1

Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 1

DPCM - Overview
n Principle of Differential Pulse Code Modulation (DPCM)
n Characteristics of DPCM quantization errors
n Adaptive intra-interframe DPCM
n Conditional Replenishment
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 2
Principle of DPCM
quantizer
entropy
coder
predictor
+
+
-
+
e'
s'
e
input
channel
s
+
+
predictor
entropy
decoder
s'
e'
output
s
channel
s
coder
decoder
Prediction error Reconstruction Reconstruction error =
quantization error
e = s s s = e + s
s s = e e = q
2
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 3
Quantization error feedback
in the DPCM coder
n Assuming a linear predictor, the DPCM coder is equivalent
to the following structure:
n Transfer function of the prefilter:
n Transfer function of quantization error feedback:
quantizer
s e
+
- -
s (s) s (q)
q(e)
-
+
e e'
~
predictor predictor
transfer function of the predictor

E () = 1 P() [ ]S()
E () =

E () + 1 P() [ ]Q()
- abbreviation for
frequency vector,
e.g.,
x
,
y
( )
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 4
Power spectrum of the DPCM
quantization error
n Power spectral density of the quantization error q measured
for intraframe DPCM with a 16 level quantizer

0
0
3
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 5
Signal distortions due to
intraframe DPCM codi ng
n Granular noise: random
noise in flat areas of the
picture
n Edge busyness: jittery
appearance of edges
(for video)
n Slope overload: blur of
high-contrast edges,
Moire patterns in periodic
structures.
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 6
Example of intraframe DPCM coding
prediction error coding
1 bit/pixel 2 bit/pixel 3 bit/pixel
4 bit/pixel original
slope overload
edge busyness
granular noise
n Linear predictor:
n Lloyd-Max quantizers
n Fixed-length coding
1/2
0 1/4 1/4
4
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 7
Interframe coding of video signals
n Interframe coding exploits:
l similarity of temporally successive pictures
l temporal properties of human vision
n Important interframe coding methods:
l Adaptive intra-interframe coding
l Conditional replenishment
l Motion-compensated prediction
l Motion-compensated interpolation
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 8
Principle of adaptive intra-interframe DPCM
n Predictor is switched between two states:
A: Intraframe prediction for
moving or changed areas.
B: Interframe prediction (previous frame
prediction) for still areas of the picture.
S
22
S
23
S
24
S
21
S
20
S
25
FRAME N - 1
S
2
S
3
S
4
S
1
S
0
FRAME N
S
22
S
23
S
24
S
21
S
20
S
25
FRAME N - 1
S
2
S
3
S
4
S
1
S
0
FRAME N
4
0
m
s

S
intra
= a
1
S
1
+ a
2
S
2
+ a
3
S
3
+ a
4
S
4

S
inter
= S
20
frame
interval
5
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 9
Intra-interframe DPCM: feedback adaptation
Interframe
predictor
Intraframe
predictor
Variable
length decoder
Variable
length coder
Quantizer
Interframe
predictor
Intraframe
predictor
Predictor
adaptation
e s
+
-
s
^
s
^
inter
s
^
intra
s'
e'
e'
s
^
s
^
inter
s
^
intra
s'
Predictor
adaptation
Coder Decoder
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 10
Intra-interframe DPCM: feedforward adaptation
Interframe
predictor
Intraframe
predictor
Variable
length decoder
Variable
length coder
Quantizer
Interframe
predictor
Intraframe
predictor
Predictor
adaptation
e s
+
-
s
^
s
^
inter
s
^
intra
s'
e'
e'
s
^
s
^
inter
s
^
intra
s'
intra-/interframe
switching information
Coder Decoder
6
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 11
Conditional replenishment
n Still areas: repeat from frame store
n Moving areas: encode and transmit address and waveform
SIGNAL
INPUT
SEGMENTER
(MOVEMENT
DETECTOR)
CODING,
ADDRESSING,
BUFFERING
BUFFERING,
DECODING,
ADDRESSING
FRAME DELAY
(1PICTURE MEMORY)
FRAME DELAY
(1PICTURE MEMORY)
TRANSMISSION
CHANNEL
SIGNAL
OUTPUT
CODER DECODER
Change
detector
Frame store Frame store
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 12
Change detection
n Example of a pixel-wise change detector
n Example of a block-wise change detector
ABS
Average
of 3x3
window
Eliminate isolated
points or pairs of
points
+
-
Current
frame
Previous
frame
Decision
changed/
unchanged
Threshold
ABS
Accumulate
over NxN
blocks
+
-
Current
frame
Previous
frame
Decision
changed/
unchanged
Threshold
7
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 13
The Dirty Window effect
n Conditional replenishment scheme with change detection
threshold set too high leads to the subjective impression of
looking through a dirty window.
Background
Moving area
picked up by
change detector
Moving areas
missed by
change detector
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 14
Crawford noise reduction filter
+
-
+
+

f( )
NL
frame
store
clean video
signal
noisy video
signal

f( )
Non-
linearity
8
Bernd Gi rod: EE368b I mage and Vi deo Compressi on DPCM no. 15
DPCM - Summary
n DPCM: Prediction from previously coded/transmitted samples (known at
transmitter and receiver)
n Typical signal distortions for intraframe DPCM: granular noise, edge
busyness, slope overload
n Adaptive Intra-Interframe-DPCM: forward adaptation vs. backward
adaptation
n Conditional replenishment: only transmit frame-to-frame changes
n Temporal noise reduction by nonlinear, recursive frame differencing