摘要:
Techniques and tools for selecting search ranges and/or motion vector ranges during motion estimation are described. For example, a video encoder performs motion estimation constrained by a first search range, which results in multiple motion vectors. The encoder computes motion vector distribution information for the motion vectors. To compute the distribution information, the encoder can track the motion vectors in a histogram and count how many of the motion vectors fall within each of multiple intervals for the distribution information. The encoder then selects a second search range and performs motion estimation constrained by the second search range. Selecting the second search range can include selecting a motion vector range, which in some cases in effect determines the second search range.
摘要:
Techniques and tools for selecting search ranges and/or motion vector ranges during motion estimation are described. For example, a video encoder performs motion estimation constrained by a first search range, which results in multiple motion vectors. The encoder computes motion vector distribution information for the motion vectors. To compute the distribution information, the encoder can track the motion vectors in a histogram and count how many of the motion vectors fall within each of multiple intervals for the distribution information. The encoder then selects a second search range and performs motion estimation constrained by the second search range. Selecting the second search range can include selecting a motion vector range, which in some cases in effect determines the second search range.
摘要:
Techniques and tools for adaptive selection of picture quantization parameters (“QPs”) for predicted pictures are described. For example, a video encoder adaptively selects a delta QP for a B-picture based on spatial complexity, temporal complexity, whether differential quantization is active, whether the B-picture is available as a reference picture, or some combination or subset of these or other factors. The delta QP can then be used to adjust the picture QP for the B-picture (e.g., to reduce bit rate for the B-picture without appreciably reducing the perceived quality of a video sequence.
摘要:
Techniques and tools are described for adaptive deadzone (“DZ”) resizing during quantization. For example, in some embodiments, for quantization of an AC frequency coefficient of a block, a video encoder adjusts DZ size of a selected quantizer depending on the texture of the block. In other embodiments, a video encoder adjusts DZ size depending on the frequency of a coefficient being quantized. In still other embodiments, for quantization of an AC frequency coefficient of a block, a video encoder adjusts DZ size depending on the texture of the block and the frequency of the coefficient being quantized.
摘要:
Multiple-pass video encoding systems and techniques are described which utilize statistics taken during a first-pass encoding to create complexity measurements for video data which is to be encoded. By analyzing these complexity measurements, preprocessing decisions, such as, for example, the determination of strength of denoise filters, can be made with greater accuracy. In one implementation, these complexity measurements take the form of calculation of temporal and spatial complexity parameters, which are then used to compute a unified complexity parameter for each group of pictures being encoded.
摘要:
Techniques and tools are described for compensating for rounding when estimating sample-domain distortion in the transform domain. For example, a video encoder estimates pixel-domain distortion in the transform domain for a block of transform coefficients after compensating for rounding in the DC coefficient of the block. In this way, the video encoder improves the accuracy of pixel-domain distortion estimation but retains the computational advantages of performing the estimation in the transform domain. Rounding compensation includes, for example, looking up an index (from a de-quantized transform coefficient) in a rounding offset table to determine a rounding offset, then adjusting the coefficient by the offset. Other techniques and tools described herein are directed to creating rounding offset tables and encoders that make encoding decisions after considering rounding effects that occur after an inverse frequency transform on de-quantized transform coefficient values.
摘要:
Techniques and tools are described for adaptive deadzone (“DZ”) resizing during quantization. For example, in some embodiments, for quantization of an AC frequency coefficient of a block, a video encoder adjusts DZ size of a selected quantizer depending on the texture of the block. In other embodiments, a video encoder adjusts DZ size depending on the frequency of a coefficient being quantized. In still other embodiments, for quantization of an AC frequency coefficient of a block, a video encoder adjusts DZ size depending on the texture of the block and the frequency of the coefficient being quantized.
摘要:
Techniques and tools for adaptive selection of picture quantization parameters (“QPs”) for predicted pictures are described. For example, a video encoder adaptively selects a delta QP for a B-picture based on spatial complexity, temporal complexity, whether differential quantization is active, whether the B-picture is available as a reference picture, or some combination or subset of these or other factors. The delta QP can then be used to adjust the picture QP for the B-picture (e.g., to reduce bit rate for the B-picture without appreciably reducing the perceived quality of a video sequence.
摘要:
Multiple-pass video encoding systems and techniques are described which utilize statistics taken during a first-pass encoding to create complexity measurements for video data which is to be encoded. By analyzing these complexity measurements, preprocessing decisions, such as, for example, the determination of strength of denoise filters, can be made with greater accuracy. In one implementation, these complexity measurements take the form of calculation of temporal and spatial complexity parameters, which are then used to compute a unified complexity parameter for each group of pictures being encoded.
摘要:
Techniques and tools for content classification and adaptive quantization are described. In an example implementation, a video encoding tool classifies blocks as textured, dark smooth or other smooth. The tool classifies a block as textured or non-textured by comparing the energy of AC coefficients for the block to a texture threshold, which can be set using a non-linear mapping of possible texture classification levels to possible texture thresholds. If a block is not textured, the tool further classifies the block as dark smooth or smooth depending on average intensity value for the block. Using the classification information and one or more control parameters to control bit allocation for dark smooth content relative to other smooth content, the tool encodes the video and outputs encoded video information. Example multi-pass approaches to setting the control parameters are also described.