Lecture 2 – Binary Data Representation

CNG111 - Lecture 2 1/23

 Objective

• Binary Data Representation

– .. of integers, real numbers, characters etc.

CNG111 - Lecture 2 2
 Data Representation

• Based on 1s and 0s
– So, everything is represented as a set of binary
numbers
• We will now see how we can represent:
– Integers: 3, 1234435, -12945 etc.
– Floating point numbers: 4.5, 124.3458, -1334.234 etc.
– Characters: /, &, +, -, A, a, ^, 1, etc.
– …

CNG111 - Lecture 2 3
Binary Representation of Numeric Information
• Decimal numbering system
– Base-10
– Each position is a power of 10
3052 = 3 x 103 + 0 x 102 + 5 x 101 + 2 x 100

• Binary numbering system

– Base-2
– Uses ones and zeros
– Each position is a power of 2
1101 = 1 x 23 + 1 x 22 + 0 x 21 + 1 x 20

CNG111 - Lecture 2 4
 Decimal-to-binary Conversion
• Divide the number until zero:
Example: Convert decimal value 35 to binary
–35 / 2 = 17x2 + 1
–17 / 2 = 8x2 + 1
–8 / 2 = 4x2 + 0
–4 / 2 = 2x2 + 0
–2 / 2 = 1x2 + 0
–1 / 2 = 0x2 + 1

• Therefore, 35 has the binary representation: 100011

CNG111 - Lecture 2 5
 Binary Representation of Numeric Information
(continued)
• Representing integers
– Decimal integers are converted to binary integers

– Question: given k bits, what is the value of the largest

integer that can be represented?

– Answer: 2k - 1
• Ex: given 4 bits, the largest binary number is 1111 (in
decimal 15)

• So, the largest is 24-1 = 15

CNG111 - Lecture 2 6
 Binary Representation of Numeric Information (continued)

• Signed integers must also represent the sign (positive or

negative) – How ?

• Method 1: Sign/magnitude notation

1 101 = - 5
0 101 = + 5
• Problems:
– Two different representations for 0:
• 1 000 = -0
• 0 000 = +0
– Addition & subtraction require a watch for the sign!
Otherwise, you get wrong results:
• 0 010 (+2) + 1 010 (-2) = 1 100 (-4)
CNG111 - Lecture 2 7
 Binary Representation of Numeric
Information (continued)
• Method 2: Two’s complement instead of sign-magnitude
representation
– Positive numbers have a leading 0.
• 5 => 0101
– The representation for negative numbers is found by
subtracting the absolute value from 2N for an N-bit system(2N
–(absolute value of the number):
• Example: Assume that N=4. Then, -5 => 24 – 5 = 16 – 5 =
1110 => 10112
• Advantages:
– 0 has a single representation: +0 = 0000, -0 = 0000
– Arithmetic works fine without checking the sign bit:
• 1011 (-5) + 0110 (6) = 0001 (1)
• 1011 (-5) + 0011 (3) = 1110 (-2)
CNG111 - Lecture 2 8
 Binary Decimal Value in
Binary Representation of Number Value Two’s
Compleme
Numeric Information (continued) nt
0000 0 0
• Shortcut to convert from “two’s 0001 1 1
complement” : 0010 2 2
– If the leading bit is zero, no need to 0011 3 3
convert. 0100 4 4
– If the leading bit is one, invert the 0101 5 5

number and add 1. 0110 6 6

0111 7 7
• What is our range? 1000 8 -8
– With 2’s complement we can 1001 9 -7
represent numbers from - 1010 10 -6
2N-1 to 2N-1 - 1 using N bits. 1011 11 -5
– 8 bits: -128 to +127. 1100 12 -4

– 16 bits: -32,768 to +32,767. 1101 13 -3

1110 14 -2
– 32 bits: -2,147,483,648 to
1111 15 -1
+2,147,483,647. CNG111 - Lecture 2 9
 Binary Representation of Numeric
Information (continued)
• Example:
– We want to compute: 12 – 6
– 12 => 01100
– -6 => -(00110) => (11001)+1 => (11010)
• 12 – 6 =
01100
+ 11010
So, addition and subtraction operations
----------- are simpler in the Two’s Complement
00110 => 6 representation

CNG111 - Lecture 2 10
 Binary Representation of Numeric Information
(continued)

• Due to its advantages, two’s complement is the most

common way to represent integers on computers.

CNG111 - Lecture 2 11
Binary Representation of Real Numbers

Conversion of the digits after the dot into binary:

• 1st Way:
– 0.375  0x½ + 1x¼ + 1x1/8  011

•2nd Way:
– Multiply by 2 and get the integer part until we get ‘0’
after the dot:
• 0.375 x 2 = 0.750 = 0 + 0.750
• 0.750 x 2 = 1.500 = 1 + 0.500
• 0.500 x 2 = 1.000 = 1 + 0.000
CNG111 - Lecture 2 12
 Binary Representation of Real Numbers
• Approach 1: Use fixed-point
– Similar to integers, except that there is a decimal point.
– E.g: using 8 bits:

1 1 1 1 1 1 1 1 = 1*23 + 1*22 + 1*21 + 1*20 +

1*2-1 + 1*2-2 + 1*2-3 + 1*2-4 = 15.9375

Assumed decimal
point

CNG111 - Lecture 2 13
 Binary Representation of Real Numbers
• Location of the decimal point changes the value of the
number.
– E.g.: using 8 bits:

1 1 1 1 1 1 1 1 = 1*24 + 1*23 + 1*22 + 1*21 + 1*20

1*2-1 + 1*2-2 + 1*2-3 = 31.875

Assumed decimal
point

CNG111 - Lecture 2 14
 Binary Representation of Real Numbers
• Solution: Use scientific notation: a x 2b (or ±M x B±E)
– Example: 5.75
• 5  101
• 0.75  ½ + ¼  2-1+2-2  11
• 5.75  101.11 x 20
• Number is then normalized so that the first significant
digit is immediately to the left of the binary point
– Example: 1.0111 x 22
• We take and store the mantissa and the exponent.
CNG111 - Lecture 2 15
 Binary Representation of Real Numbers
• This needs some standardization for:
– where to put the decimal point
– how to represent negative numbers
– how to represent numbers less than 1

CNG111 - Lecture 2 16
 IEEE 32bit Floating-Point Number
Representation (8bit exp, 23 bit fraction)

• M x 2E
• Exponent (E): 8 bits
• Add 127 to the exponent value before storing it
• E can be 0 to 255 with 127 representing the real
zero.
• Fraction (M - Mantissa): 23 bits
CNG111 - Lecture 2 17
 IEEE 32bit Floating-Point Number
Representation (1 bit sign, 8 bits exponent, 23 bits fraction)

• Example: 12.375
• The digits before the dot:
• (12)10  (1100)2
• The digits after the dot:
• 1st Way: 0.375  0x½ + 1x¼ + 1x1/8  011
• 2nd Way: Multiply by 2 and get the integer part until 0:
• 0.375 x 2 = 0.750 = 0 + 0.750
• 0.750 x 2 = 1.50 = 1 + 0.50
• 0.50 x 2 = 1.0 = 1 + 0.0
• (12.375)10 = (1100.011)2
• NORMALIZE: (1100.011)2= (1.100011)2x23
• Exponent: 3, adding 127 to it, we get 1000 0010
• Fraction: 100011
• Then our number is: 0 10000010 10001100000000000000000
2013 G. Ucoluk & S. Kalkan - CEng 111 18
 IEEE 32bit Floating-Point Number
Representation
• What is the maximum positive IEEE floating
point value that can be stored?
– Just less than 2129.

• Check out these useful links:

– http://steve.hollasch.net/cgindex/coding/ieeefloat.ht
ml
– http://babbage.cs.qc.edu/IEEE-754/Decimal.html
– http://www.cs.uaf.edu/~cs301/notes/Chapter4/node
13.html
CNG111 - Lecture 2 19
 Binary Representation of Textual Information

• Characters are mapped onto binary numbers

– ASCII (American Standard Code for Information Interchange) code set
• Originally: 7 bits per character; 128 character codes
– Unicode code set
• 16 bits per character
– UTF-8 (UCS Transformation Format) code set.
– UTF-8 is capable of encoding all 1,112,064 valid character code
points in unicode using one to four one-byte (8-bit) code units.
• Variable number of 8-bits.

CNG111 - Lecture 2 20
 Binary Representation of Textual Information (cont’d)
Decima Binary Val. Dec. Unicode Charac.
l 0x30 0x0030 0
48 0011000 0 0x31 0x0031 1
0
0x32 0x0032 2
49 0011000 1
1 0x33 0x0033 3

50 0011001 2 0x34 0x0034 4

0 0x35 0x0035 5
51 0011001 3 0x36 0x0036 6
1
ASCII 52 0011010 4
0x37 0x0037 7 Unicode
7 bits long 0 0x38 0x0038 8 16 bits long
53 0011010 5 0x39 0x0039 9
1 0x3A 0x003A :
54 0011011 6 0x3B 0x003B ;
0
0x3C 0x003C <
55 0011011 7
1 0x3D 0x003D =
56 0011100 8 0x3E 0x003E >
0 Partial
0x3F 0x003F ?
57 0011100 9 0x40 0x0040 @
listings
1
0x41 0x0041 A
only!
58 0011101 :
0 0x42 0x0042 B
59 0011101 ;CNG111 - Lecture 2 21
1
 Binary Representation of Sound and Images

• Multimedia data is sampled to store a digital

form, with or without detectable differences

• Representing sound data

– Sound data must be digitized for storage in a computer

– Digitizing means periodic sampling of amplitude values

CNG111 - Lecture 2 22
Binary Representation of Sound and Images
(continued)

– From samples, original sound may be approximated

– To improve the approximation:

• Sample more frequently (increase sampling rate)

• Use more bits for each sample value ( bit depth)

CNG111 - Lecture 2 23
 Digitization of an Analog Signal

(a) Sampling the Original

Signal

(b) Recreating the

Signal from the Sampled
Values

CNG111 - Lecture 2 24
Binary Representation of Sound and Images
(continued)
• Representing image data

– Images are sampled by reading

color and intensity values at
even intervals across the image

– Each sampled point is a pixel

– Image quality depends on

number of bits at each pixel

– More image information: http://c

at.xula.edu/tutorials/imaging/gra
yscale.php
CNG111 - Lecture 2 25
http://photo.net/equipment/digital/basics/