Methods and apparatus for characterizing media
View Patent ↗Methods and apparatus for characterizing media are described. An example apparatus includes a transformer to convert at least a portion of a block of audio into a frequency domain representation including a plurality of frequency components; a decision metric processor to: define a band of frequency components having real and imaginary spectral components, define a plurality of frequency bins within the band, determine respective difference functions for each of the frequency bins using a product of the real and imaginary spectral components of a corresponding group of the frequency bins, the corresponding group for a first one of the frequency bins including at least the first one of the frequency bins and a second one of the frequency bins selected based on the first frequency bin, and determine a decision metric by summing the difference functions; and a signature determiner to determine a signature based on the decision metric.
1. An apparatus, comprising:
a transformer to convert at least a portion of a block of audio into a frequency domain representation including a plurality of frequency components;
a decision metric processor to:
define a band of frequency components having real and imaginary spectral components;
define a plurality of frequency bins within the band;
determine respective functions for each of the frequency bins using a product of the real and imaginary spectral components of a corresponding group of the frequency bins, wherein the corresponding group for a first one of the frequency bins includes at least the first one of the frequency bins and a second one of the frequency bins selected based on the first frequency bin; and
determine a decision metric for the band by summing the functions;
a signature determiner to determine a signature based on a value of the decision metric; and
a processor to implement at least one of the transformer, the decision metric processor, or the signature determiner.
2. An apparatus as defined in claim 1 , wherein determining the function for the first one of the frequency bins is based on the following equation:
D
[
m
]
=
∑
m
-
w
≤
j
,
k
,
r
,
s
,
u
,
v
≤
m
+
w
[
α
jk
X
R
[
j
]
X
I
[
k
]
+
β
rs
X
R
[
r
]
X
R
[
s
]
+
γ
uv
X
I
[
u
]
X
I
[
v
]
]
,
wherein m is an index of the first one of the frequency bins, D[m] is the function for the first one of the frequency bins, α jk , β rs , and γ uv are representative of coefficients to be determined by the decision metric computer, j, k, r, s, u, and v are indexes of respective ones of the plurality of frequency bins within the band, X R is a real spectral component of a respective frequency bin, and X I is an imaginary spectral component of a respective frequency bin.
3. An apparatus as defined in claim 1 , further comprising a sample generator to capture the block of audio by digitally sampling an audio signal and storing the digital samples in a buffer.
4. An apparatus as defined in claim 3 , wherein the sample generator is to capture a second block of audio by shifting a number of old samples from the buffer and shifting a number of new samples into the buffer.
5. An apparatus to create a signature representative of media, the apparatus comprising:
a transformer to convert at least a portion of a block of audio into a frequency domain representation including a plurality of frequency components;
a decision metric processor to:
define a band of frequency components having real and imaginary spectral components;
define vectors for respective frequency components, the vector for a first one of the frequency components comprising the real and imaginary spectral components for the first one of the frequency components; and
determine a decision metric using linear combinations of dot products of the vectors for the frequency components; and
a signature determiner to determine the signature based on a value of the decision metric.
6. An apparatus as defined in claim 5 , wherein determining the decision metric is based on the following equation:
D
B
[
p
]
=
∑
p
s
≤
j
,
k
,
r
,
s
,
u
,
v
≤
p
E
[
λ
jk
X
R
[
j
]
X
I
[
k
]
+
μ
rs
X
R
[
r
]
X
R
[
s
]
+
η
uv
X
I
[
u
]
X
I
[
v
]
]
,
wherein p is an index of the band of frequency components, D B is the decision metric function for a selected one of the frequency bins, λ jk , μ rs , and η uv are representative of coefficients to be determined, j, k, r, s, u, and v are indexes of respective ones of the plurality of frequency bins within the band, X R is a real spectral component of a respective vector, and X I is a imaginary spectral component of a respective vector.
7. An apparatus as defined in claim 6 , wherein the signature is based on whether a value of the decision metric D B is positive or negative.
8. A method, comprising:
converting at least a portion of a block of audio into a frequency domain representation including a plurality of frequency components;
defining a band of frequency components having real and imaginary spectral components;
defining a plurality of frequency bins within the band;
determining, using a processor, respective functions for each of the frequency bins using a product of the real and imaginary spectral components of a corresponding group of the frequency bins, wherein the corresponding group for a first one of the frequency bins includes at least the first one of the frequency bins and a second one of the frequency bins selected based on the first frequency bin;
determining, using the processor, a decision metric by summing the functions; and
determining a bit for a signature based on a value of the decision metric.
9. A method as defined in claim 8 , wherein determining the function for the first one of the frequency bins is based on the following equation:
D
[
m
]
=
∑
m
-
w
≤
j
,
k
,
r
,
s
,
u
,
v
≤
m
+
w
[
α
jk
X
R
[
j
]
X
I
[
k
]
+
β
rs
X
R
[
r
]
X
R
[
s
]
+
γ
uv
X
I
[
u
]
X
I
[
v
]
]
,
wherein m is an index of the first one of the frequency bins, D[m] is the function for the first one of the frequency bins, α jk , β rs , and γ uv are representative of coefficients to be determined by the decision metric computer, j, k, r, s, u, and v are indexes of respective ones of the plurality of frequency bins within the band, X R is a real spectral component of a respective frequency bin, and X I is a imaginary spectral component of a respective frequency bin.
10. A method as defined in claim 8 , wherein capturing the block of audio comprises digitally sampling an audio signal and storing the digital samples in a buffer.
11. A method as defined in claim 10 , further comprising capturing a second block of audio by shifting a number of old samples from the buffer and shifting a number of new samples into the buffer.
12. A method as defined in claim 11 , further comprising generating a second bit of the signature by:
converting at least a portion of the second block of audio into a second frequency domain representation including a second plurality of frequency components;
defining a second band of frequency components having real and imaginary spectral components;
defining a second plurality of frequency bins within the second band;
determining respective second functions for each of the second frequency bins using a product of the real and imaginary spectral components of a corresponding group of the second frequency bins, wherein the corresponding group for a first one of the second frequency bins includes at least the first one of the second frequency bins and a second one of the second frequency bins selected based on the first one of the second frequency bins;
determining a second decision metric by summing the second functions; and
determining the second bit for a signature based on a value of the decision metric.
13. A tangible article of manufacture comprising machine readable instructions which, when executed, cause a machine to at least:
convert at least a portion of a block of audio into a frequency domain representation including a plurality of frequency components;
define a band of frequency components having real and imaginary spectral components;
define a plurality of frequency bins within the band;
determine respective functions for each of the frequency bins using a product of the real and imaginary spectral components of a corresponding group of the frequency bins, wherein the corresponding group for a first one of the frequency bins includes at least the first one of the frequency bins and a second one of the frequency bins selected based on the first frequency bin;
determine a decision metric by summing the functions; and
determine a bit for a signature based on a value of the decision metric.
14. An article of manufacture as defined in claim 13 , wherein determining the function for the first one of the frequency bins is based on the following equation:
D
[
m
]
=
∑
m
-
w
≤
j
,
k
,
r
,
s
,
u
,
v
≤
m
+
w
[
α
jk
X
R
[
j
]
X
I
[
k
]
+
β
rs
X
R
[
r
]
X
R
[
s
]
+
γ
uv
X
I
[
u
]
X
I
[
v
]
]
,
wherein m is an index of the first one of the frequency bins, D[m] is the function for the first one of the frequency bins, α jk , β rs , and γ uv are representative of coefficients to be determined by the decision metric computer, j, k, r, s, u, and v are indexes of respective ones of the plurality of frequency bins within the band, X R is a real spectral component of a respective frequency bin, and X I is a imaginary spectral component of a respective frequency bin.
15. An article of manufacture as defined in claim 13 , wherein capturing the block of audio comprises digitally sampling an audio signal and storing the digital samples in a buffer.
16. An article of manufacture as defined in claim 15 , further comprising capturing a second block of audio by shifting a number of old samples from the buffer and shifting a number of new samples into the buffer.
17. An article of manufacture as defined in claim 16 , further comprising generating a second bit of the signature by:
converting at least a portion of the second block of audio into a second frequency domain representation including a second plurality of frequency components;
defining a second band of frequency components having real and imaginary spectral components;
defining a second plurality of frequency bins within the second band;
determining respective second functions for each of the second frequency bins using a product of the real and imaginary spectral components of a corresponding group of the second frequency bins, wherein the corresponding group for a first one of the second frequency bins includes at least the first one of the second frequency bins and a second one of the second frequency bins selected based on the first one of the second frequency bins;
determining a second decision metric by summing the second functions; and
determining the second bit for a signature based on a value of the decision metric.