Bi-directional motion compensation method to reduce the number of registers

A method for bi-directional motion compensation using a register to calculate a pixel data of a present frame based on a first predictor, a second predictor, and a predicted error, said method comprising:

• adding said predicted error twice, said first predictor, and one to obtain a first calculation result;
• wherein if said first calculation result is larger than a maximum positive value that is storable in said register, either said maximum positive value or a value of said maximum positive value minus one is stored in said register, otherwise said first calculation result is stored in said register;
• adding said value stored in said register and said second predictor to obtain a second calculation result;
• wherein if said second calculation result is larger than said maximum positive value that is storable in said register, either said maximum positive value or a value of said maximum positive value minus one is stored in said register, and if said second calculation result is smaller than zero, a zero is stored in said register, otherwise said second calculation result is stored in said register;
• and
• right-shifting said value one bit to obtain said pixel data.