Coding Strategies and Standards - Digital Speech: Coding for Low Bit Rate Communication Systems - page 15

Table 2.5 INMARSAT speech coding standards
Bit rate Noise Delay

Speech coder (kb/s) VAD reduction (ms) Quality Year

IMBE 4.15 No No 120 Communication 1990

AMBE 3.6 No No - - -

Quality measurements based on SNR can be used to evaluate coders that
preserve the waveform similarity, usually coders operating at bit rates above
16 kb/s. Low bit-rate parametric coders do not preserve the waveform simi-
larity and SNR-based quality measures become meaningless. For parametric
coders, perception-based subjective measures are more reliable. The Mean
Opinion Score (MOS) [49] scale shown in Table 2.6 is a widely-used subjective
quality measure.
Table 2.7 compares some of the most well-known speech coding standards
in terms of their bit rate, algorithmic delay and Mean Opinion Scores and
Figure 2.2 illustrates the performance of those standards in terms of speech
quality against bit rate [50, 51].
Linear PCM at 128 kb/s offers transparent speech quality and its A-law
companded 8 bits/sample (64 kb/s) version (which provides the standard
for the best (narrowband) quality) has a MOS score higher than 4, which
is described as Toll quality. In order to find the MOS score for a given

Excellent

Linear PCM

G.726

G.711

GSM EFR

G.728

ITU 4

Good

G.729

G.723.1

GSM FR

IS54

GSM/2

JDC/2

IS96

JDC

In-M

Quality

New FS 2.4

FS1016

Fair

FS1015

Poor

2 4 8 16 32 64 128

Bit rate (kb/s)

Figure 2.2 Performance of telephone band speech coding standards (only the top
fourpointsoftheMOSscalehavebeenused)