Method of reflecting time/language distortion in objective speech quality assessment
Disclosed is an objective speech quality assessment technique that reflects the impact of distortions which can dominate overall speech quality assessment by modeling the impact of such distortions on subjective speech quality assessment, thereby, accounting for language effects in objective speech quality assessment.
1. A method for objectively assessing speech quality comprising the steps of:
detecting distortions in an interval of speech activity using envelope information;
modifying an objective speech quality assessment value associated with the speech activity to reflect the impact of the distortions on subjective speech quality assessment; and
prior to the step of detecting, determining the interval of speech activity using the envelope information.
2. The method of claim 1 , wherein the step of modifying includes the step of determining the objective speech quality assessment value for the speech activity.
3. The method of claim 1 , wherein the distortions being detected are impulsive noise, abrupt stop or abrupt start.
4. The method of claim 1 , wherein the step of detecting includes the step of determining a distortion type.
5. A method of claim 1 , wherein the distortion type is determined to be impulsive noise if the envelope information indicates that the speech activity can be perceived by a human listener to be noise and if the interval is of a duration long enough to be perceived by a human listener but not too long for a short burst.
6. The method of claim 4 , wherein the distortion type is determined to be impulsive noise if the envelope information indicates that the speech activity can be perceived by a human listener to be noise, if a ratio of the objective speech quality assessment value to a modulation noise reference unit indicates a human listener would perceive annoying noise, and if the interval is of a duration long enough to be perceived by a human listener but not too long for a short burst.
7. The method of claim 4 , wherein the objective speech quality assessment value associated with the speech activity is modified in accordance with the following equation to obtain a modified objective speech quality assessment value if the distortion type is impulsive noise:
v
~
s
(
m
)
=
v
s
(
m
)
1
+
exp
[
-
8.2
(
m
-
m
I
)
/
Δ
e
(
l
I
)
-
10
]
where v s (m) is the objective speech quality assessment value, {tilde over (v)} s (m) is the modified objective speech quality assessment value, “m” is a frame of the interval of speech activity, “l I ” is an impulsive noise frame, “m I ” is the frame m impacted most by impulsive noise frame “l I ”, and “e(l I )” is a frame envelope for impulsive noise frame “l I ”.
8. The method of claim 4 , wherein the distortion type is determined to be abrupt stop if the envelope information indicates that there was an sufficient negative change in frame energy from one frame to another to be considered an abrupt stop and if the interval is of a duration longer than a short burst.
9. The method of claim 4 , wherein the distortion type is determined to be abrupt stop if the envelope information indicates that a maximum frame envelope had sufficient energy prior to ending the interval, and if the interval is of a duration longer than a short burst.
10. The method of clam 4 , wherein the objective speech quality assessment value associated with the speech activity is modified in accordance with the following equation to obtain a modified objective speech quality assessment value if the distortion type is impulsive noise:
v
~
s
(
m
)
=
Δ
e
(
l
M
)
[
6
1
+
exp
[
-
2
(
m
-
m
M
-
3
]
-
6
]
where v s (m) is the objective speech quality assessment value, {tilde over (v)} s (m) is the modified objective speech quality assessment value, “m” is a frame of the interval of speech activity, “l M ” is an abrupt stop frame, “m M ” is the frame m impacted most by abrupt stop frame “l M ”, and “Δe(l M )” is a delta frame envelope for abrupt stop frame “l M ”.
11. The method of claim 4 , wherein the distortion type is determined to be abrupt start if the envelope information indicates that there was an sufficient positive change in frame energy from one frame to another to be considered an abrupt start and if the interval is of a duration longer than a short burst.
12. The method of claim 4 , wherein the distortion type is determined to be abrupt stop if the envelope information indicates that a maximum frame envelope had sufficient energy towards a beginning of the interval, and if the interval is of a duration longer than a short burst.
13. The method of claim 4 , wherein the objective speech quality assessment value associated with the speech activity is modified in accordance with the following equation to obtain a modified objective speech quality assessment value if the distortion type is impulsive noise:
v
~
s
(
m
)
=
v
s
(
m
)
1
+
exp
[
-
0.4
(
m
-
m
S
)
/
Δ
e
(
l
S
)
-
10
]
where v s (m) is the objective speech quality assessment value, {tilde over (v)} s (m) is the modified objective speech quality assessment value, “m” is a frame of the interval of speech activity, “l S ” is an abrupt start frame, “m S ” is the frame m most impacted by abrupt start frame “l S ”, and “Δe(l S )” is a delta frame envelope for abrupt start frame “l S ”.
14. An objective speech quality assessment system comprising:
means for detecting distortions in an interval of speech activity using envelope information; and
means for modifying an objective speech quality assessment value associated with the speech activity to reflect the impact of the distortions on subjective speech quality assessment, wherein
the means for detecting includes a means for determining a distortion type, and
the means for detecting includes a voice activity detector for detecting intervals of speech activity, wherein the means for determining a distortion type examines intervals of speech activities detected by the voice activity detector.
15. The objective speech quality assessment system of claim 14 , wherein the means for modifying includes a means for determining the objective speech quality assessment values without accounting for distortions for the speech activity.
16. The objective speech quality assessment system of claim 14 , wherein the distortion being detected are impulsive noise, abrupt stop or abrupt start.