Bit-rate reduction for multimedia data streams

• US 20070058716A1
• Filed: 09/08/2006
• Published: 03/15/2007
• Est. Priority Date: 09/09/2005
• Status: Active Grant

• Current Assignee: Broadcast International, Inc. (Creative Realities, Inc.)
• Original Assignee: Broadcast International, Inc. (Creative Realities, Inc.)
• Inventors: Blum, David
• Granted Patent: US 8,160,160 B2
• Time in Patent Office: Days
• US Class Current: 375/240.30
• CPC Class Codes: H04N 19/115 Selection of the code volume for a coding unit prior to coding
• CPC Class Codes: H04N 19/124 Quantisation
• CPC Class Codes: H04N 19/146 Data rate or code amount at the encoder output
• CPC Class Codes: H04N 19/15 by monitoring actual compressed data size at the memory before deciding storage at the transmission buffer
• CPC Class Codes: H04N 19/154 Measured or subjectively estimated visual quality after decoding, e.g. measurement of distortion use of rate-distortion criteria H04N19/147
• CPC Class Codes: H04N 19/194 involving only two passes
• CPC Class Codes: H04N 19/523 with sub-pixel accuracy
• CPC Class Codes: H04N 19/533 Motion estimation using multistep search, e.g. 2D-log search or one-at-a-time search [OTS]
• CPC Class Codes: H04N 19/583 Motion compensation with overlapping blocks
• CPC Class Codes: H04N 19/61 in combination with predictive coding
• CPC Class Codes: H04N 19/63 using sub-band based transform, e.g. wavelets
• CPC Class Codes: H04N 19/80 Details of filtering operations specially adapted for video compression, e.g. for pixel interpolation H04N19/635, H04N19/86 take precedence
• CPC Class Codes: H04N 19/85 using pre-processing or post-processing specially adapted for video compression

Abstract

Systems and methods for reducing a bit-rate in a multimedia data stream while maintaining high image quality. Two consecutive frames of the multimedia data stream are spatially filtered, temporally filtered, and combined into a pre-processed frame. In one embodiment, the two consecutive frames are consecutive fields of an interlaced data stream that are deinterlaced by the spatial and temporal filtering process. The pre-processed frame is encoded multiple times using different quantization values to produce a plurality of encoded frames. One of the encoded frames is selected based on its image quality and/or size for inclusion in an encoded data stream. In one embodiment, the pre-processed frame undergoes a motion estimation method that performs an integer level search of overlapping locations around a co-located macroblock in a reference frame followed by a sub-pixel level diamond pattern search.

Claims

28. A method for deinterlacing a multimedia data stream, the method comprising:

creating a three-dimensional array comprising a first field and a second field corresponding to an interlaced frame of a multimedia data stream;

filtering and decimating the three-dimensional array along each of the dimensions to create a plurality of sub-fields;

reducing temporal noise in a first sub-field of the plurality of sub-fields; and

combining the plurality of sub-fields to create a deinterlaced frame.

29. The method of claim 28, further comprising spatially filtering the first field and the second field.

30. The method of claim 28, wherein reducing temporal noise comprises Wiener filtering the first sub-field.

31. The method of claim 28, wherein reducing temporal noise further comprises:

transforming a second sub-field of the plurality of sub-fields to the frequency domain; and

deleting frequency samples from the second sub-field that are less than a frequency threshold.

33. The method of claim 28, further comprising deleting a second sub-field of the plurality of sub-fields.

34. The method of claim 28, further comprising:

determining an average sample value of a second sub-field; and

replacing all of the sample values in the second sub-field with the average sample value.

35. The method of claim 28, further comprising:

determining a maximum sample value for a second sub-field; and

replacing sample values in the second sub-field that are greater than a predetermined portion of the maximum sample value with a replacement value that substantially equals the predetermined portion of the maximum sample value.

36. The method of claim 35, wherein the predetermined portion of the maximum sample value comprises approximately one-half the maximum sample value.

32-1. The method of claim 32, further comprising:

performing an inverse discrete cosine transform on the second sub-field; and

Wiener filtering the second sub-field.

37. A method for low computational motion estimation of a multimedia data stream, the method comprising:

comparing a macroblock of a temporally filtered frame to a co-located macroblock of a reference frame using an error measurement; and

generating a vector related to a displacement between the macroblock of the temporally filtered frame and the co-located macroblock of the reference frame;

wherein the comparison comprises;

performing an integer pixel level search at overlapping locations around the co-located macroblock in the reference frame to find a first minimum error measurement; and

performing a first sub-pixel level search around a first displaced location corresponding to the first minimum error measurement.

38. The method of claim 37, further comprising:

determining a second minimum error measurement from the first sub-pixel level search; and

performing a second sub-pixel level search around a second displaced location corresponding to the second minimum error measurement.

39. The method of claim 38, further comprising:

determining a third minimum error measurement from the second sub-pixel level search; and

if the third minimum error measurement is equal to the second minimum error measurement, associating the vector with the second displaced location.

40. The method of claim 39, wherein if the third minimum error measurement is less than the second minimum error measurement, performing a third sub-pixel level search around a third displaced location corresponding to the third minimum error measurement.

41. The method of claim 37, wherein the error measurement comprises a sum of absolute differences.

42. The method of claim 37, wherein the first sub-pixel search comprises searching quarter and half pixels in a diamond pattern around the first displaced location.

43. The method of claim 37, wherein the temporally filtered frame comprises two consecutive frames of the multimedia data stream that have been spatially and temporally combined.