CODEWITHSUNDEEP
vhdlfpga
emmaw
emmaw

Reputation: 35

How can i use floating point numbers in VHDL?

I'm trying to create RGB to YUV converter in VHDL. 

Y = 0.299 x R + 0.587 x G + 0.114 x B

U = -0.147 x R - 0.289 x G + 0.436 x B

V = 0.615 x R - 0.515 x G -0.100 x B

In this formula I have to use and multiple floating point numbers like 0.299, 0.587, etc. How do I create synthesizable VHDL with floating point numbers?

Upvotes: 2

Views: 6343

Answers (1)

Holminge
Holminge

Reputation: 120

you can this using integer multiplication.

If you have a FPGA with 18 bit multipliers you multiply and then use the upper bits.

E.g. : 0.299, R=20 =>

0.299 * 2^17 = 39191

39191 * 20 then use the result's upper bits ( top downto 17) that is the same as divide by 2^17

Since your application is RGB to YUV I guess this is time critical so the above example is the best choice, otherwise I usually use Instant SoC to implement float algorithms that is not time critical.

Upvotes: 4

Related Questions