Method, apparatus, and medium for video processing
The method adaptively controls video encoding complexity to balance latency and efficiency in live streaming, addressing high computational demands through real-time detection and parameter adjustment, enhancing transcoding performance and resource utilization.
Patent Information
- Application Number
- PCT/CN2025/110189
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2024-07-23
- Filing Date
- 2025-07-23
- Publication Date
- 2026-01-29
AI Technical Summary
Existing video coding technologies face challenges in achieving real-time transcoding with low latency and efficient resource utilization, particularly in live streaming scenarios, due to high computational complexity and the need for adaptive complexity control.
A method and apparatus for video processing that adaptively control encoding complexity by detecting real-time transcoding conditions and adjusting parameters using a complexity model and logistic regression classifier to balance encoding complexity and efficiency.
This approach enables real-time transcoding with reduced latency and improved CPU utilization, ensuring seamless live streaming experiences by dynamically adjusting encoding parameters based on complexity detection.
Smart Images

Figure CN2025110189_29012026_PF_FP_ABST
Abstract
Description
METHOD, APPARATUS, AND MEDIUM FOR VIDEO PROCESSINGFIELDS
[0001] Embodiments of the present disclosure relate generally to video processing techniques, and more particularly, to complexity-aware transcoding.BACKGROUND
[0002] In nowadays, digital video capabilities are being applied in various aspects of peoples’ lives. Multiple types of video compression technologies, such as motion picture expert group (MPEG) -2, MPEG-4, international telecommunication union -telecommunication standardization sector (ITU-T) H. 263, ITU-T H. 264 / MPEG-4 Part 10 advanced video coding (AVC) , ITU-T H. 265 high efficiency video coding (HEVC) standard, versatile video coding (VVC) standard, have been proposed for video encoding / decoding. However, coding efficiency of video coding techniques is generally expected to be further improved.SUMMARY
[0003] Embodiments of the present disclosure provide a solution for video processing.
[0004] In a first aspect, a method for video processing is proposed. The method comprises: determining, for a conversion between a current video unit of a video and a bitstream of the video, complexity information associated with the conversion; obtaining at least one parameter associated with the conversion based on the complexity information; and performing the conversion based on the at least one parameter. The method in accordance with the first aspect of the present disclosure enabling adjusting the parameter based on complexity, and thus the complexity may be adaptively controlled.
[0005] In a second aspect, an apparatus for video processing is proposed. The apparatus comprises a processor and a non-transitory memory with instructions thereon. The instructions upon execution by the processor, cause the processor to perform a method in accordance with the first aspect of the present disclosure.
[0006] In a third aspect, a non-transitory computer-readable storage medium is proposed. The non-transitory computer-readable storage medium stores instructions that cause a processor to perform a method in accordance with the first aspect of the present disclosure.
[0007] In a fourth aspect, another non-transitory computer-readable recording medium is proposed. The non-transitory computer-readable recording medium stores a bitstream of a video which is generated by a method performed by an apparatus for video processing. The method comprises: determining complexity information associated with generating of the bitstream of the video from a current video unit of the video; obtaining at least one parameter for generating the bitstream based on the complexity information; and generating the bitstream based on the at least one parameter.
[0008] In a fifth aspect, a method for storing a bitstream of a video is proposed. The method comprises: determining complexity information associated with generating of the bitstream of the video from a current video unit of the video; obtaining at least one parameter for generating the bitstream based on the complexity information; generating the bitstream based on the at least one parameter; and storing the bitstream in a non-transitory computer-readable recording medium.
[0009] This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.BRIEF DESCRIPTION OF THE DRAWINGS
[0010] Through the following detailed description with reference to the accompanying drawings, the above and other objectives, features, and advantages of example embodiments of the present disclosure will become more apparent. In the example embodiments of the present disclosure, the same reference numerals usually refer to the same components.
[0011] Fig. 1 illustrates a block diagram of an example video coding system in accordance with some embodiments of the present disclosure;
[0012] Fig. 2 illustrates a block diagram of an example video encoder in accordance with some embodiments of the present disclosure;
[0013] Fig. 3 illustrates a block diagram of an example video decoder in accordance with some embodiments of the present disclosure;
[0014] Fig. 4 illustrates a group of picture (GOP) structure of size equal to 16;
[0015] Fig. 5 illustrates a GOP structure of size equal to 4;
[0016] Fig. 6 illustrates a frame window in accordance with some embodiments of the present disclosure;
[0017] Fig. 7 illustrates a multi-stage logistic regression classifier in accordance with some embodiments of the present disclosure;
[0018] Fig. 8 illustrates a flowchart of a complexity-aware video transcoding method in accordance with some embodiments of the present disclosure;
[0019] Fig. 9 illustrates a flowchart of a method for video processing in accordance with some embodiments of the present disclosure; and
[0020] Fig. 10 illustrates a block diagram of a computing device in which various embodiments of the present disclosure can be implemented.
[0021] Throughout the drawings, the same or similar reference numerals usually refer to the same or similar elements.DETAILED DESCRIPTION
[0022] Principle of the present disclosure will now be described with reference to some embodiments. It is to be understood that these embodiments are described only for the purpose of illustration and help those skilled in the art to understand and implement the present disclosure, without suggesting any limitation as to the scope of the disclosure. The disclosure described herein can be implemented in various manners other than the ones described below.
[0023] In the following description and claims, unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skills in the art to which this disclosure belongs.
[0024] References in the present disclosure to “one embodiment, ” “an embodiment, ” “an example embodiment, ” and the like indicate that the embodiment described may include a particular feature, structure, or characteristic, but it is not necessary that every embodiment includes the particular feature, structure, or characteristic. Moreover, such phrases are not necessarily referring to the same embodiment. Further, when a particular feature, structure, or characteristic is described in connection with an example embodiment, it is submitted that it is within the knowledge of one skilled in the art to affect such feature, structure, or characteristic in connection with other embodiments whether or not explicitly described.
[0025] It shall be understood that although the terms “first” and “second” etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element could be termed a second element, and similarly, a second element could be termed a first element, without departing from the scope of example embodiments. As used herein, the term “and / or” includes any and all combinations of one or more of the listed terms.
[0026] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments. As used herein, the singular forms “a” , “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms “comprises” , “comprising” , “has” , “having” , “includes” and / or “including” , when used herein, specify the presence of stated features, elements, and / or components etc., but do not preclude the presence or addition of one or more other features, elements, components and / or combinations thereof. Example Environment
[0027] Fig. 1 is a block diagram that illustrates an example video coding system 100 that may utilize the techniques of this disclosure. As shown, the video coding system 100 may include a source device 110 and a destination device 120. The source device 110 can be also referred to as a video encoding device, and the destination device 120 can be also referred to as a video decoding device. In operation, the source device 110 can be configured to generate encoded video data and the destination device 120 can be configured to decode the encoded video data generated by the source device 110. The source device 110 may include a video source 112, a video encoder 114, and an input / output (I / O) interface 116.
[0028] The video source 112 may include a source such as a video capture device. Examples of the video capture device include, but are not limited to, an interface to receive video data from a video content provider, a computer graphics system for generating video data, and / or a combination thereof.
[0029] The video data may comprise one or more pictures. The video encoder 114 encodes the video data from the video source 112 to generate a bitstream. The bitstream may include a sequence of bits that form a coded representation of the video data. The bitstream may include coded pictures and associated data. The coded picture is a coded representation of a picture. The associated data may include sequence parameter sets, picture parameter sets, and other syntax structures. The I / O interface 116 may include a modulator / demodulator and / or a transmitter. The encoded video data may be transmitted directly to destination device 120 via the I / O interface 116 through the network 130A. The encoded video data may also be stored onto a storage medium / server 130B for access by destination device 120.
[0030] The destination device 120 may include an I / O interface 126, a video decoder 124, and a display device 122. The I / O interface 126 may include a receiver and / or a modem. The I / O interface 126 may acquire encoded video data from the source device 110 or the storage medium / server 130B. The video decoder 124 may decode the encoded video data. The display device 122 may display the decoded video data to a user. The display device 122 may be integrated with the destination device 120, or may be external to the destination device 120 which is configured to interface with an external display device.
[0031] The video encoder 114 and the video decoder 124 may operate according to a video compression standard, such as the High Efficiency Video Coding (HEVC) standard, Versatile Video Coding (VVC) standard and other current and / or further standards.
[0032] Fig. 2 is a block diagram illustrating an example of a video encoder 200, which may be an example of the video encoder 114 in the system 100 illustrated in Fig. 1, in accordance with some embodiments of the present disclosure.
[0033] The video encoder 200 may be configured to implement any or all of the techniques of this disclosure. In the example of Fig. 2, the video encoder 200 includes a plurality of functional components. The techniques described in this disclosure may be shared among the various components of the video encoder 200. In some examples, a processor may be configured to perform any or all of the techniques described in this disclosure.
[0034] In some embodiments, the video encoder 200 may include a partition unit 201, a prediction unit 202 which may include a mode select unit 203, a motion estimation unit 204, a motion compensation unit 205 and an intra-prediction unit 206, a residual generation unit 207, a transform unit 208, a quantization unit 209, an inverse quantization unit 210, an inverse transform unit 211, a reconstruction unit 212, a buffer 213, and an entropy encoding unit 214.
[0035] In other examples, the video encoder 200 may include more, fewer, or different functional components. In an example, the prediction unit 202 may include an intra block copy (IBC) unit. The IBC unit may perform prediction in an IBC mode in which at least one reference picture is a picture where the current video block is located.
[0036] Furthermore, although some components, such as the motion estimation unit 204 and the motion compensation unit 205, may be integrated, but are represented in the example of Fig. 2 separately for purposes of explanation.
[0037] The partition unit 201 may partition a picture into one or more video blocks. The video encoder 200 and the video decoder 300 may support various video block sizes.
[0038] The mode select unit 203 may select one of the coding modes, intra or inter, e.g., based on error results, and provide the resulting intra-coded or inter-coded block to a residual generation unit 207 to generate residual block data and to a reconstruction unit 212 to reconstruct the encoded block for use as a reference picture. In some examples, the mode select unit 203 may select a combined inter and intra prediction (CIIP) mode in which the prediction is based on an inter prediction signal and an intra prediction signal. The mode select unit 203 may also select a resolution for a motion vector (e.g., a sub-pixel or integer pixel precision) for the block in the case of inter-prediction.
[0039] To perform inter prediction on a current video block, the motion estimation unit 204 may generate motion information for the current video block by comparing one or more reference frames from buffer 213 to the current video block. The motion compensation unit 205 may determine a predicted video block for the current video block based on the motion information and decoded samples of pictures from the buffer 213 other than the picture associated with the current video block.
[0040] The motion estimation unit 204 and the motion compensation unit 205 may perform different operations for a current video block, for example, depending on whether the current video block is in an I-slice, a P-slice, or a B-slice. As used herein, an “I-slice” may refer to a portion of a picture composed of macroblocks, all of which are based upon macroblocks within the same picture. Further, as used herein, in some aspects, “P-slices” and “B-slices” may refer to portions of a picture composed of macroblocks that are not dependent on macroblocks in the same picture.
[0041] In some examples, the motion estimation unit 204 may perform uni-directional prediction for the current video block, and the motion estimation unit 204 may search reference pictures of list 0 or list 1 for a reference video block for the current video block. The motion estimation unit 204 may then generate a reference index that indicates the reference picture in list 0 or list 1 that contains the reference video block and a motion vector that indicates a spatial displacement between the current video block and the reference video block. The motion estimation unit 204 may output the reference index, a prediction direction indicator, and the motion vector as the motion information of the current video block. The motion compensation unit 205 may generate the predicted video block of the current video block based on the reference video block indicated by the motion information of the current video block.
[0042] Alternatively, in other examples, the motion estimation unit 204 may perform bi-directional prediction for the current video block. The motion estimation unit 204 may search the reference pictures in list 0 for a reference video block for the current video block and may also search the reference pictures in list 1 for another reference video block for the current video block. The motion estimation unit 204 may then generate reference indexes that indicate the reference pictures in list 0 and list 1 containing the reference video blocks and motion vectors that indicate spatial displacements between the reference video blocks and the current video block. The motion estimation unit 204 may output the reference indexes and the motion vectors of the current video block as the motion information of the current video block. The motion compensation unit 205 may generate the predicted video block of the current video block based on the reference video blocks indicated by the motion information of the current video block.
[0043] In some examples, the motion estimation unit 204 may output a full set of motion information for decoding processing of a decoder. Alternatively, in some embodiments, the motion estimation unit 204 may signal the motion information of the current video block with reference to the motion information of another video block. For example, the motion estimation unit 204 may determine that the motion information of the current video block is sufficiently similar to the motion information of a neighboring video block.
[0044] In one example, the motion estimation unit 204 may indicate, in a syntax structure associated with the current video block, a value that indicates to the video decoder 300 that the current video block has the same motion information as the another video block.
[0045] In another example, the motion estimation unit 204 may identify, in a syntax structure associated with the current video block, another video block and a motion vector difference (MVD) . The motion vector difference indicates a difference between the motion vector of the current video block and the motion vector of the indicated video block. The video decoder 300 may use the motion vector of the indicated video block and the motion vector difference to determine the motion vector of the current video block.
[0046] As discussed above, video encoder 200 may predictively signal the motion vector. Two examples of predictive signaling techniques that may be implemented by video encoder 200 include advanced motion vector prediction (AMVP) and merge mode signaling.
[0047] The intra prediction unit 206 may perform intra prediction on the current video block. When the intra prediction unit 206 performs intra prediction on the current video block, the intra prediction unit 206 may generate prediction data for the current video block based on decoded samples of other video blocks in the same picture. The prediction data for the current video block may include a predicted video block and various syntax elements.
[0048] The residual generation unit 207 may generate residual data for the current video block by subtracting (e.g., indicated by the minus sign) the predicted video block (s) of the current video block from the current video block. The residual data of the current video block may include residual video blocks that correspond to different sample components of the samples in the current video block.
[0049] In other examples, there may be no residual data for the current video block, for example in a skip mode, and the residual generation unit 207 may not perform the subtracting operation.
[0050] The transform unit 208 may generate one or more transform coefficient video blocks for the current video block by applying one or more transforms to a residual video block associated with the current video block.
[0051] After the transform unit 208 generates a transform coefficient video block associated with the current video block, the quantization unit 209 may quantize the transform coefficient video block associated with the current video block based on one or more quantization parameter (QP) values associated with the current video block.
[0052] The inverse quantization unit 210 and the inverse transform unit 211 may apply inverse quantization and inverse transforms to the transform coefficient video block, respectively, to reconstruct a residual video block from the transform coefficient video block. The reconstruction unit 212 may add the reconstructed residual video block to corresponding samples from one or more predicted video blocks generated by the prediction unit 202 to produce a reconstructed video block associated with the current video block for storage in the buffer 213.
[0053] After the reconstruction unit 212 reconstructs the video block, loop filtering operation may be performed to reduce video blocking artifacts in the video block.
[0054] The entropy encoding unit 214 may receive data from other functional components of the video encoder 200. When the entropy encoding unit 214 receives the data, the entropy encoding unit 214 may perform one or more entropy encoding operations to generate entropy encoded data and output a bitstream that includes the entropy encoded data.
[0055] Fig. 3 is a block diagram illustrating an example of a video decoder 300, which may be an example of the video decoder 124 in the system 100 illustrated in Fig. 1, in accordance with some embodiments of the present disclosure.
[0056] The video decoder 300 may be configured to perform any or all of the techniques of this disclosure. In the example of Fig. 3, the video decoder 300 includes a plurality of functional components. The techniques described in this disclosure may be shared among the various components of the video decoder 300. In some examples, a processor may be configured to perform any or all of the techniques described in this disclosure.
[0057] In the example of Fig. 3, the video decoder 300 includes an entropy decoding unit 301, a motion compensation unit 302, an intra prediction unit 303, an inverse quantization unit 304, an inverse transform unit 305, a reconstruction unit 306 and a buffer 307. The video decoder 300 may, in some examples, perform a decoding pass generally reciprocal to the encoding pass described with respect to video encoder 200.
[0058] The entropy decoding unit 301 may retrieve an encoded bitstream. The encoded bitstream may include entropy coded video data (e.g., encoded blocks of video data) . The entropy decoding unit 301 may decode the entropy coded video data, and from the entropy decoded video data, the motion compensation unit 302 may determine motion information including motion vectors, motion vector precision, reference picture list indexes, and other motion information. The motion compensation unit 302 may, for example, determine such information by performing the AMVP and merge mode. AMVP is used, including derivation of several most probable candidates based on data from adjacent PBs and the reference picture. Motion information typically includes the horizontal and vertical motion vector displacement values, one or two reference picture indices, and, in the case of prediction regions in B slices, an identification of which reference picture list is associated with each index. As used herein, in some aspects, a “merge mode” may refer to deriving the motion information from spatially or temporally neighboring blocks.
[0059] The motion compensation unit 302 may produce motion compensated blocks, possibly performing interpolation based on interpolation filters. Identifiers for interpolation filters to be used with sub-pixel precision may be included in the syntax elements.
[0060] The motion compensation unit 302 may use the interpolation filters as used by the video encoder 200 during encoding of the video block to calculate interpolated values for sub-integer pixels of a reference block. The motion compensation unit 302 may determine the interpolation filters used by the video encoder 200 according to the received syntax information and use the interpolation filters to produce predictive blocks.
[0061] The motion compensation unit 302 may use at least part of the syntax information to determine sizes of blocks used to encode frame (s) and / or slice (s) of the encoded video sequence, partition information that describes how each macroblock of a picture of the encoded video sequence is partitioned, modes indicating how each partition is encoded, one or more reference frames (and reference frame lists) for each inter-encoded block, and other information to decode the encoded video sequence. As used herein, in some aspects, a “slice” may refer to a data structure that can be decoded independently from other slices of the same picture, in terms of entropy coding, signal prediction, and residual signal reconstruction. A slice can either be an entire picture or a region of a picture.
[0062] The intra prediction unit 303 may use intra prediction modes for example received in the bitstream to form a prediction block from spatially adjacent blocks. The inverse quantization unit 304 inverse quantizes, i.e., de-quantizes, the quantized video block coefficients provided in the bitstream and decoded by entropy decoding unit 301. The inverse transform unit 305 applies an inverse transform.
[0063] The reconstruction unit 306 may obtain the decoded blocks, e.g., by summing the residual blocks with the corresponding prediction blocks generated by the motion compensation unit 302 or intra-prediction unit 303. If desired, a deblocking filter may also be applied to filter the decoded blocks in order to remove blockiness artifacts. The decoded video blocks are then stored in the buffer 307, which provides reference blocks for subsequent motion compensation / intra prediction and also produces decoded video for presentation on a display device.
[0064] Some example embodiments of the present disclosure will be described in detailed hereinafter. It should be understood that section headings are used in the present document to facilitate ease of understanding and do not limit the embodiments disclosed in a section to only that section. Furthermore, while certain embodiments are described with reference to Versatile Video Coding or other specific video codecs, the disclosed techniques are applicable to other video coding technologies also. Furthermore, while some embodiments describe video coding steps in detail, it will be understood that corresponding steps decoding that undo the coding will be implemented by a decoder. Furthermore, the term video processing encompasses video coding or compression, video decoding or decompression and video transcoding in which video pixels are represented from one compressed format into another compressed format or at a different compressed bitrate. 1. Brief summary
[0065] This disclosure is related to video coding technologies. Specifically, it is related to complexity control in a video encoder. It may be applied to the encoders compatible with existing video coding standards like VVC, AV1, HEVC, and AVC. It may be also used as a technology for future video coding standards or video codecs, and could be extended to other fields involving low-complexity and low-latency video encoding tasks. 2. Introduction
[0066] The video and live streaming industry has experienced a rapid growth and evolution in recent years. The emergence of live streaming platforms like Twitch, YouTube Live, TikTok, and Facebook Live has enabled widespread real-time video broadcasting, which has become hugely popular for gaming, sports, events, and other immediate content.
[0067] At the same time, consumer expectations for video quality have steadily risen. High definition 1080p video has become the standard, and 4K ultra-HD video is gaining mainstream adoption. Beyond just increasing resolution, new video formats like 360-degree, virtual reality (VR) , and augmented reality (AR) videos are also emerging, presenting new technical challenges.
[0068] To meet the demands of these evolving video applications, video codecs have to become more advanced. Key requirements include higher compression efficiency to enable streaming of high-resolution video at reasonable bitrates, support for features like HDR, wide color gamut, and high frame rates, low encoding / decoding latency for live streaming and interactive applications, and hardware acceleration for power-efficient implementation on mobile devices.
[0069] Standards like AVC / H. 264, HEVC / H. 265, and the latest AV1 and VVC are attempting to address these needs and provide the video quality and features required by modern video services and consumer devices. The video industry's constant push for higher quality and new capabilities will continue driving advances in video compression technology. 2.1 The development of video coding
[0070] The evolution of video coding standards has primarily been driven by the development of the well-known ITU-T and ISO / IEC standards. The ITU-T produced H. 261 and H. 263, while ISO / IEC produced MPEG-1 and MPEG-4 Visual. The two organizations jointly developed the H. 262 / MPEG-2 Video and H. 264 / MPEG-4 Advanced Video Coding (AVC) as well as the H. 265 / HEVC standards. Since H. 262, the video coding standards have been based on the hybrid video coding structure, which utilizes temporal prediction and transform coding.
[0071] To explore future video coding technologies beyond HEVC, the Joint Video Exploration Team (JVET) was founded jointly by VCEG and MPEG in 2015. In April 2018, the Joint Video Expert Team (JVET) between VCEG (Q6 / 16) and ISO / IEC JTC1 SC29 / WG11 (MPEG) was created to work on the VVC standard, targeting a 50%bitrate reduction compared to HEVC.
[0072] In mid-2013, Google launched and deployed the VP9 video codec, which achieved around 50%higher compression performance than the prior H. 264 / AVC codec, and was quickly adopted by the industry. As the demand for high-performance video compression continued growing, the Alliance for Open Media (AOM) was formed in 2015 as a consortium for the development of open, royalty-free technology for multimedia delivery. Its first video compression format, AV1, released in 2018, enabled about a 30%compression gain over its predecessor, VP9. AV1 is already supported by many web platforms, including Android, Chrome, Microsoft Edge, and Firefox, and multiple web-based video service providers, including YouTube, Netflix, Vimeo, and Bitmovin, have begun rolling out AV1 streaming services at scale.
[0073] To enhance coding efficiency, each successive generation of video coding standards has incorporated numerous novel coding tools beyond those of the preceding generation. For example, VVC has added the Quaternary Tree Plus Multi-Type Tree (QT+MTT) partition structure of the Coding Unit (CU) , Position-Dependent Prediction Combination (PDPC) , Multiple Reference Line Intra Prediction (MRL) , Affine Motion Compensation (AMC) , History-based Motion Vector Prediction (HMVP) , Adaptive Loop Filter (ALF) , etc. These new technologies introduced in VVC have achieved great benefits in coding efficiency, but the coding complexity brought by these coding tools is also much higher than that of HEVC. In the JVET-T0003 report, the coding complexity of the VVC reference software VTM-10.0 in All Intra mode (AI) is 26 times higher than that of HEVC on average, and the coding complexity in Random Access mode (RA) is 8.5 times higher than that of HEVC.
[0074] The increased coding complexity of VVC and AV1 presents significant implications for live video streaming scenarios, where low-latency encoding and decoding are essential. The computational demands of VVC encoding may limit the scalability and real-time performance of live video transcoding, potentially leading to increased latency or reduced video quality. Overcoming these challenges will require further advancements in hardware acceleration and optimization of VVC encoding and decoding algorithms. 2.2 Multi-channel transcoding of real-time streams
[0075] In various application scenarios, the encoder needs to meet specific requirements. Two common transcoding scenarios are offline transcoding and live transcoding.
[0076] Offline transcoding is the process of transcoding existing video files, without real-time or latency requirements, allowing more time for optimized encoding. Offline transcoding can employ more complex encoding algorithms and parameter optimizations, as well as more sophisticated bitrate control strategies, to achieve higher compression efficiency. Offline transcoding is suitable for video-on-demand, video storage, and other non-real-time video applications.
[0077] Live transcoding is performed in real-time, with the requirement to complete the encoding immediately after the video stream is fed into the encoder. This use case faces a very tight timing constraint, typically requiring latency within a few hundred milliseconds to ensure a satisfactory viewing experience for the audience. To maintain a good video quality, live transcoding must dynamically adapt to network conditions, performing real-time bitrate adjustments. As a result, this technique is primarily utilized for real-time video streaming scenarios, such as live video broadcasting, where the need for low latency takes precedence over optimization for compression efficiency. 2.2.1 Real-time performance and latency
[0078] The core value of live streaming lies in its ability to capture and disseminate events in real-time, allowing audience to witness the content as it happens. Live streaming requires real-time transcoding to accommodate different devices and network conditions. The transcoded streams need to be received in a timely manner to ensure a seamless viewing experience.
[0079] The real-time transcoding performance refers to the ability to process and generate the transcoded streams without introducing excessive delays. This is essential to maintain the live and interactive nature of the content. Factors affecting real-time transcoding performance include the complexity of the transcoding algorithms, the computing power of the transcoding infrastructure, and the utilization of system resources. Efficient resource allocation and optimized transcoding algorithms are necessary to achieve high real-time performance, ensuring that the transcoded streams are available with minimal latency.
[0080] Despite the emphasis on real-time performance, some level of latency in the transcoding process is unavoidable. This end-to-end latency can be introduced by the encoding, decoding, and the actual transcoding operations. Excessive transcoding latency can negatively impact the overall live streaming experience, especially in scenarios where interactivity is crucial, such as live events or interactive live streams. Viewers may experience a noticeable time lag between the live action and the transcoded stream received, which can degrade the sense of immediacy and engagement.
[0081] Minimizing the transcoding latency is crucial to provide a seamless and synchronized viewing experience for the audience. This can be achieved through techniques like hardware-acceleration, optimized software algorithms, and efficient resource management. Live streaming platforms often need to achieve a trade-off between real-time transcoding performance and transcoding latency, considering factors like content type, audience size, and device diversity. 2.2.2 Low latency GOP structure
[0082] Normally, video streams are encoded using repeating compression patterns that utilize different types of frames, forming a structure known as a group of pictures (GOP) . Namely, · I-frames, or keyframes, are standalone images and do not reference earlier or later images. · P-frames reference earlier frames and represent encoding of the current frame using changes from a (sometimes multiple) previous frames, and usually compress better than I-frames. · B-frames can reference earlier and later frames, and usually compress better than P-frames and much better than I-frames.
[0083] In an encoder, to improve inter-frame prediction, a hierarchical structure is typically used for encoding, distributing frames across different temporal layers. However, this approach can introduce encoding delay. When there is no latency requirement, most GOPs of HEVC or VVC consist of one I frame, a few P frames, and mostly B frames for best compression. One example GOP structure might be GOP16. Fig. 4 illustrates a GOP structure of size equal to 16.
[0084] When aiming for low-latency video compression, certain trade-offs must be considered. Since B-frames rely on future frames acquired after them for compression, they cannot be encoded immediately. If a B-frame may only reference the next frame, it will add 1 frame of delay. If a B-frame may reference up to 4 future frames, it will add 4 frames of delay. For streaming applications, this can also create uneven processing bottlenecks -for example, in a GOP16 structure, compression of the first B-frame must wait until the data for the first P-frame is ready, thus incurring an additional 16 frames of input lag, plus processing time for each B-frame, further increasing the latency.
[0085] If ultra-low latency is required and a slight reduction in compression efficiency can be tolerated, using only P-frames can be a solution, as they do not suffer from the above effects. However, for real-time live scenarios, we need to maintain low latency while minimizing the loss of compression efficiency, so a trade-off between quality and latency needs to be found. Typically, a GOP structure with a size of 4 is used, as shown in Fig. 5. In this GOP, there are 3 B-frames, which greatly improves coding efficiency compared to not using B-frames, while only incurring a 4-frame delay, which is 1 / 4 of the delay of a GOP16 structure. 2.2.3 Random Access in live streaming
[0086] Random Access is an important feature in video encoding that provides users with the ability to quickly locate and jump to specific content within a video. This functionality plays a crucial role in various video application scenarios.
[0087] The principle behind Random Access is the insertion of I-frames within the GOP structure. This allows the video to be decoded starting from any I-frame, enabling users to freely jump to their desired video location and begin watching immediately, without the need to start from the beginning. The distribution of key frames, or I-frames, is the key to Random Access, and it is typically optimized based on the importance of the content and user behavior patterns. The distance between two consecutive I-frames is commonly referred to as the intra-period.
[0088] The implementation of Random Access relies heavily on the length of the intra-period. Generally, the shorter the intra-period, the lower the random-access latency, but the compression efficiency will also decrease. Video applications often need to balance the trade-off between latency and compression efficiency.
[0089] In the context of live streaming, users may join the stream at any given time, so periodic insertion of key frames is necessary to enable random access. Typically, the intra-period is set to 2 seconds or 4 seconds, striking a balance between initial delay and compression efficiency. 2.2.4 Multi-channel transcoding
[0090] Multi-channel video transcoding refers to the process of simultaneously running multiple video transcoding tasks on the one machine. This approach can fully utilize the machine's CPU and memory resources, thereby improving the efficiency and throughput of video encoding. The main advantages of multi-channel transcoding include increased resource utilization, reduced costs, and more convenient centralized management and control of multiple video encoding tasks.
[0091] CPU utilization is an important metric for multi-channel transcoding. Due to the need to process multiple video streams simultaneously, multi-channel transcoding can significantly increase CPU utilization, which is typically in the range of 70%-90%, depending on factors such as video encoding parameters, resolution, and frame rate. To maximize CPU utilization, it is necessary to properly allocate CPU resources to each video encoding task, but over-optimization can also lead to CPU resource constraints.
[0092] The number of concurrent channels is a key factor affecting CPU utilization. The more video encoding tasks running simultaneously on a machine, the higher the CPU utilization will be. For example, on a 128-core CPU, when the CPU is not fully utilized, adding an additional encoding task typically results in an increase in CPU utilization. In real-time multi-channel transcoding scenarios, the primary requirement is to ensure transcoding timeliness. The goal is to maximize CPU utilization while maintaining real-time performance, which means finding the optimal number of concurrent channels. Excessively high CPU utilization can degrade the real-time performance of individual video streams. Therefore, when selecting the number of concurrent channels, it is necessary to balance CPU utilization and real-time performance to find the optimal trade-off.
[0093] Within multi-channels transcoding tasks, a single machine may simultaneously transcode both simple and complex video sequences, and it may also run encoders based on different encoding standards. As a result, the CPU utilization of different transcoding tasks on the same machine can vary, and they can influence each other. In situations with high CPU utilization, complex transcoding tasks may cause their own or other transcoding tasks to lose real-time performance. Therefore, appropriately controlling the complexity of transcoding tasks can significantly help improve the CPU utilization in multi-stream transcoding. 3 Problems 1. Complex videos are prone to non-real-time situations when the CPU utilization is high. In order to achieve real-time, the number of concurrent channels must be reduced, which will reduce the CPU uti-lization. 2. The encoder can choose a faster preset to improve real-time performance, but this will lead to a decrease in the efficiency of all video encoding. A complexity-adaptive encoding scheme is needed to achieve a trade-off between encoding complexity and encoding efficiency. 4 Detailed solutions
[0094] In this disclosure, by detecting the complexity of the encoder's current transcoding and whether it is real-time during the transcoding process, the encoding complexity can be adaptively controlled by adjusting parameters in real time. This can not only solve the problem of complex sequences not being real-time, but also avoid the loss of simple sequence transcoding efficiency. 4.1 Real-time detection
[0095] When most machines are idle or the number of concurrent channels is relatively small, even particularly complex transcoding tasks are likely to be real-time. Non-real-time situations usually occur when the CPU is under high load situation. Therefore, it is necessary to first detect whether the transcoding task is real-time. Only when it is not real-time, adaptive encoding complexity control will be performed.
[0096] Encoders usually use frame parallel processing to achieve effective encoding, so the time difference between input and output of a single frame cannot well reflect real-time performance. Such a frame window is defined, as shown in Fig. 6.
[0097] When each frame is encoded, a queue is used to record the time when each frame finishes encoding. The encoding time of the frame window consisting of those frames between Nth frame and Mth frame in the encoding order is defined as (TM -TN) . speed_factor is calculated as
[0098] Where dtsM and dtsN indicates decoding time stamp of frame M and N. In the real-time live transcoding scenario, the ideal standard for real-time is that speed_factor equals 1. When speed_factor is less than 1, the current frame window will be marked as non-real-time.
[0099] In fact, the calculated speed_factor will fluctuate to a certain extent. Therefore, in order to reduce the impact of fluctuations, we need to have L consecutive frame windows that do not meet the real-time requirement before we determine that the current transcoding is in a non-real-time state. 4.2 Complexity model
[0100] When the transcoding task meets the non-real-time condition specified in section 4.1, the encoder needs to control the complexity of the transcoding task. To what extent the complexity should be adjusted, the complexity of the current transcoding task needs to be evaluated.
[0101] The complexity model is to evaluate all frames within an intra period interval. First, the Single frame complexity model is used to evaluate the complexity of each temporal level frame, including I frame (CI) , temporal level 0 frame (CTL0) , temporal level 1 frame (CTL1) and temporal level 2 frame (CTL2) , in the case of GOP4. Then the intra period interval complexity model is used to estimate the final intra period interval complexity based on the average complexity calculated for each temporal level frame. 4.2.1 Single frame complexity model
[0102] A multivariate linear function is used to predict the complexity of a single frame. the single frame complexity is defined as Cf = F (Xsf) , where F is a linear function and Xsf is the input vector. The Xsf = [cu_num, avg_depth, avg_qp, cbf0_num, skip_num, merge_num, amvp_num, intra_num, bits] is as follows: · cu_num: the total number of cu in the current frame. · avg_depth: the average partition depth in the current frame. · avg_qp: the average qp in the current frame. · cbf0_num: the number of CUs whose cbf is equal to 0 in the current frame. · skip_num: the number of skip modes selected by CU in the current frame. · merge_num: the number of merge modes selected by CU in the current frame. · amvp_num: the number of amvp modes selected by CU in the current frame. · intra_num: the number of intra modes selected by CU in the current frame. · bits: the number of bits required to encode the current frame.
[0103] For I slice, skip_num, merge_num, amvp_num and intra_num all equal to 0.
[0104] The coefficients of the linear function F are obtained by linear regression. In the linear regression dataset, Cf is the encoding time of a single frame in a single-threaded encoding task. This ensures the accuracy of single-frame complexity mode and will not be affected by thread waiting.
[0105] In addition, the input vector can also be obtained from the decoder, so it is more flexible to choose where to perform complexity estimation. 4.2.2 Intra period interval complexity classification model
[0106] A multi-stage logistic regression classifier is used to classify intra period interval complexity into K levels, where K is an integer power of 2. The complexity increases gradually from level 0 to K-1. According to different complexity levels, different parameters are adjusted. When the value of K is 4, multi-stage logistic regression classifier is shown in Fig. 7.
[0107] Here each classification uses a logistic regression classifier, but they are trained separately. Xip is the input vector. The X = [CI, avg_CTL0, avg_CTL1, avg_CTL2] is as follows: · CI: The complexity of I frame in the current intra period interval. · avg_CTL0: the average complexity of temporal level 0 in the current intra period interval. · avg_CTL1: the average complexity of temporal level 1 in the current intra period interval. · avg_CTL2: the average complexity of temporal level 2 in the current intra period interval.
[0108] When obtaining the training dataset, sequence was performed multi-channels transcoding on a single machine, control all sequences to use the same number of concurrent channels, and evaluate the real-time performance of each intra period interval under the corresponding number of concurrent channels. Analyze the real-time performance of each intra period interval in different channels and find their maximum channels under real-time conditions. The larger the maximum channels by each task, the higher the transcoding complexity. the ground truth of the complexity level is obtained by classifying the maximum channels by each task. 4.3 Parameter adjustment
[0109] When the transcoding task meets the non-real-time condition specified in section 4.1, and after obtaining the complexity level of the transcoding task, the parameters will be adjusted to control the complexity.
[0110] There are two methods to adjust parameters. The first method does not require any changes to the SPS syntax. It can be achieved by reducing the number of coding mode searches, such as reducing the number of RDOs, using more aggressive fast algorithms, etc. The advantage of this method is that the parameters can be adjusted almost immediately without waiting for the arrival of the next I frame. This solution is suitable for transcoding tasks that are not particularly complex.
[0111] The second method is a more aggressive method, which greatly reduces the coding complexity by turning off some coding tools. These encoding tools are generally controlled by the SPS syntax. Flexible switching can be achieved by transmitting multiple SPS syntaxes, but this will introduce additional transmission bits. One way to avoid adding extra bits is to wait for the next I frame to arrive and adjust the parameters. However, this will introduce a certain parameter adjustment delay. For real-time live broadcast scenarios, the intra period is small, so the delay in parameter adjustment is within acceptable range. Therefore, for particularly complex sequences, more aggressive parameter adjustments will be performed when the next I frame arrives.
[0112] K parameter groups are constructed, one for each of the K complexity levels. Some low-complexity sequences are selected as the benchmark, and the default parameters are used for multi-channels transcoding. The average number of CPU cores occupied by each task under real-time conditions is calculated. These K parameter groups are tuned by turning off the coding tool, adjusting the partition depth, the number of RDOs, and the threshold of the fast algorithm. This allows the transcoding task to achieve the same CPU core usage at the corresponding complexity level.
[0113] Fig. 8 shows a flowchart 800 of the complexity-adaptive video transcoding solution. First, at 810, real-time detection and single-frame complexity estimation will be performed on the current transcoding task. At 820, whether the transcoding task meets the real-time condition (s) is determined. If the current transcoding task meets the real-time conditions, no changes are required because the CPU resources are sufficient. If the real-time conditions cannot be met, at 830, the intra period interval complexity classification model will be used to classify the complexity of the current transcoding task, and then at 840, the corresponding parameters according to the complexity level will be adjusted. 5 Embodiment
[0114] The detailed embodiments below should be considered as examples to explain general concepts. These embodiments should not be interpreted in a narrow way. Furthermore, these embodiments can be combined in any manner.A real-time performance detection method for video transcoding is provided. 1. Detecting real-time performance in a frame window. a) A frame window may contain all frames within a time period. i. In one example, a frame window contains all frames from N + 1 to M in the encoding order, as shown in Fig. 6. 1. In one example, the size of M -N may be equal to 2 times the frame rate. b) Encoding time of a frame window calculation: The encoding time of the frame window consisting of those frames between the Nth frame and the Mth frame in the encoding order is defined as (TM -TN) . i. In one example, TM and TN may be the time when the encoding of the Mth frame and the Nth frame is completed. ii. In one example, a queue is used to record the time when each frame finish encoding. TM and TN may be the head and tail of this queue. 2. Speed factor of a frame window calculation: a) The speed_factor is calculated as expected encoding time divided by actual encoding time. i. In one example, expected coding time of a frame window may be equal to dtsM -dtsN. ii. In one example, expected coding time of a frame window may be equal to (M -N) / frame_rate. iii. In one example, actual encoding time of a frame window may be equal to (TM -TN) . b) In one example, the speed_factor may be equal to 1 or less than 1. i. In one example, the speed_factor is equal to or greater than 1 means that the coding of frame window is real-time. ii. In one example, the speed_factor less than 1 may mean that the frame window is non-real- time. 3. Detecting real-time performance in transcoding task. a) In one example, L consecutive frame windows that do not meet the real-time requirement may mean that the current transcoding is in a non-real-time state. i. In one example, L may be equal to 25.A single-frame complexity estimation model for video coding is provided. 1. A multivariate linear function is used to predict the complexity of a single frame. a) In one example, the input of multivariate linear function may be X = [cu_num, avg_depth, avg_qp, cbf0_num, skip_num, merge_num, amvp_num, intra_num, bits] . i. In one example, cu_num may be the total number of CUs in the current frame. ii. In one example, avg_depth may be the average partition depth in the current frame. iii. In one example, avg_qp may be the average qp in the current frame. iv. In one example, cbf0_num may be the number of CUs whose cbf is equal to 0 in the current frame. v. In one example, skip_num may be the number of skip modes selected by CU in the current frame. vi. In one example, merge_num may be the number of merge modes selected by CU in the current frame. vii. In one example, amvp_num may be the number of amvp modes selected by CU in the current frame. viii. In one example, intra_num may be the number of intra modes selected by CU in the current frame. ix. In one example, bits may be the number of bits required to encode the current frame. x. Alternatively, in one example, for I slice, skip_num, merge_num, amvp_num and intra_num may equal to 0. xi. Alternatively, in one example, the input vector may be obtained from the decoder. b) In one example, the coefficients of the linear function F are obtained by linear regression. i. In one example, in linear regression dataset, single-frame complexity ground truth may be the encoding time of a single frame in a single-threaded encoding task. ii. In one example, frames from different temporal layers may be trained separately.A complexity estimation model for video transcoding is provided. 1 .The complexity estimation mode is used to estimate complexity for an intra period interval. a) In one example, intra period interval may contain all frames between two adjacent I frames (in- cluding the first I frame) . b) A multi-stage logistic regression classifier is used to classify intra period interval complexity into K levels. i. In one example, K may be equal to 4. ii. Alternatively, in one example, K may be equal to an integer power of 2. iii. In one example, A multi-stage logistic regression classifier may consist of three identical but independently trained classifiers, as shown in Fig. 7. 1. In one example, each classification may use a logistic regression classifier. 2. In one example, the input of classification may be X = [CI, avg_CTL0, avg_CTL1, avg_CTL2] . a) In one example, CI may be the complexity of I frame in the current intra period in- terval. b) In one example, avg_CTL0: may be the average complexity of temporal level 0 in the current intra period interval. c) In one example, avg_CTL1: may be the average complexity of temporal level 1 in the current intra period interval. d) In one example, avg_CTL2: may be the average complexity of temporal level 2 in the current intra period interval. e) In one example, the grand truth of the complexity level is obtained by classifying the maximum channels by each task. i. In one example, sequence was performed multi-channels transcoding on a single machine, control all sequences to use the same number of concurrent channels, and evaluate the real-time performance of each intra period inter-val under the corresponding number of concurrent channels. Analyze the real-time performance of each intra period interval in different channels and find their maximum channels under real-time conditions.A complexity-aware video transcoding method for real-time video streaming 1. When real-time performance detection method for video transcoding and single-frame complexity es- timation model for video coding is applied, the real-time performance and complexity of current video transcoding task may be obtained, as shown in Fig. 8. a) In one example, real-time performance of current video transcoding task may be real-time or non- real-time. If current video transcoding is real-time, no actions need to take. Otherwise, a complexity estimation model for video transcoding will be applied and then the following parameters accord-ing to the complexity level will be adjusted. i. In one example, the number of RDO may be reduced. ii. Alternatively, in one example, some more aggressive fast algorithms may be used. iii. Alternatively, in one example, some coding tools may be turned off.
[0115] Fig. 9 illustrates a flowchart of a method 900 for video processing in accordance with embodiments of the present disclosure. The method 900 is implemented during a conversion between a video unit of a video and a bitstream of the video.
[0116] At block 910, for a conversion between a current video unit of a video and a bitstream of the video, complexity information associated with the conversion is determined. In some embodiments, the conversion comprises a transcoding process (also referred to as a “transcoding task” . The transcoding process comprises: obtaining the current video unit from a first bitstream of the video, and obtaining a second bitstream based on the current video unit. For example, complexity of the transcoding process may be determined.
[0117] Video transcoding refers to the process of converting a previously compressed video stream into another video stream to accommodate different network bandwidths, diverse terminal processing capabilities, and various user requirements. Video transcoding is essentially a process of decoding an input bitstream into a raw reconstructed video and subsequently re-encoding the video into another bitstream according to different settings, such as standard, resolution, and bitrate etc. Thus, the resulting stream before and after the conversion may belong to different video coding standards. For example, the conversion may include a video transcoding process from first decoded information such sa a first bitstream of a first coder to a second bitstream of a second coder. For example, the first coder may be an HEVC coder such as HEVC decoder, and the second coder may be a VVC coder such as VVC encoder, or the like. The decoded information from the first coder may be sent to the second coder to encode the video. It is to be understood that the first and second coders may be any suitable coder according to any suitable standard. Scope of the present disclosure is not limited here.
[0118] At block 920, at least one parameter associated with the conversion is obtained based on the complexity information.
[0119] At block 930, the conversion is performed based on the at least one parameter. In some embodiments, the conversion includes encoding the current video unit into the bitstream. Alternatively, or in addition, the conversion includes decoding the current video unit from the bitstream.
[0120] The method 900 enables obtaining or adjusting the parameter for the conversion such as for the transcoding based on the complexity information such as complexity level. In this manner, the coding complexity can be adaptively controlled by adjusting parameters in real time.
[0121] In some embodiments, the method 900 further comprises: obtaining at least one frame window of the video, a frame window comprising a set of frames within a time period; obtaining at least one encoding time of the at least one frame window; obtaining at least one decoding time stamp information of the at least one frame window; determining at least one speed factor for the at least one frame window based on the at least one encoding time and the at least one decoding time stamp information; and determining real-time information based on the at least one speed factor, the real-time information indicating whether coding the at least one frame window is real-time. For example, a frame window may be shown in Fig. 6, which may include (N+1) th frame to Mth frame. A next frame window may include (N+2) th frame to (M+1) th frame. A previous frame window may include Nth frame to (M-1) th frame.
[0122] In some embodiments, for a first frame window comprising (N+1) th frame to Mth frame, N and M being integers, M being greater than N, a respective speed factor of the first frame window is determined by: where speed_factor denotes the speed factor, TM denotes a completion time for encoding the Mth frame, TN denotes a completion time for encoding the Nth frame, dtsM denotes a decoding time stamp of the Mth frame, and dtsN denotes a decoding time stamp of the Nth frame.
[0123] In some embodiments, a size of (M-N) may be equal to a frame rate of the video times 2, or any other suitable value.
[0124] In some embodiments, a queue records encoding time information of frames in the first frame window, TM is stored in a head of the queue, and TN is stored in a tail of the queue.
[0125] In some embodiments, the at least one speed factor is determined by at least one expected coding time divided by the at least one encoding time.
[0126] In some embodiments, for a first frame window comprising (N+1) th frame to Mth frame, N and M being integers, M being greater than N, a respective expected coding time is determined by one of: (M -N) / frame_rate, or dtsM -dtsN, where frame_rate denotes a frame rate of the first frame window, dtsM denotes a decoding time stamp of the Mth frame, and dtsN denotes a decoding time stamp of the Nth frame.
[0127] In some embodiments, in response to a first speed factor of a first frame window being greater than or equal to 1, coding of the first frame window is real time, and in response to a second speed factor of a second frame window being less than 1, coding of the second frame window is non-real-time.
[0128] In some embodiments, the method 900 further comprises: based on a plurality of speed factors of a plurality of consecutive frame windows being less than 1, determining that a transcoding process of the video is non-real-time. For example, the number of the plurality of consecutive frame windows may be 25, or any other suitable value.
[0129] In some embodiments, determining the complexity information comprises: in response to a transcoding process of the video is non-real-time, determining the complexity information based on at least one model.
[0130] In some embodiments, determining the complexity information comprises: evaluating, with a single frame complexity model, a plurality of complexity levels of a plurality of temporal level frames of the video within an intra period interval; and determining, with an intra period interval complexity model, a final intra period interval complexity based on the plurality of complexity levels.
[0131] In some embodiments, the plurality of temporal level frames comprises an I frame, temporal level 0 frame, temporal level 1 frame and temporal level 2 frame in the case of group of pictures (GOP) 4.
[0132] In some embodiments, a complexity level of a current frame in the plurality of temporal level frames is determined by: Cf = F (Xsf) , wherein F is a linear function and Xsf is an input vector corresponding to the temporal level frame, where Xsf = [cu_num, avg_depth, avg_qp, cbf0_num, skip_num, merge_num, amvp_num, intra_num, bits] . cu_num denotes a total number of coding units (CUs) in the current frame, avg_depth denotes an average partition depth in the current frame, avg_qp denotes an average quantization parameter (QP) in the current frame, cbf0_num denotes the number of CUs whose cbf is equal to 0 in the current frame, skip_num denotes the number of skip modes selected by CU in the current frame, merge_num denotes the number of merge modes selected by a CU in the current frame, amvp_num denotes the number of amvp modes selected by a CU in the current frame, intra_num denotes the number of intra modes selected by a CU in the current frame, and bits denotes the number of bits required to encode the current frame. In some embodiments, for an I slice, skip_num, merge_num, amvp_num and intra_num are equal to 0. In some embodiments, the input vector may be obtained from a decoder.
[0133] In some embodiments, coefficients of the linear function are obtained by a linear regression based on a linear regression dataset, the linear regression dataset comprising encoding time information of at least one single frame in a single-threaded encoding task. In some embodiments, the coefficients of the linear function are trained for frames from a plurality of temporal layers separately.
[0134] In some embodiments, the intra period interval complexity model comprises a multi-stage logistic regression classifier used to classify intra period interval complexity into a set of levels, the number of the set of levels being an integer power of 2. Obtaining at least one parameter associated with the conversion based on the complexity information may include determining the at least one parameter based on a level of the intra period interval complexity.
[0135] In some embodiments, the multi-stage logistic regression classifier is trained separately for a plurality of temporal layers, and an input vector for the multi-stage logistic regression classifier comprises [CI, avg_CTL0, avg_CTL1, avg_CTL2] . CI denotes complexity of I frame in the current intra period interval, avg_CTL0 denotes average complexity of temporal level 0 in the current intra period interval, avg_CTL1 denotes average complexity of temporal level 1 in the current intra period interval, avg_CTL2 denotes average complexity of temporal level 2 in the current intra period interval.
[0136] In some embodiments, a training dataset for training the multi-stage logistic regression classifier is obtained by: performing multi-channels transcoding on at least one sequence on a single machine with the same number of concurrent channels; evaluating real-time performance of at least one intra period interval under the corresponding number of concurrent channels to obtain at least one maximum channel under a real-time condition; and obtaining ground truth of at least one complexity level by classifying the at least one maximum channel by corresponding task.
[0137] In some embodiments, obtaining at least one parameter associated with the conversion based on the complexity information comprises: reducing the at least one parameter based on a complexity level indicated by the complexity information, the at least one parameter comprising at least one of: the number of coding mode searches, or the number of rate-distortion optimization, and / or increasing the number of aggressive fast algorithms based on the complexity level indicated by the complexity information.
[0138] In some embodiments, obtaining at least one parameter associated with the conversion based on the complexity information comprises: disabling at least one coding tool based on a complexity level indicated by the complexity information.
[0139] In some embodiments, at least one sequence parameter set (SPS) syntax in the bitstream indicates the disabling of the at least one coding tool. In some embodiments, the at least one coding tool may be disabled for a next I frame.
[0140] In some embodiments, a set of parameters are obtained for a set of complexity levels by at least one of: turning off at least one coding tool, adjusting a partition depth, adjusting the number of rate-distortion optimization, or adjusting a threshold of a fast algorithm.
[0141] In some embodiments, obtaining at least one parameter associated with the conversion based on the complexity information comprises: in response to a transcoding process satisfying a real-time condition, keeping the at least one parameter unchanged; and in response to the transcoding process dissatisfying the real-time condition, adjusting the at least one parameter based on the complexity information.
[0142] According to further embodiments of the present disclosure, a non-transitory computer-readable recording medium is provided. The non-transitory computer-readable recording medium stores a bitstream of a video which is generated by a method performed by an apparatus for video processing. In the method, complexity information associated with generating of the bitstream of the video from a current video unit of the video is determined. At least one parameter for generating the bitstream is determined based on the complexity information. The bitstream is generated based on the at least one parameter.
[0143] According to still further embodiments of the present disclosure, a method for storing bitstream of a video is provided. In the method, complexity information associated with generating of the bitstream of the video from a current video unit of the video is determined. At least one parameter for generating the bitstream is determined based on the complexity information. The bitstream is generated based on the at least one parameter. The bitstream is stored in a non-transitory computer-readable recording medium.
[0144] Implementations of the present disclosure can be described in view of the following clauses, the features of which can be combined in any reasonable manner.
[0145] Clause 1. A method for video processing, comprising: determining, for a conversion between a current video unit of a video and a bitstream of the video, complexity information associated with the conversion; obtaining at least one parameter associated with the conversion based on the complexity information; and performing the conversion based on the at least one parameter.
[0146] Clause 2. The method of clause 1, wherein the conversion comprises a transcoding process, the transcoding process comprises: obtaining the current video unit from a first bitstream of the video, and obtaining a second bitstream based on the current video unit.
[0147] Clause 3. The method of clause 1 or 2, further comprising: obtaining at least one frame window of the video, a frame window comprising a set of frames within a time period; obtaining at least one encoding time of the at least one frame window; obtaining at least one decoding time stamp information of the at least one frame window; determining at least one speed factor for the at least one frame window based on the at least one encoding time and the at least one decoding time stamp information; and determining real-time information based on the at least one speed factor, the real-time information indicating whether coding the at least one frame window is real-time.
[0148] Clause 4. The method of clause 3, wherein for a first frame window comprising (N+1) th frame to Mth frame, N and M being integers, M being greater than N, a respective speed factor of the first frame window is determined by: wherein speed_factor denotes the speed factor, TM denotes a completion time for encoding the Mth frame, TN denotes a completion time for encoding the Nth frame, dtsM denotes a decoding time stamp of the Mth frame, and dtsN denotes a decoding time stamp of the Nth frame.
[0149] Clause 5. The method of clause 4, wherein a size of (M-N) is equal to a frame rate of the video times 2.
[0150] Clause 6. The method of clause 4, wherein a queue records encoding time information of frames in the first frame window, TM is stored in a head of the queue, and TN is stored in a tail of the queue.
[0151] Clause 7. The method of clause 3, wherein the at least one speed factor is determined by at least one expected coding time divided by the at least one encoding time.
[0152] Clause 8. The method of clause 7, wherein for a first frame window comprising (N+1) th frame to Mth frame, N and M being integers, M being greater than N, a respective expected coding time is determined by one of: (M -N) / frame_rate, or dtsM -dtsN, wherein frame_rate denotes a frame rate of the first frame window, dtsM denotes a decoding time stamp of the Mth frame, and dtsN denotes a decoding time stamp of the Nth frame.
[0153] Clause 9. The method of any of clauses 3 to 8, wherein in response to a first speed factor of a first frame window being greater than or equal to 1, coding of the first frame window is real time, and in response to a second speed factor of a second frame window being less than 1, coding of the second frame window is non-real-time.
[0154] Clause 10. The method of any of clauses 3 to 8, further comprising: based on a plurality of speed factors of a plurality of consecutive frame windows being less than 1, determining that a transcoding process of the video is non-real-time.
[0155] Clause 11. The method of clause 10, wherein a number of the plurality of consecutive frame windows is 25.
[0156] Clause 12. The method of any of clauses 3 to 11, wherein determining the complexity information comprises: in response to a transcoding process of the video is non-real-time, determining the complexity information based on at least one model.
[0157] Clause 13. The method of any of clauses 1 to 12, wherein determining the complexity information comprises: evaluating, with a single frame complexity model, a plurality of complexity levels of a plurality of temporal level frames of the video within an intra period interval; and determining, with an intra period interval complexity model, a final intra period interval complexity based on the plurality of complexity levels.
[0158] Clause 14. The method of clause 13, wherein the plurality of temporal level frames comprises an I frame, temporal level 0 frame, temporal level 1 frame and temporal level 2 frame in the case of group of pictures (GOP) 4.
[0159] Clause 15. The method of clause 13, wherein a complexity level of a current frame in the plurality of temporal level frames is determined by: Cf = F (Xsf) , wherein F is a linear function and Xsf is an input vector corresponding to the temporal level frame, wherein Xsf = [cu_num, avg_depth, avg_qp, cbf0_num, skip_num, merge_num, amvp_num, intra_num, bits] , and wherein: cu_num denotes a total number of coding units (CUs) in the current frame, avg_depth denotes an average partition depth in the current frame, avg_qp denotes an average quantization parameter (QP) in the current frame, cbf0_num denotes the number of CUs whose cbf is equal to 0 in the current frame, skip_num denotes the number of skip modes selected by CU in the current frame, merge_num denotes the number of merge modes selected by a CU in the current frame, amvp_num denotes the number of amvp modes selected by a CU in the current frame, intra_num denotes the number of intra modes selected by a CU in the current frame, and bits denotes the number of bits required to encode the current frame.
[0160] Clause 16. The method of clause 15, wherein for an I slice, skip_num, merge_num, amvp_num and intra_num are equal to 0.
[0161] Clause 17. The method of clause 15, wherein the input vector is obtained from a decoder.
[0162] Clause 18. The method of clause 15, wherein coefficients of the linear function are obtained by a linear regression based on a linear regression dataset, the linear regression dataset comprising encoding time information of at least one single frame in a single-threaded encoding task.
[0163] Clause 19. The method of clause 18, wherein the coefficients of the linear function are trained for frames from a plurality of temporal layers separately.
[0164] Clause 20. The method of clause 13, wherein the intra period interval complexity model comprises a multi-stage logistic regression classifier used to classify intra period interval complexity into a set of levels, the number of the set of levels being an integer power of 2, and wherein obtaining at least one parameter associated with the conversion based on the complexity information comprises determining the at least one parameter based on a level of the intra period interval complexity.
[0165] Clause 21. The method of clause 20, wherein the multi-stage logistic regression classifier is trained separately for a plurality of temporal layers, and an input vector for the multi-stage logistic regression classifier comprises [CI, avg_CTL0, avg_CTL1, avg_CTL2] , wherein: CI denotes complexity of I frame in the current intra period interval, avg_CTL0 denotes average complexity of temporal level 0 in the current intra period interval, avg_CTL1 denotes average complexity of temporal level 1 in the current intra period interval, avg_CTL2 denotes average complexity of temporal level 2 in the current intra period interval.
[0166] Clause 22. The method of clause 21, wherein a training dataset for training the multi-stage logistic regression classifier is obtained by: performing multi-channels transcoding on at least one sequence on a single machine with the same number of concurrent channels; evaluating real-time performance of at least one intra period interval under the corresponding number of concurrent channels to obtain at least one maximum channel under a real-time condition; and obtaining ground truth of at least one complexity level by classifying the at least one maximum channel by corresponding task.
[0167] Clause 23. The method of any of clauses 1 to 22, wherein obtaining at least one parameter associated with the conversion based on the complexity information comprises: reducing the at least one parameter based on a complexity level indicated by the complexity information, the at least one parameter comprising at least one of: the number of coding mode searches, or the number of rate-distortion optimization, and / or increasing the number of aggressive fast algorithms based on the complexity level indicated by the complexity information.
[0168] Clause 24. The method of any of clauses 1 to 22, wherein obtaining at least one parameter associated with the conversion based on the complexity information comprises: disabling at least one coding tool based on a complexity level indicated by the complexity information.
[0169] Clause 25. The method of clause 24, wherein at least one sequence parameter set (SPS) syntax in the bitstream indicates the disabling of the at least one coding tool.
[0170] Clause 26. The method of clause 24, wherein the at least one coding tool is disabled for a next I frame.
[0171] Clause 27. The method of any of clauses 1 to 22, wherein a set of parameters are obtained for a set of complexity levels by at least one of: turning off at least one coding tool, adjusting a partition depth, adjusting the number of rate-distortion optimization, or adjusting a threshold of a fast algorithm.
[0172] Clause 28. The method of any of clauses 1 to 22, wherein obtaining at least one parameter associated with the conversion based on the complexity information comprises: in response to a transcoding process satisfying a real-time condition, keeping the at least one parameter unchanged; and in response to the transcoding process dissatisfying the real-time condition, adjusting the at least one parameter based on the complexity information.
[0173] Clause 29. The method of any of clauses 1-28, wherein the conversion comprises encoding the current video unit into the bitstream.
[0174] Clause 30. The method of any of clauses 1-28, wherein the conversion comprises decoding the current video unit from the bitstream.
[0175] Clause 31. An apparatus for video processing comprising a processor and a non-transitory memory with instructions thereon, wherein the instructions upon execution by the processor, cause the processor to perform a method in accordance with any of clauses 1-30.
[0176] Clause 32. A non-transitory computer-readable storage medium storing instructions that cause a processor to perform a method in accordance with any of clauses 1-30.
[0177] Clause 33. A non-transitory computer-readable recording medium storing a bitstream of a video which is generated by a method performed by an apparatus for video processing, wherein the method comprises: determining complexity information associated with generating of the bitstream of the video from a current video unit of the video; obtaining at least one parameter for generating the bitstream based on the complexity information; and generating the bitstream based on the at least one parameter.
[0178] Clause 34. A method for storing a bitstream of a video, comprising: determining complexity information associated with generating of the bitstream of the video from a current video unit of the video; obtaining at least one parameter for generating the bitstream based on the complexity information; generating the bitstream based on the at least one parameter; and storing the bitstream in a non-transitory computer-readable recording medium. Example Device
[0179] Fig. 10 illustrates a block diagram of a computing device 1000 in which various embodiments of the present disclosure can be implemented. The computing device 1000 may be implemented as or included in the source device 110 (or the video encoder 114 or 200) or the destination device 120 (or the video decoder 124 or 300) .
[0180] It would be appreciated that the computing device 1000 shown in Fig. 10 is merely for purpose of illustration, without suggesting any limitation to the functions and scopes of the embodiments of the present disclosure in any manner.
[0181] As shown in Fig. 10, the computing device 1000 includes a general-purpose computing device 1000. The computing device 1000 may at least comprise one or more processors or processing units 1010, a memory 1020, a storage unit 1030, one or more communication units 1040, one or more input devices 1050, and one or more output devices 1060.
[0182] In some embodiments, the computing device 1000 may be implemented as any user terminal or server terminal having the computing capability. The server terminal may be a server, a large-scale computing device or the like that is provided by a service provider. The user terminal may for example be any type of mobile terminal, fixed terminal, or portable terminal, including a mobile phone, station, unit, device, multimedia computer, multimedia tablet, Internet node, communicator, desktop computer, laptop computer, notebook computer, netbook computer, tablet computer, personal communication system (PCS) device, personal navigation device, personal digital assistant (PDA) , audio / video player, digital camera / video camera, positioning device, television receiver, radio broadcast receiver, E-book device, gaming device, or any combination thereof, including the accessories and peripherals of these devices, or any combination thereof. It would be contemplated that the computing device 1000 can support any type of interface to a user (such as “wearable” circuitry and the like) .
[0183] The processing unit 1010 may be a physical or virtual processor and can implement various processes based on programs stored in the memory 1020. In a multi-processor system, multiple processing units execute computer executable instructions in parallel so as to improve the parallel processing capability of the computing device 1000. The processing unit 1010 may also be referred to as a central processing unit (CPU) , a microprocessor, a controller or a microcontroller.
[0184] The computing device 1000 typically includes various computer storage medium. Such medium can be any medium accessible by the computing device 1000, including, but not limited to, volatile and non-volatile medium, or detachable and non-detachable medium. The memory 1020 can be a volatile memory (for example, a register, cache, Random Access Memory (RAM) ) , a non-volatile memory (such as a Read-Only Memory (ROM) , Electrically Erasable Programmable Read-Only Memory (EEPROM) , or a flash memory) , or any combination thereof. The storage unit 1030 may be any detachable or non-detachable medium and may include a machine-readable medium such as a memory, flash memory drive, magnetic disk or another other media, which can be used for storing information and / or data and can be accessed in the computing device 1000.
[0185] The computing device 1000 may further include additional detachable / non-detachable, volatile / non-volatile memory medium. Although not shown in Fig. 10, it is possible to provide a magnetic disk drive for reading from and / or writing into a detachable and non-volatile magnetic disk and an optical disk drive for reading from and / or writing into a detachable non-volatile optical disk. In such cases, each drive may be connected to a bus (not shown) via one or more data medium interfaces.
[0186] The communication unit 1040 communicates with a further computing device via the communication medium. In addition, the functions of the components in the computing device 1000 can be implemented by a single computing cluster or multiple computing machines that can communicate via communication connections. Therefore, the computing device 1000 can operate in a networked environment using a logical connection with one or more other servers, networked personal computers (PCs) or further general network nodes.
[0187] The input device 1050 may be one or more of a variety of input devices, such as a mouse, keyboard, tracking ball, voice-input device, and the like. The output device 1060 may be one or more of a variety of output devices, such as a display, loudspeaker, printer, and the like. By means of the communication unit 1040, the computing device 1000 can further communicate with one or more external devices (not shown) such as the storage devices and display device, with one or more devices enabling the user to interact with the computing device 1000, or any devices (such as a network card, a modem and the like) enabling the computing device 1000 to communicate with one or more other computing devices, if required. Such communication can be performed via input / output (I / O) interfaces (not shown) .
[0188] In some embodiments, instead of being integrated in a single device, some or all components of the computing device 1000 may also be arranged in cloud computing architecture. In the cloud computing architecture, the components may be provided remotely and work together to implement the functionalities described in the present disclosure. In some embodiments, cloud computing provides computing, software, data access and storage service, which will not require end users to be aware of the physical locations or configurations of the systems or hardware providing these services. In various embodiments, the cloud computing provides the services via a wide area network (such as Internet) using suitable protocols. For example, a cloud computing provider provides applications over the wide area network, which can be accessed through a web browser or any other computing components. The software or components of the cloud computing architecture and corresponding data may be stored on a server at a remote position. The computing resources in the cloud computing environment may be merged or distributed at locations in a remote data center. Cloud computing infrastructures may provide the services through a shared data center, though they behave as a single access point for the users. Therefore, the cloud computing architectures may be used to provide the components and functionalities described herein from a service provider at a remote location. Alternatively, they may be provided from a conventional server or installed directly or otherwise on a client device.
[0189] The computing device 1000 may be used to implement video encoding / decoding in embodiments of the present disclosure. The memory 1020 may include one or more video coding modules 1025 having one or more program instructions. These modules are accessible and executable by the processing unit 1010 to perform the functionalities of the various embodiments described herein.
[0190] In the example embodiments of performing video encoding, the input device 1050 may receive video data as an input 1070 to be encoded. The video data may be processed, for example, by the video coding module 1025, to generate an encoded bitstream. The encoded bitstream may be provided via the output device 1060 as an output 1080.
[0191] In the example embodiments of performing video decoding, the input device 1050 may receive an encoded bitstream as the input 1070. The encoded bitstream may be processed, for example, by the video coding module 1025, to generate decoded video data. The decoded video data may be provided via the output device 1060 as the output 1080.
[0192] While this disclosure has been particularly shown and described with references to example embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the present application as defined by the appended claims. Such variations are intended to be covered by the scope of this present application. As such, the foregoing description of embodiments of the present application is not intended to be limiting.
Claims
1.A method for video processing, comprising:determining, for a conversion between a current video unit of a video and a bitstream of the video, complexity information associated with the conversion;obtaining at least one parameter associated with the conversion based on the complexity information; andperforming the conversion based on the at least one parameter.2.The method of claim 1, wherein the conversion comprises a transcoding process, the transcoding process comprises: obtaining the current video unit from a first bitstream of the video, and obtaining a second bitstream based on the current video unit.3.The method of claim 1 or 2, further comprising:obtaining at least one frame window of the video, a frame window comprising a set of frames within a time period;obtaining at least one encoding time of the at least one frame window;obtaining at least one decoding time stamp information of the at least one frame window;determining at least one speed factor for the at least one frame window based on the at least one encoding time and the at least one decoding time stamp information; anddetermining real-time information based on the at least one speed factor, the real-time information indicating whether coding the at least one frame window is real-time.4.The method of claim 3, wherein for a first frame window comprising (N+1) th frame to Mth frame, N and M being integers, M being greater than N, a respective speed factor of the first frame window is determined by: wherein speed_factor denotes the speed factor, TM denotes a completion time for encoding the Mth frame, TN denotes a completion time for encoding the Nth frame, dtsM denotes a decoding time stamp of the Mth frame, and dtsN denotes a decoding time stamp of the Nth frame.5.The method of claim 4, wherein a size of (M-N) is equal to a frame rate of the video times 2.6.The method of claim 4, wherein a queue records encoding time information of frames in the first frame window, TM is stored in a head of the queue, and TN is stored in a tail of the queue.7.The method of claim 3, wherein the at least one speed factor is determined by at least one expected coding time divided by the at least one encoding time.8.The method of claim 7, wherein for a first frame window comprising (N+1) th frame to Mth frame, N and M being integers, M being greater than N, a respective expected coding time is determined by one of:(M -N) / frame_rate, ordtsM -dtsN,wherein frame_rate denotes a frame rate of the first frame window, dtsM denotes a decoding time stamp of the Mth frame, and dtsN denotes a decoding time stamp of the Nth frame.9.The method of any of claims 3 to 8, wherein in response to a first speed factor of a first frame window being greater than or equal to 1, coding of the first frame window is real time, andin response to a second speed factor of a second frame window being less than 1, coding of the second frame window is non-real-time.10.The method of any of claims 3 to 8, further comprising:based on a plurality of speed factors of a plurality of consecutive frame windows being less than 1, determining that a transcoding process of the video is non-real-time.11.The method of claim 10, wherein a number of the plurality of consecutive frame windows is 25.12.The method of any of claims 3 to 11, wherein determining the complexity information comprises:in response to a transcoding process of the video is non-real-time, determining the complexity information based on at least one model.13.The method of any of claims 1 to 12, wherein determining the complexity information comprises:evaluating, with a single frame complexity model, a plurality of complexity levels of a plurality of temporal level frames of the video within an intra period interval; anddetermining, with an intra period interval complexity model, a final intra period interval complexity based on the plurality of complexity levels.14.The method of claim 13, wherein the plurality of temporal level frames comprises an I frame, temporal level 0 frame, temporal level 1 frame and temporal level 2 frame in the case of group of pictures (GOP) 4.15.The method of claim 13, wherein a complexity level of a current frame in the plurality of temporal level frames is determined by: Cf = F (Xsf) , wherein F is a linear function and Xsf is an input vector corresponding to the temporal level frame, wherein Xsf =[cu_num, avg_depth, avg_qp, cbf0_num, skip_num, merge_num, amvp_num, intra_num, bits] , andwherein:cu_num denotes a total number of coding units (CUs) in the current frame,avg_depth denotes an average partition depth in the current frame,avg_qp denotes an average quantization parameter (QP) in the current frame,cbf0_num denotes the number of CUs whose cbf is equal to 0 in the current frame,skip_num denotes the number of skip modes selected by CU in the current frame,merge_num denotes the number of merge modes selected by a CU in the current frame,amvp_num denotes the number of amvp modes selected by a CU in the current frame,intra_num denotes the number of intra modes selected by a CU in the current frame, andbits denotes the number of bits required to encode the current frame.16.The method of claim 15, wherein for an I slice, skip_num, merge_num, amvp_num and intra_num are equal to 0.17.The method of claim 15, wherein the input vector is obtained from a decoder.18.The method of claim 15, wherein coefficients of the linear function are obtained by a linear regression based on a linear regression dataset, the linear regression dataset comprising encoding time information of at least one single frame in a single-threaded encoding task.19.The method of claim 18, wherein the coefficients of the linear function are trained for frames from a plurality of temporal layers separately.20.The method of claim 13, wherein the intra period interval complexity model comprises a multi-stage logistic regression classifier used to classify intra period interval complexity into a set of levels, the number of the set of levels being an integer power of 2, andwherein obtaining at least one parameter associated with the conversion based on the complexity information comprises determining the at least one parameter based on a level of the intra period interval complexity.21.The method of claim 20, wherein the multi-stage logistic regression classifier is trained separately for a plurality of temporal layers, and an input vector for the multi-stage logistic regression classifier comprises [CI, avg_CTL0, avg_CTL1, avg_CTL2] ,wherein:CI denotes complexity of I frame in the current intra period interval,avg_CTL0 denotes average complexity of temporal level 0 in the current intra period interval,avg_CTL1 denotes average complexity of temporal level 1 in the current intra period interval,avg_CTL2 denotes average complexity of temporal level 2 in the current intra period interval.22.The method of claim 21, wherein a training dataset for training the multi-stage logistic regression classifier is obtained by:performing multi-channels transcoding on at least one sequence on a single machine with the same number of concurrent channels;evaluating real-time performance of at least one intra period interval under the corresponding number of concurrent channels to obtain at least one maximum channel under a real-time condition; andobtaining ground truth of at least one complexity level by classifying the at least one maximum channel by corresponding task.23.The method of any of claims 1 to 22, wherein obtaining at least one parameter associated with the conversion based on the complexity information comprises:reducing the at least one parameter based on a complexity level indicated by the complexity information, the at least one parameter comprising at least one of: the number of coding mode searches, or the number of rate-distortion optimization, and / orincreasing the number of aggressive fast algorithms based on the complexity level indicated by the complexity information.24.The method of any of claims 1 to 22, wherein obtaining at least one parameter associated with the conversion based on the complexity information comprises:disabling at least one coding tool based on a complexity level indicated by the complexity information.25.The method of claim 24, wherein at least one sequence parameter set (SPS) syntax in the bitstream indicates the disabling of the at least one coding tool.26.The method of claim 24, wherein the at least one coding tool is disabled for a next I frame.27.The method of any of claims 1 to 22, wherein a set of parameters are obtained for a set of complexity levels by at least one of: turning off at least one coding tool, adjusting a partition depth, adjusting the number of rate-distortion optimization, or adjusting a threshold of a fast algorithm.28.The method of any of claims 1 to 22, wherein obtaining at least one parameter associated with the conversion based on the complexity information comprises:in response to a transcoding process satisfying a real-time condition, keeping the at least one parameter unchanged; andin response to the transcoding process dissatisfying the real-time condition, adjusting the at least one parameter based on the complexity information.29.The method of any of claims 1-28, wherein the conversion comprises encoding the current video unit into the bitstream.30.The method of any of claims 1-28, wherein the conversion comprises decoding the current video unit from the bitstream.31.An apparatus for video processing comprising a processor and a non-transitory memory with instructions thereon, wherein the instructions upon execution by the processor, cause the processor to perform a method in accordance with any of claims 1-30.32.A non-transitory computer-readable storage medium storing instructions that cause a processor to perform a method in accordance with any of claims 1-30.33.A non-transitory computer-readable recording medium storing a bitstream of a video which is generated by a method performed by an apparatus for video processing, wherein the method comprises:determining complexity information associated with generating of the bitstream of the video from a current video unit of the video;obtaining at least one parameter for generating the bitstream based on the complexity information; andgenerating the bitstream based on the at least one parameter.34.A method for storing a bitstream of a video, comprising:determining complexity information associated with generating of the bitstream of the video from a current video unit of the video;obtaining at least one parameter for generating the bitstream based on the complexity information;generating the bitstream based on the at least one parameter; andstoring the bitstream in a non-transitory computer-readable recording medium.
Citation Information
Patent Citations
Content classification for multimedia processing
EP1921866A2
Method and system for controlling video transcoding
EP3322189A1
Encoding / decoding apparatus and method for parallel correction of in-loop pixels based on measured complexity, using video parameter
US20130215976A1
Intra rate control for video encoding based on sum of absolute transformed difference
US20140301460A1
Method and apparatus for video coding
US20220385905A1