Method and arrangement for transcoding a video bitstream
An example method includes receiving a video bitstream with a high fidelity input format and side information related to the video bitstream. The side information includes a video bitstream representing the original video source with a low fidelity input format, coding parameters optimized based on knowledge of the original video source, transform coefficients, indicative of a residual between the original video source and a decoded version of the high fidelity format. The method further includes decoding the side information to generate transcoding guiding information, estimating a representation of video bitstream with a low fidelity output format, based on the received bit stream and the generated transcoding guiding information, improving the estimated low-fidelity representation by adding the transform coefficients of the side information to transform coefficients generated from the estimated low-fidelity representation, and encoding the improved estimated representation of the low-fidelity bitstream.
1. A method of operation by a transcoder arrangement, the method comprising:
receiving a base layer of a scalable video bitstream, the base layer representing an original video source in a high fidelity format;
receiving side information for an additional layer of the scalable video bitstream, the additional layer being encoded in dependency on the base layer and representing the original video source in a low fidelity format, and the side information including pixel prediction parameters and transform coefficients for the additional layer;
decoding the base layer to obtain reconstructed pixel data;
generating transform coefficients based on prediction errors between the reconstructed pixel data and predicted pixel data, as predicted from the reconstructed pixel data using motion estimation parameters associated with the base layer and the pixel prediction parameters included in the side information;
refining the generated transform coefficients by adding the transform coefficients included in the side information for the additional layer to the generated transform coefficients, to obtain refined transform coefficients; and
generating an output video bitstream representing the original video source in the low fidelity format, based on the refined transform coefficients and the reconstructed pixel data.
2. The method of claim 1 , wherein receiving the side information comprises receiving the side information together with receiving the base layer of the scalable video bitstream.
3. The method of claim 1 , wherein receiving the side information comprises receiving the additional layer of the scalable video bitstream, in conjunction with receiving the base layer of the scalable video bitstream.
4. The method of claim 1 , wherein decoding the base layer to obtain the reconstructed pixel data includes down sampling pixel data decoded from the base layer to match a resolution of the additional layer.
5. The method of claim 1 , wherein the generated transform coefficients and the transform coefficients included in the side information are in quantized form, for adding them to obtain the refined transform coefficients.
6. The method of claim 1 , further comprising applying a clipping operation to the refined transform coefficients, to ensure that the refined transform coefficients do not exceed limits associated with inverse quantization and transformation of the refined transform coefficients, for use in generating the output video bitstream.
7. A transcoder arrangement comprising:
input/output circuitry; and
processing circuitry configured to:
receive, via the input/output circuitry, a base layer of a scalable video bitstream, the base layer representing an original video source in a high fidelity format, and side information for an additional layer of the scalable video bitstream, the additional layer being encoded in dependency on the base layer and representing the original video source in a low fidelity format, and the side information including pixel prediction parameters and transform coefficients for the additional layer;
decode the base layer to obtain reconstructed pixel data;
generate transform coefficients based on prediction errors between the reconstructed pixel data and predicted pixel data, as predicted from the reconstructed pixel data using motion estimation parameters associated with the base layer and the pixel prediction parameters included in the side information;
refine the generated transform coefficients by adding the transform coefficients included in the side information for the additional layer to the generated transform coefficients, to obtain refined transform coefficients; and
generate an output video bitstream, based on the refined transform coefficients and the reconstructed pixel data, the output video bitstream representing the original video source in the low fidelity format, for output from the transcoding arrangement via the input/output circuitry.
8. The method of claim 7 , wherein receiving the side information comprises receiving the side information together with receiving the base layer of the scalable video bitstream.
9. The method of claim 7 , wherein receiving the side information comprises receiving the additional layer of the scalable video bitstream, in conjunction with receiving the base layer of the scalable video bitstream.
10. The method of claim 7 , wherein decoding the base layer to obtain the reconstructed pixel data includes down sampling pixel data decoded from the base layer to match a resolution of the additional layer.
11. The method of claim 7 , wherein the generated transform coefficients and the transform coefficients included in the side information are in quantized form, for adding them to obtain the refined transform coefficients.
12. The method of claim 7 , further comprising applying a clipping operation to the refined transform coefficients, to ensure that the refined transform coefficients do not exceed limits associated with inverse quantization and transformation of the refined transform coefficients, for use in generating the output video bitstream.