floating point to fixed point conversion
in [ADA]
|
From: riya1012 on 21 Feb 2006 17:15 hello guys, I need some help from you. I am doing a DSP project and for that I need to do some C coding for the conversion of sample data which is in floating point representation to fixed point representation. the sample data is in floating point like 0.224128 2.299965 0.448350 -1.779926 My DSP algorithm is implemented in C and is supposed to be using fixed point representation. The above data is intended to be converted to fixed integer format.I request you to help me out regarding this conversion.I will be very glad if u give me some hints or algorithms for this conversion.
From: Jeffrey R. Carter on 21 Feb 2006 18:49 riya1012(a)gmail.com wrote: > I need some help from you. I am doing a DSP project and for that I need > to do some C coding for the conversion of sample data which is in > floating point representation to fixed point representation. This is c.l.ada. We're for engineering SW in Ada, not for "C coding". > the sample data is in floating point like > 0.224128 > 2.299965 > 0.448350 > -1.779926 > My DSP algorithm is implemented in C and is supposed to be using fixed > point representation. > The above data is intended to be converted to fixed integer format.I > request you to help me out regarding this conversion.I will be very > glad if u give me some hints or algorithms for this conversion. I have no idea how to convert floating-point values to fixed point in C, and this does not seem to be the right place to be asking about that. As far as I know, C does not have fixed-point types. Ada does; yet another reason to use Ada. If you get a useful answer about C coding here, it will be pure luck. I suggest you try c.l.c. If I'm misinterpreting your question, and this does have something to do with Ada, perhaps you can rephrase it in a manner that makes that clear. -- Jeff Carter "Death awaits you all, with nasty, big, pointy teeth!" Monty Python & the Holy Grail 20
From: Keith Thompson on 21 Feb 2006 18:51 riya1012(a)gmail.com writes: > I need some help from you. I am doing a DSP project and for that I need > to do some C coding for the conversion of sample data which is in > floating point representation to fixed point representation. > the sample data is in floating point like > 0.224128 > 2.299965 > 0.448350 > -1.779926 > My DSP algorithm is implemented in C and is supposed to be using fixed > point representation. > The above data is intended to be converted to fixed integer format.I > request you to help me out regarding this conversion.I will be very > glad if u give me some hints or algorithms for this conversion. Um, why are you asking about C coding in comp.lang.ada? -- Keith Thompson (The_Other_Keith) kst-u(a)mib.org <http://www.ghoti.net/~kst> San Diego Supercomputer Center <*> <http://users.sdsc.edu/~kst> We must do something. This is something. Therefore, we must do this.
From: Keith Thompson on 21 Feb 2006 19:51 Keith Thompson <kst-u(a)mib.org> writes: > riya1012(a)gmail.com writes: >> I need some help from you. I am doing a DSP project and for that I need >> to do some C coding for the conversion of sample data which is in >> floating point representation to fixed point representation. >> the sample data is in floating point like [snip] > > Um, why are you asking about C coding in comp.lang.ada? As well as comp.arch.embedded, comp.lang.c, comp.graphics.algorithms, comp.arch.arithmetic, and comp.dsp. -- Keith Thompson (The_Other_Keith) kst-u(a)mib.org <http://www.ghoti.net/~kst> San Diego Supercomputer Center <*> <http://users.sdsc.edu/~kst> We must do something. This is something. Therefore, we must do this.
From: stuart clark on 22 Feb 2006 06:01
Not sure if this is what you are after but an example from a real IDD, where floats are represented as 16 bit signed integers, the 16th bit is the sign bit. Eg from a real life navigational computer IDD magnetic heading = +/-180 degrees. bits to use = 15, the 16th bit the sign bit pass integers across an interface, which are then converted to floats using a known scaling factor. scaling factor = 180 / (2^15 ) = 0.005493.... we define the integer value of 1 = 0.005493.... and we call this value the LSB. >From basics where 15 bits can represent a max value of 2^15 -1 we now have max value of 180 - LSB (iw 2^15 =180 and 1 =LSB) therefore to send 180 - LSB across an interface what integer value do you need = 180 / scaling factor = (180 - 0.005493..) / 0.005493 = 32767 (the old 2^15-1) = 0x7FFF 90 degrees = 90 / 0.005493 = 16383.5 = 0x3FFF but note there is an accuracy loss as 0x3FFF * 0.005493 = 89.997 Thats hows its done in real life, hope it helps! Stuart Clark |