4X4 transform for media coding
In general, techniques are described that provide for 4×4 transforms for media coding. A number of different 4×4 transforms are described that adhere to these techniques. As one example, an apparatus includes a 4×4 discrete cosine transform (DCT) hardware unit. The DCT hardware unit implements an orthogonal 4×4 DCT having an odd portion that applies first and second internal factors (C, S) that are related to a scaled factor (ξ) such that the scaled factor equals a square root of a sum of a square of the first internal factor (C) plus a square of the second internal factor (S). The 4×4 DCT hardware unit applies the 4×4 DCT implementation to media data to transform the media data from a spatial domain to a frequency domain. As another example, an apparatus implements a non-orthogonal 4×4 DCT to improve coding gain.
1. An apparatus for encoding media data comprising:
a 4×4 discrete cosine transform (DCT) hardware unit that implements an orthogonal 4×4 DCT having an odd portion that applies first and second internal factors (C, S) that are related to a scaled factor (ξ) such that the scaled factor equals a square root of a sum of a square of the first internal factor (C) plus a square of the second internal factor (S), the first and second internal factors (C, S) are co-prime and greater than or equal to two,
wherein the 4×4 DCT hardware unit applies the 4×4 DCT implementation to media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients, and
wherein the 4×4 DCT hardware unit includes a quantization unit that multiplies the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.
2. The apparatus of claim 1 ,
wherein the internal factor C equals two, and
wherein the internal factor S equals five.
3. The apparatus of claim 1 ,
wherein the internal factor C equals three, and
wherein the internal factor S equals seven.
4. The apparatus of claim 1 ,
wherein the internal factor C equals five, and
wherein the internal factor S equals twelve.
5. The apparatus of claim 1 ,
wherein the internal factor C equals seventeen, and
wherein the internal factor S equals forty one.
6. The apparatus of claim 1 , wherein the 4×4 DCT hardware unit comprises a processor that executes software to implement the orthogonal 4×4 DCT.
7. A method for encoding media data comprising:
applying an orthogonal 4×4 discrete cosine transform (DCT) implementation with a 4×4 DCT hardware unit to media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients,
wherein the orthogonal 4×4 DCT implementation includes an odd portion that applies first and second internal factors (C, S) that are related to a scaled factor (ξ) such that the scaled factor equals a square root of a sum of a square of the first internal factor (C) plus a square of the second internal factor (S), the first and second internal factors (C, S) are co-prime and greater than or equal to two, and
wherein the 4×4 DCT hardware unit includes a quantization unit that multiplies the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.
8. The method of claim 7 ,
wherein the internal factor C equals two, and
wherein the internal factor S equals five.
9. The method of claim 7 ,
wherein the internal factor C equals three, and
wherein the internal factor S equals seven.
10. The method of claim 7 ,
wherein the internal factor C equals five, and
wherein the internal factor S equals twelve.
11. The method of claim 7 ,
wherein the internal factor C equals seventeen, and
wherein the internal factor S equals forty one.
12. An apparatus for encoding media data comprising:
means for applying an orthogonal 4×4 discrete cosine transform (DCT) implementation to media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients,
wherein the orthogonal 4×4 DCT implementation includes an odd portion that applies first and second internal factors (C, S) that are related to a scaled factor (ξ) such that the scaled factor equals a square root of a sum of a square of the first internal factor (C) plus a square of the second internal factor (S), the first and second internal factors (C, S) are co-prime and greater than or equal to two, and
wherein the means for applying an orthogonal 4×4 DCT includes a means for quantization that multiplies the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.
13. The apparatus of claim 12 ,
wherein the internal factor C equals two, and
wherein the internal factor S equals five.
14. The apparatus of claim 12 ,
wherein the internal factor C equals three, and
wherein the internal factor S equals seven.
15. The apparatus of claim 12 ,
wherein the internal factor C equals five, and
wherein the internal factor S equals twelve.
16. The apparatus of claim 12 ,
wherein the internal factor C equals seventeen, and
wherein the internal factor S equals forty one.
17. A non-transitory computer-readable storage medium that stores instructions for encoding media data that, when executed by a processor, cause the processor to:
apply an orthogonal 4×4 discrete cosine transform (DCT) implementation with a 4×4 DCT hardware unit to media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients,
wherein the orthogonal 4×4 DCT implementation includes an odd portion that applies first and second internal factors (C, S) that are related to a scaled factor (ξ) such that the scaled factor equals a square root of a sum of a square of the first internal factor (C) plus a square of the second internal factor (S), the first and second internal factors (C, S) are co-prime and greater than or equal to two, and
wherein the 4×4 DCT hardware unit includes a quantization unit that multiplies the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.
18. An apparatus for encoding media data comprising:
a 4×4 discrete cosine transform (DCT) hardware unit, wherein the DCT hardware unit implements a non-orthogonal 4×4 DCT having an odd portion that applies first and second internal factors (C, S) that are related to a scaled factor (ξ) by the following equation:
ξ
=
C
+
S
ω
+
ψ
,
wherein variables ω and ψ denote irrational internal transform factors and variables C and S denote dyadic rational internal transform factors used in place of variables ω and ψ in integer implementations of the non-orthogonal 4×4 DCT, the first and second internal factors (C, S) are co-prime and greater than or equal to two,
wherein the 4×4 DCT hardware unit applies the 4×4 DCT implementation to media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients, and
wherein the 4×4 DCT hardware unit includes a quantization unit that multiplies the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.
19. The apparatus of claim 18 ,
wherein the internal factor C equals two, and
wherein the internal factor S equals five.
20. The apparatus of claim 18 ,
wherein the internal factor C equals five, and
wherein the internal factor S equals twelve.
21. The apparatus of claim 18 , wherein the 4×4 DCT hardware unit comprises a processor that executes software to implement the non-orthogonal 4×4 DCT.
22. A method for encoding media data comprising:
applying an orthogonal 4×4 discrete cosine transform (DCT) with a 4×4 DCT hardware unit to DCT coefficients representative of media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients,
wherein the 4×4 DCT comprises an DCT of a non-orthogonal 4×4 DCT having an odd portion that applies first and second internal factors (C, S) that are related to a scaled factor (ξ) by the following equation:
ξ
=
C
+
S
ω
+
ψ
,
wherein variables ω and ψ denote irrational internal transform factors and variables C and S denote dyadic rational internal transform factors used in place of variables ω and ψ in integer implementations of the non-orthogonal 4×4 DCT, the first and second internal factors (C, S) are co-prime and greater than or equal to two, and
wherein the 4×4 DCT hardware unit includes a quantization unit that multiplies the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.
23. The method of claim 22 ,
wherein the internal factor C equals two, and
wherein the internal factor S equals five.
24. The method of claim 22 ,
wherein the internal factor C equals five, and
wherein the internal factor S equals twelve.
25. An apparatus for encoding media data comprising:
means for applying an orthogonal 4×4 discrete cosine transform (DCT) with a 4×4 DCT hardware unit to DCT coefficients representative of media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients,
wherein the 4×4 DCT comprises a DCT of a non-orthogonal 4×4 DCT having an odd portion that applies first and second internal factors (C, S) that are related to a scaled factor (ξ) by the following equation:
ξ
=
C
+
S
ω
+
ψ
,
wherein variables ω and ψ denote irrational internal transform factors and variables C and S denote dyadic rational internal transform factors used in place of variables ω and ψ in integer implementations of the non-orthogonal 4×4 DCT, the first and second internal factors (C, S) are co-prime and greater than or equal to two, and
wherein the means for applying a 4×4 DCT includes a means for quantization that multiplies of the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.
26. The apparatus of claim 25 ,
wherein the internal factor C equals two, and
wherein the internal factor S equals five.
27. The apparatus of claim 25 ,
wherein the internal factor C equals five, and
wherein the internal factor S equals twelve.
28. A non-transitory computer-readable storage medium that stores instructions for encoding media data that, when executed by a processor, cause the processor to:
apply a non-orthogonal 4×4 discrete cosine transform (DCT) with a 4×4 DCT hardware unit to media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients,
wherein the non-orthogonal 4×4 DCT includes an odd portion that applies first and second variables (C, S) that are related to a scaled factor (ξ) by the following equation:
ξ
=
C
+
S
ω
+
ψ
,
wherein variables ω and ψ denote irrational internal transform factors and variables C and S denote dyadic rational internal transform factors used in place of variables ω and ψ in integer implementations of the non-orthogonal 4×4 DCT, the first and second internal factors (C, S) are co-prime and greater than or equal to two, and
wherein the 4×4 DCT hardware unit includes a quantization unit that multiplies the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.
29. An apparatus for encoding media data comprising,
a 4×4 discrete cosine transform (DCT) hardware unit that implements a non-orthogonal 4×4 DCT having an odd portion that applies first and second internal factors (A, B) as a part of the 4×4 inverse DCT that are related to a scaled factor (ξ) by the following equation:
ξ
=
A
+
B
1
+
1
/
2
,
wherein the scaled factor (ξ) equals a sum of the first internal factor (A) plus the second internal factor (B) divided by one plus one divided by the square root of two, the first and second internal factors (A, B) are co-prime and greater than or equal to two,
wherein the 4×4 DCT hardware unit applies the 4×4 DCT implementation to media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients, and
wherein the 4×4 DCT hardware unit includes a quantization unit that multiplies the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.
30. The apparatus of claim 29 ,
wherein the internal factor A equals three, and
wherein the internal factor B equals two.
31. The apparatus of claim 29 ,
wherein the internal factor A equals seven, and
wherein the internal factor B equals five.
32. The apparatus of claim 29 ,
wherein the internal factor A equals twenty nine, and
wherein the internal factor B equals forty one.
33. The apparatus of claim 29 , wherein the 4×4 DCT hardware unit comprises a processor that executes software to implement the non-orthogonal 4×4 DCT.
34. A method for encoding media data comprising:
applying a non-orthogonal 4×4 discrete cosine transform (DCT) with a 4×4 DCT hardware unit to media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients,
wherein the non-orthogonal 4×4 DCT includes an odd portion that applies first and second internal factors (A, B) that are related to a scaled factor (ξ) by the following equation:
ξ
=
A
+
B
1
+
1
/
2
,
wherein the scaled factor (ξ) equals a sum of the first internal factor (A) plus the second internal factor (B) divided by one plus one divided by the square root of two, the first and second internal factors (A, B) are co-prime and greater than or equal to two, and
wherein the 4×4 DCT hardware unit includes a quantization unit that multiplies the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.
35. The method of claim 34 ,
wherein the internal factor A equals three, and
wherein the internal factor B equals two.
36. The method of claim 34 ,
wherein the internal factor A equals seven, and
wherein the internal factor B equals five.
37. The method of claim 34 ,
wherein the internal factor A equals twenty nine, and
wherein the internal factor B equals forty one.
38. An apparatus for encoding media data comprising:
means for applying a non-orthogonal 4×4 discrete cosine transform (DCT) with a 4×4 DCT hardware unit to media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients,
wherein the non-orthogonal 4×4 DCT includes an odd portion that applies first and second internal factors (A, B) that are related to a scaled factor (ξ) by the following equation:
ξ
=
A
+
B
1
+
1
/
2
,
wherein the scaled factor (ξ) equals a sum of the first internal factor (A) plus the second internal factor (B) divided by one plus one divided by the square root of two, the first and second internal factors (A, B) are co-prime and greater than or equal to two, and
wherein the means for applying a non-orthogonal 4×4 DCT includes a means for quantization that multiplies the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.
39. The apparatus of claim 38 ,
wherein the internal factor A equals three, and
wherein the internal factor B equals two.
40. The apparatus of claim 38 ,
wherein the internal factor A equals seven, and
wherein the internal factor B equals five.
41. The apparatus of claim 38 ,
wherein the internal factor A equals twenty nine, and
wherein the internal factor B equals forty one.
42. A non-transitory computer-readable storage medium that stores instructions for encoding media data that, when executed by a processor, cause the processor to:
apply a non-orthogonal 4×4 discrete cosine transform (DCT) with a 4×4 DCT hardware unit to media data to transform the media data from a spatial domain to a frequency domain and to determine one or more DCT coefficients,
wherein the non-orthogonal 4×4 DCT includes an odd portion that applies first and second internal factors (A, B) as a part of the non-orthogonal 4×4 DCT that are related to a scaled factor (ξ) by the following equation:
ξ
=
A
+
B
1
+
1
/
2
,
wherein the scaled factor (ξ) equals a sum of the first internal factor (A) plus the second internal factor (B) divided by one plus one divided by the square root of two, the first and second internal factors (A, B) are co-prime and greater than or equal to two, and
wherein the 4×4 DCT hardware unit includes a quantization unit that multiplies the scaled factor (ξ) with one or more DCT coefficients to determine quantized DCT coefficients.