Bitshifting std_logic_vector while keep precision and conversion to signed

Question

In VHDL I want to take a 14 bit input and append '00' on the end to give me a 16 bit number which is the 14 bit input multiplied by 4 and then put this into a 17 bit signed variable such that it is positive (the input is always positive). How should I go about this?

like this? shiftedInput <= to_signed('0' & input & '00', 17);

Or maybe like this? shiftedInput <= to_signed(input sll 2, 17);

Or this? shiftedInput <= to_signed(input & '00', 17);

Does it see that the std_logic_vector it's getting is 16 bit and the signed variable is 17 bit and therefore assume the most significant bit (the singing bit) is 0?

Or do I have to do this? shiftedInput <= to_signed('0' & input sll 2, 17);

e.g. If I read in the 14 bit number 17 as a std_logic_vector [i.e. (00 0000 0001 0001)] it should be converted to the signed number +68. [i.e. (0 0000 0000 0100 0100)]

What is the type of input? What errors do you encounter with your various attempts? — Renaud Pacalet
1st option gives: no function declarations for operator "&" 2nd option gives: no function declarations for operator "sll" 3rd option gives: prefix is neither a function name nor can it be sliced or indexed 4th option gives: no function declarations for operator "sll" — SomeRandomPhysicist

Renaud Pacalet Renaud Pacalet · Accepted Answer · 2016-08-21T17:05:10

std_logic_vector is compatible with the type signed of numeric_std. So, the type conversion function is signed (not to_signed that converts between integers and vectors):

shiftedInput <= signed('0' & input & "00");

should make it. Note the "00" instead of your '00'. Bit strings are double-quoted while bits are single-quoted.

Bitshifting std_logic_vector while keep precision and conversion to signed

1 Answers