To my knowledge, the protocol we call NEC1 is specified in official documents. It is also described in a number of places on the Internet, for example [Link: sbprojects.com] . (I even think that the firm NEC (now called Renesas) asks for a license fee if you implement it in a commercial product with D != 0.)

The procotol NEC1 contains 32 "payload" bits, which can be decomposed into four bytes. First one is device address, called D by IrScrutinizer and DecodeIR. The second one was originally just the (one-) complement of the first, but as more "addresses" were sought, it aquired its owm "life" as "subdevice", S. The third byte is the command number, F. The forth byte the complement of the third, although some manufacturers (Yamaha, possibly others) has used it differently. Note that by convention AND specification, these bytes are interpreted least significant bit first.

Some projects interpret the payload bits differently. For example Arduino-IRremote [Link: github.com] considers the number of bits variable (up to 32) and lumps them all into one number, interpreted most-significant-bit-first.

Why Lyngdorf selected their format you have to ask them. Possibly the last two "numbers" of the "short pronto form"?

I have put a file in Girr format (can be directly read into IrScrutinizer, and then transformed to other formats) here:  [Link: raw.githubusercontent.com]