Databases Reference
In-Depth Information
x
n
^
^
x
n
d
n
d
n
+
d
n
+
+
Q
+
-
+
-
+
p
n
p
n
+
p
n
+
P
P
P
Decoder
P
+
Encoder
+
^
x
n
P
p
F I GU R E 11 . 16
The DPCM structure with a pitch predictor.
3
2
1
0
−1
−
2
−3
500
1000
1500
2000
2500
3000
3500
4000
F I GU R E 11 . 17
The residual sequence using the DPCM system with a pitch
predictor.
error. The level of distortion we perceive is often related to the level of the speech signal. In
regions where the speech signal is of higher amplitude, we have a harder time perceiving the
distortion, but the same amount of distortion in a different frequency band, where the speech
is of lower amplitude, might be very perceptible. We can take advantage of this by shaping
the quantization error so that most of the error lies in the region where the signal has a higher
amplitude. This variation of DPCM is called
noise feedback coding
(NFC) (see [
134
]for
details).
11.7.1 G.726
The International Telecommunications Union has published several recommendations for a
standard ADPCM system, including recommendations G.721, G.723, and G.726. G.726 su-
persedes G.721 and G.723. In this section we will describe the G.726 recommendation for
ADPCM systems at rates of 40, 32, 24, and 16 kbits.
