Table of Contents
Advertisement
MB86R02 'Jade-D' Hardware Manual V1.64
18.9.2 Data Format
18.9.2.1
Data format
The supported data formats are 32-bit single-precision floating-point format, 32-bit fixed-point format,
integer packed format and RGB packed format. All internal processing is performed in the floating-
point format. For this reason, the integer packed format, fixed-point format and RGB packed format
must be converted to the floating-point format. The processing speeds in these formats are slightly
lower than in the 32-bit single-precision floating-point format.
The data format to use is selected by setting the GMDR0 register.
(1) 32-bit single-precision floating-point format
31 30
s
s: Sign bit (1 bit)
e: Exponent part (8 bits)
f: Mantissa (23 bits): '1.f' shows the fraction. '1' is a hidden bit.
The numerical value of the floating-point format becomes (-1)
(2) Signed fixed-point format (SFIX16.16)
31 30
s
s: Sign bit (1 bit)
int: Integer (15 bits)
frac: Fraction (16 bits)
(3) Signed integer packed format (SINT16.SINT16)
31 30
s
s: Sign bit (1 bit)
int: Integer (15 bits)
(4) RGB packed format
31
reserved
R, G, B: Color bits (8 bits)
(5) ARGB packed format
31
A
A: Alpha bits (8 bits)
R, G, B: Color bits (8 bits)
23 22
e
Int
Y.int
24 23
R
24 23
R
f
s
(e-127)
(1.f)2
(0 < e < 255).
16 15
Frac
16 15 14
s
X.int
16 15
8 7
G
16 15
8 7
G
0
0
0
0
B
0
B
18-147
Advertisement
Table of Contents
