Method and arrangement for estimating the quality degradation of a processed signal
An objective quality assessment method for obtaining an improved estimate of a perceptual quality degradation of a processed signal, and an arrangement for executing such a method, is provided, which is executed on a processed signal and an associate reference signal. Both signals are split up into associated frame-pairs after which either all or selected frame-pairs are processed further, by creating a reference residual signal and a processed residual signal for each frame-pair, calculating separate ratios of p-norms on both residual signals, and by calculating and storing a per-frame quality estimate on the basis of the ratios of p-norms for each selected frame-pair. An objective per-signal quality estimate that is proportional to the perceptual quality degradation is then provided by aggregating the calculated per-frame-pair quality estimates.
1. An objective quality assessment method for estimating a perceptual quality degradation of a processed signal, the method comprising the following steps to be executed on the processed signal and a reference signal:
a) splitting the reference signal and the processed signal into associated frame-pairs;
b) selecting a first frame-pair;
c) creating a reference residual signal and a processed residual signal for the selected frame-pair;
d) calculating separate ratios of p-norms on both residual signals for the selected frame-pair;
e) calculating and storing a per-frame quality estimate based on the ratios of p-norms for the selected frame-pair;
f) iteratively selecting additional frame-pairs, and repeating steps c) to e) for each additional frame-pair; and
g) aggregating the calculated per-frame-pair quality estimates to provide an objective per-signal quality estimate that is proportional to the perceptual quality degradation of the processed signal.
2. The quality assessment method of claim 1 , wherein the processed signal has been processed by a bandwidth extension scheme or noise-fill scheme.
3. The quality assessment method of claim 1 , further comprising:
h) repeatedly providing and storing objective per-signal quality estimates; and
i) iteratively adjusting at least one parameter of a network node that is used for distribution of the processed signal on the basis of at least one of the objective per-signal quality estimates.
4. The quality assessment method of claim 1 , wherein step f) comprises selecting a subsequent frame-pair.
5. The quality assessment method of claim 1 , wherein the step of iteratively selecting additional frame-pairs comprises selecting subsequent frame-pairs for which the energy of the respective reference signal frame exceeds a predefined threshold.
6. The quality assessment method of claim 1 , wherein the step of iteratively selecting additional frame-pairs comprises selecting subsequent frame-pairs for which the difference in energy between the reference signal having maximum energy and the energy of the reference signal frame of the respective frame-pair is below a predefined threshold.
7. The quality assessment method of claim 1 , wherein the step of calculating separate ratios of p-norms comprises calculating a ratio of p-norms, L r (n), for the reference signal, and a ratio of p-norms, L p (n), for the processed signal for frame-pair n, wherein:
L
r
(
n
)
=
{
1
K
∑
k
=
1
K
e
r
(
k
)
S
}
1
S
{
1
K
∑
k
=
1
K
e
r
(
k
)
Q
}
1
Q
and
L
P
(
n
)
=
{
1
K
∑
k
=
1
K
e
p
(
k
)
S
}
1
S
{
1
K
∑
k
=
1
K
e
p
(
k
)
Q
}
1
Q
where e r (k) is the residual reference signal for sample k, e p (k) is the processed residual signal for sample k, K is the total number of samples of frame-pair n, and S and Q are optimization parameters with S being less than Q.
8. The quality assessment method of claim 7 , wherein the per-frame-pair quality estimate, D(n), for frame n is defined as:
D
(
n
)
=
L
r
(
n
)
-
L
p
(
n
)
L
r
(
n
)
+
L
p
(
n
)
.
9. The quality assessment method of claim 1 , wherein the per-signal quality estimate, D res , is defined as:
D
res
=
1
N
∑
n
=
1
N
D
(
n
)
2
where N is the total number of selected frame-pairs.
10. A network node for providing an estimate of a perceptual quality degradation of a processed signal, by further processing the processed signal and an associated reference signal, the network node comprising:
a receiver configured to receive the processed signal from a communications network and the reference signal from a signal source;
an estimating unit connected to the receiver and configured to:
split the reference signal and the processed signal into associated frame-pairs; iteratively select frame-pairs for successive further processing; and for each selected frame-pair, to:
create a reference residual signal and a processed residual signal; calculate separate ratios of p-norms on both residual signals for the selected frame-pair; and
calculate a per-frame quality estimate on the basis of the ratios of p-norms for the selected frame-pair;
a tangible storage unit connected to the estimating unit and configured to store the calculated per-flame quality estimates; and
an aggregation unit connected to the estimating unit and to the storage unit, and configured to provide an objective per-signal quality estimate that is proportional to the perceptual quality degradation of the processed signal, by aggregating the calculated per-flame-pair quality estimates.
11. The network node of claim 10 , wherein the estimating unit is further configured to repeatedly provide objective per-signal quality estimates to a receiving device.
12. The network node of claim 10 , wherein the estimating unit is further configured to select frame-pairs by selecting each subsequent frame-pair.
13. The network node of claim 10 , wherein the estimating unit is further configured to select frame-pairs by selecting subsequent frame-pairs for which the energy of the respective reference signal frame exceeds a predefined threshold.
14. The network node of claim 10 , wherein the estimating unit is further configured to select frame-pairs by selecting subsequent frame-pairs for which the difference in energy between the reference signal having maximum energy and the energy of the reference signal frame of the respective frame-pair is below a predefined threshold.
15. The network node of claim 10 , wherein the aggregation unit is configured to provide the objective per-signal quality estimate by combining the aggregated, calculated per-frame-pair quality estimates with at least one additional per-signal quality estimate.
16. The network node of claim 10 , wherein the estimating unit is further configured to create the residual signals by filtering the processed and reference signals with a whitening filter in the time-domain.
17. The network node of claim 10 , wherein the estimating unit is further configured to create the residual signals by normalizing the processed and reference signals in the frequency-domain.
18. A perceptual quality degradation estimation system, comprising:
a receiver configured to receive a processed signal from a communications network and a reference signal from a signal source;
an estimating unit connected to the receiver and configured to:
split the reference signal and the processed signal into associated frame-pairs;
iteratively select frame-pairs for successive further processing; and for each selected frame-pair to:
create a reference residual signal and a processed residual signal; calculate separate ratios of p-norms on both residual signals for the selected frame-pair; and
calculate a per-frame quality estimate on the basis of the ratios of p-norms for the selected frame-pair;
a tangible storage unit connected to the estimating unit and configured to store the calculated per-flame quality estimates;
an aggregation unit connected to the estimating unit and to the storage unit, and configured to provide an objective per-signal quality estimate that is proportional to the perceptual quality degradation of the processed signal by aggregating the calculated per-flame-pair quality estimates, wherein the estimating unit, storage unit and aggregation unit correspond to a network node; and
a network optimizing unit connected to the aggregation unit and configured to execute configurations, re-configurations, or both of the network node on the basis of an objective per-signal quality estimate received from the aggregation unit.
19. A perceptual quality degradation estimation system, comprising:
a receiver configured to receive a processed signal from a communications network and a reference signal from a signal source;
an estimating unit connected to the receiver and configured to:
split the reference signal and the processed signal into associated frame-pairs;
iteratively select frame-pairs for successive further processing; and
for each selected frame-pair to:
create a reference residual signal and a processed residual signal; calculate separate ratios of p-norms on both residual signals for the selected frame-pair; and
calculate a per-frame quality estimate on the basis of the ratios of p-norms for the selected frame-pair;
a tangible storage unit connected to the estimating unit and configured to store the calculated per-flame quality estimates;
an aggregation unit connected to the estimating unit and to the storage unit, and configured to provide an objective per-signal quality estimate that is proportional to the perceptual quality degradation of the processed signal by aggregating the calculated per-flame-pair quality estimates, wherein the estimating unit, storage unit and aggregation unit correspond to a network node; and
a network optimizing unit connected to the aggregation unit and configured to execute configurations, re-configurations, or both of the network node on the basis of an objective per-signal quality estimate received from the aggregation unit.