Re: Number truncation, round off and saturation in Fixed Point Arithmetic



On Jun 19, 3:10 am, Abhishek <abhisheksgum...@xxxxxxxxx> wrote:
Hi,
I have the following specific doubts regarding fixed point arithmetic.
1) If I have Q(1.15) meaning one sign bit and 15 fractional bits, and
If I also have another unsigned 8 bit number represented in a 16 bit
number with leading 8 zeros in it, I then multiply these two treating
the second number as a signed 16 bit number or Q(16.0) format (.i.e. 1
sign bit and 15 intger bits. here sign bit is zero).

Whats the result size and what would be its format
after multiplication? Can anyone explain whats happening here clearly?
Now once it is multiplied, the author says it is in Q17.15
format(which you need to explain how he got it plzzz). Then he
converts Q17.15 result of multiplication to 16 bit signed integer by
right shifting the number by 15 bits and extracting the lower 16 bits
of the right shifted result.
Can anyone explain in detail whats happening here and exactly why did
he follow these steps?
Suitable examples and a clear explanation would be highly
appreciated :-)
Thanks

With Regards,
Abhishek S

see section 2 of

http://www.superkits.net/whitepapers/Fixed%20Point%20Representation%20&%20Fractional%20Math.pdf

But I don't think any explanation is going to make you happy until you
get a bit piece of paper and do the multiply yourself. Its somewhat
akin to explaining to a kid why 5 + 5 = 10, you just have to get the
kid 10 apples and work it out himself....

.



Relevant Pages