Fair enough, the term "sign bit" has specific meaning and I used it out of this meaning in a more general way to denote a bit who's value isn't numeric. Point taken.WinterMute wrote:The term "sign bit" has a very specific meaning which does not include the sign offset in an integer value. Floating point representation has a sign bit as do some DSP chips, most processors use 2s complement to represent signed numbers. This is neither semantics nor unimportant and drives me nuts when people perpetuate this kind of misinformation.
However, with regard to zeromus' previous assertion that in the naming of fixed point number formats, though "there is no real standard and there is some ambiguity", they are named without explicitly making reference to the bit that is non-numeric in 2s compliment signed intergers, there seems to be more ambiguity than I realised. In the libnds math.h file reference (http://libnds.devkitpro.org/a00130.html) in the description of functions crossf32, dotf32 and normalizef32, fixed point numbers are described in the form 1.19.12 fixed point cross product function result...
Is this a special matrix form of using fixed point numbers? Or does the first value denote a non-numeric bit?