Samsung36 extra gap

Barf · Post by **Barf** » Sun Oct 16, 2016 3:51 am

Our "official" IRP for the Samsung36 protocol reads

{38k,500}<1,-1|1,-3>(9,-9,D:8,S:8,1,-9,E:4,F:8,-68u,~F:8,1,-118)+[D:0..255,S:0..255,F:0..255,E:0..15]

It recently struck me that this description is both implausible and practically awkward for decoding.

The problem is the 68 microseconds gap ("-68u"). It is very implausible that this is a deliberate design from a protocol designer: The bulk of the signal consists of flashes of 500us (micro seconds) duration and gaps of either 500 or 1500us. Then suddenly, a gap of 568 (or 1568) microseconds appears instead of 500 or 1500! It appears very unlikely that this is deliberate. Most receivers (?) would just lump that measurement onto 500 or 1500's.

Also, it makes the recognition of a signal much harder, since the IRP objects are no longer uniquely corresponding to the durations (the last gap of F:8 is being lumped into the -68u).

Comments?

I have not been able to locate a authoritative document describing the protocol. The closest is this one, which does not support the significance of the 68us extra gap..

Post by **mathdon** » Sun Oct 16, 2016 8:13 am

Barf, I think you may have an out-of-date version of DecodeIR.html. The version in build 14 (the final build) of RMIR v2.03 and in all development versions of RMIR v2.04 has this gap omitted:

Code: Select all

{38k,500}<1,-1|1,-3>(9,-9,D:8,S:8,1,-9,E:4,F:8,~F:8,1,-118)+

Barf · Post by **Barf** » Sun Oct 16, 2016 9:23 am

Thanx Graham. I see that the version in SVN is the "sane" version.

The good news is that there is a consensus on the correct version. The bad news is that the "non-sane" version is in official places, like Wiki and the the latest version of DecodeIR.

Post by **mathdon** » Sun Oct 16, 2016 10:25 am

I don't know who deals with the Wiki. DecodeIR is currently maintained by 3FG.

mdavej · Post by **mdavej** » Sun Oct 16, 2016 10:36 am

I just fixed the wiki

3FG · Post by **3FG** » Sun Oct 16, 2016 9:31 pm

OK, I uploaded the revised DecodeIR.html into the "official" zip file.

However, I think we should be using SourceForge to keep the all of the binaries, docs, and source. When we revised some of the content of DecodeIR.html last January, I checked it in to SF. And the source has been checked in there since Jan 2015. The issue I see with multiple locations is it's easy to get them out of sync. For example, while Dave has fixed the Samsung entry in the Wiki, overall it has obsolete content.