Convert 0.008 to binary?
56 views (last 30 days)
Show older comments
Since the function dec2bin only works for integers, how can I convert 0.008 to binary?
4 Comments
Walter Roberson
on 2 Mar 2021
signed is 0 if you want unsigned values and 1 if you want signed values. https://www.mathworks.com/help/fixedpoint/ref/embedded.fi.html#mw_0d34f1fa-0f13-43f0-b8b3-5878c069fa9f
Accepted Answer
the cyclist
on 14 Sep 2012
There are several submission in the File Exchange that do this. I have not used any of them, but here is one that is highly rated: http://www.mathworks.com/matlabcentral/fileexchange/13899-floating-number-conversion-to-binary-and-vice-versa
As Walter mentions, you need to take care about competing conventions.
2 Comments
More Answers (1)
Walter Roberson
on 14 Sep 2012
There is no one way to convert a decimal fraction to binary. There are a large number of competing conventions. You will need to specify exactly what representation you want to use.
1 Comment
Walter Roberson
on 4 Oct 2017
The people who ask this question are often dealing with values that could potentially be negative. That leads into the question of how to represent negative binary values. Some of the competing conventions include:
- two's complement. This is the most commonly used convention for integers. The negative of positive integer value P is found by taking the bitwise complement of P, and adding 1 to the result.
typecast(int16(-123),'uint16')
ans = 65413
For strictly negative value N, this is
TC16 = @(N) uint16( double(intmax('uint16')) + double(N) + 1 )
- one's complement. This is not as commonly used except sometimes in some communications theory. The negative of positive integer value P is found by taking the bitwise complement of P, with nothing added to the result.
For strictly negative value N, this is
OC16 = @(N) uint16( double(intmax('uint16')) + double(N) )
- separate sign. This is the most common arrangement for floating point representations. One bit is reserved to indicate whether the value is positive (typically the bit is 0 for this) or negative (typically the bit is 1 for this), but otherwise the positive representation of the integer is used.
For strictly negative value N, this is
SS16 = @(N) uint16(2^15) + uint16(abs(N))
However, notice the above assume fixed-width representation, which is not always what people are asking for when they are asking about converting decimal fractions to binary. It is not uncommon for people to be asking for some kind of minimal representation, that ends at the point where the remaining bits would all be 0. This requires more complex representations, because the representation has to somehow encode "the following is N bits wide" where N is a changing number depending on the value; or "the representation ends at this point". And that starts to get messy. It is not uncommon for fixed-width representations to be more efficient than variable-width representations, unless there are large differences in the frequency distributions that happen to favor the shorter items.
See Also
Categories
Find more on Logical in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!