31 am Post subject:

Also, I believe, that if you don't specify a unit for the leadout time, it's treated as multiples of the time base, so in my example as I was aiming for 77810 uS I entered -251 (ie, 251 * 310 = 77810).
_________________
Rob
www.hifi-remote.com
Please don't PM me with remote questions, post them in the forums so all the experts can help!

mathdon · Posted: Thu Jul 30, 2009 11:06 am Post subject:

The Robman · Posted: Thu Jul 30, 2009 11:13 am Post subject:

I believe this protocol uses a fixed leadout time, rather than an "off as total" leadout time, so you don't want to use the "^" prefix on the leadout time.

Also, leadout times don't need to be that accurate, and you can certainly be off by more than 310 uS and still be correct, so using a multiple of 310 is fine. Having said that, I personally prefer to see the leadout time specified as a time rather than a multiple because it saves the user the extra step of needing to convert it.
_________________
Rob
www.hifi-remote.com
Please don't PM me with remote questions, post them in the forums so all the experts can help!

mathdon · Posted: Fri Jul 31, 2009 6:07 am Post subject:

I have finished up putting the IRP for Telefonica (and Sejin) in the forms

Telefonica {56.3k,310,msb}<-1|1>(3,<8:4|4:4|2:4|1:4>(3:2,D:8,S:8,F:8,C:8,-77m))+
Sejin {38.8k,310,msb}<-1|1>(3,<8:4|4:4|2:4|1:4>(3:2,D:8,S:8,F:8,C:8,-77m))+

where C = D:4:0 + D:4:4 + S:4:0 + S:4:4 + F:4:0 + F:4:4.

Rob, you may think it a step backwards to put the 3:2 at the start of the data, instead of -3,1 before the data, but this is best for decoding. If you treat -3,1 as part of the leadin then the data may start in the middle of an ON pulse, which is difficult to handle. Having 3:2 in the data both gives data that starts at an ON/OFF boundary and gives an easy check on that OFF/ON pair. Also, it isn't very obvious that -3,1 does correspond to a bit pair and this makes it clearer to any future coders of protocol executors or decoders.

I'm only giving the output from one of the files, but the others do decode properly too. Telefonica and Sejin are distinguished by frequency. The first Telefonica file decodes as

The Robman · Posted: Fri Jul 31, 2009 2:02 pm Post subject:

mathdon · Posted: Fri Jul 31, 2009 5:03 pm Post subject:

Rob, I'm completely happy with everything you suggest. I know absolutely nothing about the origin of these protocols, manufacturers etc. My names are virtually random and I've been waiting for someone to tell me better ones. My only questions are:

Is John happy with <-1|1><8:4|4:4|2:4|1:4> without any intervening brackets, and

What is a "break string"? Is that the "Extra" frame? As I've said elsewhere, due to DecodeIR not getting the Extra frame I've no way of distinguishing its presence or absence.

Please remember that, so far, I know nothing about PB or what's in the executors. I've just been working from the .ir files people have posted. So if you, or John, or anyone, can tell me the full flexibility of the $0161 executor in terms I currently understand, such as in IRP notation, I will be happy to extend/modify the decode to meet it.
_______________
Graham

The Robman · Posted: Fri Jul 31, 2009 5:48 pm Post subject:

The term "break" comes from infrared keyboards where the protocols tend to have a "make" (ie, sent once when the button is pressed) and a "break" (ie, sent when the button is released) portion.

I'm using the term to refer to the portion of the signal that gets sent when the button is released. This portion would normally show up as an "extra" in learned signals, but that doesn't mean that everything that turns up in "extra" is a "break" signal.

Actually, I think you do need a set of brackets in this IRP, so my suggestion would be:

Sejin-38 {38.8k,310,msb}<-1|1>(<8:4|4:4|2:4|1:4>(3,3:2,D:8,S:8,F:8,C:8,-77m))+
Sejin-56 {56.3k,310,msb}<-1|1>(<8:4|4:4|2:4|1:4>(3,3:2,D:8,S:8,F:8,C:8,-77m))+

So all I'm really doing is moving the '3' to the main section rather than have it all alone up front.

Also, protocols and executors don't have to have a 1-to-1 relationship. We can decide that an executor is capable of generating more than one protocol, if we wish, and we have certainly seen protocols generated by more than one executor.

I'm assuming that you can read assembler, so at some point, you should take a look at an executor to see how it works, if for no other reason than curiosity. Plus, for simple protocols you don't even need assembler as PB does it all for you.
_________________
Rob
www.hifi-remote.com
Please don't PM me with remote questions, post them in the forums so all the experts can help!

mathdon · Posted: Sat Aug 01, 2009 3:37 am Post subject:

The Robman · Posted: Sat Aug 01, 2009 11:23 am Post subject:

You've certainly been keeping yourself busy in the JP1 world! As if taking on IR.exe wasn't enough (and I know it turned out to be more than you expected) now you're taking on DecodeIR. Amazing.

You can get a copy of the S3C8 instruction sheet here:
http://www.hifi-remote.com/files/
_________________
Rob
www.hifi-remote.com
Please don't PM me with remote questions, post them in the forums so all the experts can help!

mathdon · Posted: Sat Aug 01, 2009 2:56 pm Post subject:

I got hooked! I did only start on IR.exe as it didn't cope very well with my remote, a URC-7781, but since I got to understand the program I thought I would do a bit more, to help others. And I sort of stayed! But IR.exe uses DecodeIR, and that was getting out of date since John couldn't spare the time, so one thing led to another! I just love the challenge of programming!
_____________
Graham

The Robman · Posted: Sat Aug 01, 2009 3:46 pm Post subject:

We have plenty here to keep you busy, so tell your friends and family that you may be away for a while! Smile

_________________
Rob
www.hifi-remote.com
Please don't PM me with remote questions, post them in the forums so all the experts can help!

The Robman · Posted: Mon Aug 03, 2009 9:41 am Post subject:

Hey Graham,
If you'd like to see an example of an un-needlessly complicated checksum, take a look at the 4DTV protocol.

http://www.hifi-remote.com/forums/viewtopic.php?t=8373
_________________
Rob
www.hifi-remote.com
Please don't PM me with remote questions, post them in the forums so all the experts can help!