Method and apparatus to check for wrongly decoded macroblocks in streaming multimedia applications
A method to detect (identify) wrongly decoded data (I type macro-blocks), which can be subsequently concealed by error concealment techniques, thereby improving the visual quality of a decoded stream, is provided. Additionally, a method is provided. In use, a set of macro-block data is selected for validation. Additionally, a set of suspicious macro blocks is selected from the set of macro-block data based on location information for each suspicious macro block.
1. A method, comprising:
decoding macro-block data for detection of wrongly decoded macro-blocks within the macro-block data, utilizing a video decoding system; and
determining a plurality of macro-blocks from the macro-block data as suspicious of being wrongly decoded but for which it is unknown whether the macro-blocks are wrongly decoded by:
traversing the macro-block data in reverse zig zag order,
upon traversal of a macro-block of the macro-block data, evaluating the traversed macro-block relative to a location of the traversed macro-block with respect to its packet boundary and a location of another macro-block determined to be suspicious of being wrongly decoded, and
identifying the traversed macro-block as suspicious, based on the evaluation;
wherein the determining the plurality of macro-blocks from the macro-block data as suspicious includes keeping track of a count of consecutive macro-blocks being identified as suspicious and terminating the determining when the count reaches three.
2. The method as claimed in claim 1 , further comprising detecting at least one of the plurality of macro-blocks determined as suspicious of being wrongly decoded as being a macro-block with false coding.
3. The method as claimed in claim 2 , further comprising concealing the detected at least one macro-block and generating at least one error concealed video frame.
4. The method as claimed in claim 2 , wherein the detecting includes:
evaluating a threshold parameter based on a quantization value of a selected macro-block from the plurality of macro-blocks determined as suspicious of being wrongly decoded;
searching neighboring valid macro-blocks in a vertical adjacent location followed by a horizontal adjacent location;
finding a continuity measure by taking into consideration at least two macro-block boundaries by taking a sum of an absolute difference in pixel data for the selected suspicious macro-block and the adjacent macro-blocks;
comparing respective pixel data on the boundaries; and
deciding to mark the selected suspicious macro-block as LOST by comparing the continuity measure with a dynamic threshold.
5. The method as claimed in claim 4 , in which the threshold parameter varies dynamically based on the quantization value of the selected suspicious macro-block.
6. The method as claimed in claim 1 , wherein the traversed macro-block is identified as suspicious in response to a result of the evaluation indicating that the traversed macro-block is at an end of a packet in a current frame.
7. The method as claimed in claim 1 , wherein the traversed macro-block is identified as suspicious in response to a result of the evaluation indicating that the traversed macro-block is adjacent to another macro-block identified as suspicious of being wrongly decoded.
8. The method as claimed in claim 1 , wherein the traversed macro-block is identified as suspicious in response to a result of the evaluation indicating that the macro-block precedes a LOST macro-block within a fixed number of macro-blocks of the macro-block data.
9. The method as claimed in claim 1 , further including transmitting a list of the plurality of macro-blocks determined as suspicious of being wrongly decoded, marking at least one of the plurality of macro-blocks in the list as LOST based on a comparison of a continuity measure with a dynamic threshold, and transmitting the list with the at least one marked LOST macro-blocks for performing error concealment on the at least one marked LOST macro-blocks in the list.
10. An apparatus, comprising:
an error selection and detection block coupled to a decoder for:
receiving decoded macro-block data for detection of wrongly decoded macro-blocks within the macro-block data; and
determining a plurality of macro-blocks from the macro-block data as suspicious of being wrongly decoded but for which it is unknown whether the macro-blocks in the selected set are wrongly decoded by:
traversing the macro-block data in reverse zig zag order,
upon traversal of a macro-block of the macro-block data, evaluating the traversed macro-block relative to a location of the traversed macro-block with respect to its packet boundary and a location of another macro-block determined to be suspicious of being wrongly decoded, and
identifying the traversed macro-block as suspicious, based on the evaluation;
wherein the determining the plurality of macro-blocks from the macro-block data as suspicious includes keeping track of a count of consecutive macro-blocks being identified as suspicious and terminating the determining when the count reaches three.
11. The apparatus as claimed in claim 10 , wherein the error selection and detection block further generates a refined set of suspicious macro-blocks.
12. The apparatus as claimed in claim 10 , wherein the error selection and detection block further evaluates threshold parameters of the plurality of macro-blocks determined as suspicious of being wrongly decoded based on a quantization value.
13. The apparatus as claimed in claim 10 , wherein the error selection and detection block further searches neighboring valid macro-block data, finds a continuity measure by taking a difference in pixel data for a selected macro-block from the plurality of macro-blocks determined as suspicious and adjacent macro-blocks, and decides that the selected suspicious macro-block is LOST by comparing the continuity measure with a dynamic threshold.