signal processing - Input of a fixed point DSP -


i'm new working dsps , fixed point , need know: 1. fixed point dsp converts float number q format or device before feeding dsp? 2. specifies q format used. each dsp come specified q_format or programmer in codes. 3. can have idea of how perform simple 4 4 fixed point matrix multiplication in c++?

thanks in anticipation

the format fixed given dsp, e.g. motorola dsp 56k family uses 24 bit signed fractional format (q23).

fixed point same ordinary integer there's implicit scale factor. operations makes no difference, e.g. load/store/add/subtract work same way regardless of whether data integer or fixed point.
when comes multiplication or division implicit scaling factor needs taken account - typically there shift after operation correct this. dsp instructions take care of automatically, whereas normal cpus have explicitly.

when you're doing e.g. 4x4 matrix multiply use dsp's native fixed point arithmetic instructions , scaling taken care of automatically.


Comments

Popular posts from this blog

html - How to style widget with post count different than without post count -

How to remove text and logo OR add Overflow on Android ActionBar using AppCompat on API 8? -

javascript - storing input from prompt in array and displaying the array -