#40 IRRecord quits when learning the buttons.

[e5frog]

Have a 1838 3V (that I stole from an old cable box) hooked to the audio in with its own powersource (2032 battery), running Windows XP SP3 32bit, stationary computer. Data looks fine in IRGraph and excellent when recording sound, most stable lines in IRGraph when when in 96kHz mode.
I am able to get to the part where I start typing names of buttons and holding them. Occasionally I get retries and suddenly program just quits, no visible error messages. So I have to start all over again.

It gets a bit pointless after a few times.

I quit after just one detection once and got a config file - but I guess it's a matter of copy and paste of one button at the time going through the calibration process every time.

This is the data I got for one button, values have been pretty consistent, it seems similar to the existing EUR7628030AR (looks the same apart from upper right button) , difference is mine is detected to be 48bits so it has 4004 before the power code in that config file.

# brand:                       C:\Program\WinLIRC\Recordings\EUR7628010
# model no. of remote control: 
# devices being controlled by this remote:
#

begin remote

  name  C:\Program\WinLIRC\Recordings\EUR7628010
  bits           48
  flags SPACE_ENC
  eps            30
  aeps          100

  header       3563  1688
  one           468  1274
  zero          468   401
  ptrail        463
  gap          74463
  toggle_bit_mask 0x0

      begin codes
          power                    0x40040100BCBD
      end codes

end remote

Actually they're probably the same, the AR version has this:
pre_data_bits 16
pre_data 0x4004
gap 74419
toggle_bit_mask 0x0

So it actually adds the 4004 in front of all codes...
Nice to get a confirmation, code should work but my transmitter that I tried yesterday didn't I guess.

Anyway, IRRecord crashes for me, doesn't matter if I turn anti-virus off and about everything else and set priority high on both that and WinLIRC.
Computer is AMD Opteron 180, 2.4GHz, 4GB RAM (3.25GB usable)

I have included the error report from Windows, perhaps someone knows how to interpret it.
"Error signature" is this:
AppName: irrecord.exe AppVer: 0.0.0.0 ModName: audiocapture.dll
ModVer: 0.0.0.0 Offset: 000010f2

I'll try again and see if the error pops up the same way, will in that case update...

Thanks for what seems to be a nice port.

[Ian]

Try recording in raw mode
Think -f command line
http://www.lirc.org/html/irrecord.html
The irrecord was a very rough port of the linux version. The linux version frankly was a mess, and it would be hard to debug unless I could replicate the exact error

[e5frog]

I thought I already wrote a reply here... where did that go.
I rebooted computer and started a cmd-window, there I had no problems reading the entire controller with the command "irrecord -d plugins/audiocapture.dll file".
Will try later to go via WinLIRC and see if it works.

It's not a very userfriendly collection of programs, IRGraph works nicely - no problems there.
Need to try and get a transmitter working, best was the "direct to COM-port"-version, the one that adds NPN, PNP and capacitors to the mix had terrible range on my COM-port. The laptop I used seem to have a low voltage swing, not sure what it's supposed to be.

Seems I need to write my own application to actually be able to use it to transmit. WinLIRC should have a new window opening with all defined buttons as klickable buttons instead of the slow current method, shoosing one button from drop down and then click a button.
The dream is all controllers as scanned images and you'd just scroll or press a hot key for a controller and then you'd get the image of that controller and click needed button. Maybe with some hotkeys per controller.

Thanks for the tip, will play around some more and update here if I have any more news about the IRRecord exit problem.

EDIT:
I recommend running the forced RAW mode to anyone, it worked a lot faster, it doesn't seem to do optimizations like the normal mode (extracting initial bits that are same for all codes) but that hardly matters for users I think.

[e5frog]

I had the same problem again, twice, this time I was running from cmd, actually via a .bat file that's supposed to -u (update) a file that is dropped on the -bat file, quite handy.
I managed to read 11 buttons before it crashed the first time. When trying again I get an error on a buttton that is a bit tricky and fails sometimes - and it quits. Same "Error signature" as before:
AppName: irrecord.exe AppVer: 0.0.0.0 ModName: audiocapture.dll
ModVer: 0.0.0.0 Offset: 000010f2

I'll try -f and then do the -a (analyze) I guess...