US20220394279A1 - Method of encoding video with film grain - Google Patents
Method of encoding video with film grain Download PDFInfo
- Publication number
- US20220394279A1 US20220394279A1 US17/889,645 US202217889645A US2022394279A1 US 20220394279 A1 US20220394279 A1 US 20220394279A1 US 202217889645 A US202217889645 A US 202217889645A US 2022394279 A1 US2022394279 A1 US 2022394279A1
- Authority
- US
- United States
- Prior art keywords
- frames
- video
- film grain
- score
- frame
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/184—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being bits, e.g. of the compressed video stream
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/115—Selection of the code volume for a coding unit prior to coding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/117—Filters, e.g. for pre-processing or post-processing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/119—Adaptive subdivision aspects, e.g. subdivision of a picture into rectangular or non-rectangular coding blocks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/124—Quantisation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/136—Incoming video signal characteristics or properties
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/136—Incoming video signal characteristics or properties
- H04N19/137—Motion inside a coding unit, e.g. average field, frame or block difference
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/179—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a scene or a shot
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/593—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving spatial prediction techniques
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/70—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by syntax aspects related to video coding, e.g. related to compression standards
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/85—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/85—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression
- H04N19/86—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using pre-processing or post-processing specially adapted for video compression involving reduction of coding artifacts, e.g. of blockiness
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/14—Picture signal circuitry for video frequency region
- H04N5/144—Movement detection
Definitions
- the present invention relates to a process for improving video quality when encoding video with film grain. More particularly, the present invention relates to a solution to improving video quality when film grain is present on a level similar to the “Red Lady” noise problem.
- Film grain is hard to compress in an encoder. It requires more bits to encode than many other kinds of content for any level of video quality. Film grain may be thought of as a particular kind of spatial temporal noise. As such, film grain has low temporal predictability from one frame to any other frame of video. Thus, the encoding process is limited in its ability to leverage inter-frame estimation to achieve significant compression efficiency.
- FIG. 1 A frame of the “Red Lady” video is illustrated with FIG. 1 .
- the “red lady” video shows that a lady walking alongside a soccer field with a grassy background.
- the scene itself is simple, but the entire video is filled with film grain.
- Film grain is like random noise. It requires a lot of bits to encode and is not temporally predictable, which makes high film grain video, in particular the “Red Lady” video, very difficult to encode.
- a common practice to encode video with film grain is to encode a good quality I frame as a reference frame as a prediction frame for subsequent predictive frames (P or B frames).
- P or B frames predictive frames
- the I frame and P and B frames all contain film grain this does not predict well, and many bits are needed to encode unpredicted high frequency components. If too many bits are allocated to the I frame, later P and B frames may be allocated fewer bits than they need, and their quality suffers.
- the good quality I frame thus, may not help with the subsequent P and B frames.
- the difference between two consecutive frames is mostly noise. Encoding a good quality I frame for these costs too many bits and leaves fewer bits for predictive frames. Moreover, the high quality I frame, even with additional bits allocated, is not a good reference frame because the noisy temporal differences cannot be motion predicted well. With the I frame as a reference, the film grains in the predictive frames would be poorly encoded, and create a quality disparity between I and predictive frames, as illustrated in FIG. 3 . Thus allocating more bits to the I frame creates a “Dirty Window” for future film grain elimination in P and B frames.
- Embodiments of the present invention provide a system that enables improved video quality and compression efficiency during encoding by detecting video segments having film grain approaching the “Red Lady” problem and then optimizing the bit allocation between intra- and inter-predicted pictures using bit allocation variation between I, P and B type frames.
- embodiments of the present invention encode smaller I frames and allocate more bits on P and B frames. Since allocating more bits to the I frame when the “Red Lady” film grain problem occurs does not enable better prediction for encoding in the P and B frames, additional bits to the I frame are not necessary. Thus, allocating extra bits allocated to the P and B frames and not using the extra bits in the I frame enables reduction of frame grain when the “Red Lady” like film grain problem occurs, and the “Dirty Window” I frame issue will no longer be a consideration.
- a temporal analysis of motion-prediction data available is provided.
- measurements of plotted frame-by-frame temporal differences (ME scores) of the received videos are determined.
- ME scores plotted frame-by-frame temporal differences
- two key indicators are identified: (1) The average temporal difference in frames with an intermediate motion level (i.e., greater than ME score of 20), is higher than frames of non-noisy video with intermediate motion; and (2) The fluctuation of the temporal differences between frames in a group is very small, unlike the non-noise video with natural motions which have higher motion differences without the noise.
- These two indicators are set to identify when a special bit allocation ratio between the I, P and B frames should be applied so that there will tend to be less difference between frame types for film-grain content.
- the system uses a preprocessing filter that analyzes video frames prior to the encoder.
- the preprocessing filter computes the temporal difference score and stores it in a queue of data provided with the frames to the encoder.
- the encoder analyzes the temporal difference scores. If it is detected that the average of all temporal differences is higher than a threshold and the variance of them are smaller than a threshold, it means the video contains significant film grain or noise. Based on the level of film grain or noise detected, the encoder allocates bits to I, P and B frames dynamically.
- FIG. 1 shows a frame from the “Red Lady” video
- FIG. 2 illustrates that for the “Red Lady” video frames, the difference between two consecutive frames is mostly noise
- FIG. 3 illustrates that with an I frame as a reference, the film grains in the predictive frames would be poorly encoded, and create a quality disparity between I and predictive frames;
- FIGS. 4 A- 4 F show the measured and plotted frame-by-frame temporal differences (ME scores) of various video clips
- FIG. 5 is a diagram of components for implementing embodiments of the present invention in an encoding system.
- FIG. 6 is a flow chart illustrating steps for implementing encoding with film grain according to embodiments of the present invention.
- a clip can be identified as a “Red Lady” like clip, good quality can be achieved by encoding smaller I frames and allocating more bits on P and B frames.
- FIGS. 4 A- 4 F show the measured and plotted frame-by-frame temporal differences (ME scores) of various video clips.
- the X axis is a frame index and the Y axis shows a scaled ME score ranging from 0 to 100.
- FIGS. 4 A- 4 C illustrate the ME score levels for video with noise and limited or no motion.
- the Red Lady video has a constant noise level ME score of just over 20.
- the “Sweep” video noise levels of FIG. 4 B are very high with an ME noise level of near 100.
- the “Zone Plate” noise with no motion and a set noise level has an ME score of just over 20 and can be used as a reference for ME levels. Note that in the “Blacksmith” frame of FIG. 4 D , a normal video that does not need special treatment during encoding using embodiments of the present invention that the temporal average noise difference is far below an ME level of 20 found with the “Red Lady” video of FIG. 4 C .
- the “Sprinkler Lady” of FIG. 4 E meets both the two key factors of (1) an ME score over 20 and (2) the fluctuation of the temporal differences is small even taking into account motion.
- the final video of “Basketball” in FIG. 4 F has a relatively high ME score, but it is less than 20, and the motion in the video accounts for ME levels that on average may push the total ME score over 20.
- the video of FIG. 4 F does not require embodiments of the present invention to be used during encoding.
- FIG. 5 is a diagram of one embodiment of components for implementing embodiments of the present invention in an encoding system.
- the preprocessing filter 500 computes the temporal difference score and stores it in the queue 502 .
- the encoding in encoder 506 will be delayed by the frame buffer 504 until temporal difference scores of N frames are collected in the queue 502 .
- the encoder 506 analyzes the temporal difference scores of N frames. If it is detected that the average of all temporal differences is higher than a threshold and the variance of them are smaller than a threshold, it means the video contains significant film grain or noise. Based on the level of film grain or noise detected, the encoder 506 allocates bits to I, P and B dynamically according to embodiments of the present invention described herein. Generally, if the level of film grain or noise is high, the encoder allocates more bits to P and B frames than for other content.
- FIG. 6 is a flow chart illustrating steps for implementing encoding when high film grain is detected according to embodiments of the present invention.
- the video clips are received, such as the “Red Lady” clip illustrated in FIG. 4 C .
- the temporal difference score or ME score
- ME score is determined for the video frames and the result for each frame stored in a queue to provide to the encoder.
- a determination is made to decide if film grain noise is high enough to constitute “Red Lady” type film grain that requires application of embodiments of the present invention. For the step 602 determination, if the average of a group of temporal differences is higher than a threshold and the variance is smaller than a threshold, the film grain noise is indicated to be significant for the frames of the video clip.
- step 603 the determination is reviewed. If film grain noise for the clip is determined to be significant, then the program proceeds to step 604 . If the film grain noise is determined to be insignificant, the program proceeds to step 605 . In step 604 when high film grain noise is detected encoding is performed by allocating enough bits so that the I frame at the beginning has little additional bits and the P and B frames have additional bits for encoding. In step 605 when film grain noise is not detected as high, a normal bit allocation is performed by the encoder.
- the results of applying the algorithm shown in FIG. 6 will reduce the “dirty window” effect.
- the algorithm also makes some high texture clips, such as “sprinkler” of FIG. 4 E , look sharper.
- the algorithm does not change the quality of other non-noisy clips that do not rise to the detected level of the “Red Lady” video.
- Appendix A shows an example of coding in “C” to implement the algorithm illustrated by FIG. 6 .
- each component can include a processor and memory to enable operation.
- the memory of each device stores code that is executable by the processor to enable the processor to perform the processes described herein. Further the memory can be used to provide data storage with the data accessible by the processor to store or retrieve when performing operations.
Abstract
Description
- This application is a continuation of U.S. patent application Ser. No. 14/962,814 filed Dec. 8, 2015, which claims priority under 35 U.S.C. § 119(e) from earlier filed U.S. Provisional Application Ser. No. 62/099,372 filed on Jan. 5, 2015 and incorporated herein by reference in its entirety.
- The present invention relates to a process for improving video quality when encoding video with film grain. More particularly, the present invention relates to a solution to improving video quality when film grain is present on a level similar to the “Red Lady” noise problem.
- Film grain is hard to compress in an encoder. It requires more bits to encode than many other kinds of content for any level of video quality. Film grain may be thought of as a particular kind of spatial temporal noise. As such, film grain has low temporal predictability from one frame to any other frame of video. Thus, the encoding process is limited in its ability to leverage inter-frame estimation to achieve significant compression efficiency.
- In some encoders, not enough bits are allocated to inter-predicted pictures. The result can be significant video quality artifacts such as I-frame beating and intermittent repetitive loss and recovery of spatial texture. Even encoders that can allocate significant bits might not be able to eliminate substantial noise, such as when noise is as high as in the “Red Lady” video frames.
- A frame of the “Red Lady” video is illustrated with
FIG. 1 . The “red lady” video shows that a lady walking alongside a soccer field with a grassy background. The scene itself is simple, but the entire video is filled with film grain. - Film grain is like random noise. It requires a lot of bits to encode and is not temporally predictable, which makes high film grain video, in particular the “Red Lady” video, very difficult to encode.
- A common practice to encode video with film grain is to encode a good quality I frame as a reference frame as a prediction frame for subsequent predictive frames (P or B frames). However, since the I frame and P and B frames all contain film grain this does not predict well, and many bits are needed to encode unpredicted high frequency components. If too many bits are allocated to the I frame, later P and B frames may be allocated fewer bits than they need, and their quality suffers. The good quality I frame, thus, may not help with the subsequent P and B frames.
- In the Red Lady Video, the random noise level is very high. Thus, in the Red Lady Video, a beating effect will be seen, due to the quality of the pictures varying too much between frame types. Allocating more bits to the I frame does not help with reducing film grain in subsequent P and B frames in a typical encoder.
- As shown in
FIG. 2 , the difference between two consecutive frames is mostly noise. Encoding a good quality I frame for these costs too many bits and leaves fewer bits for predictive frames. Moreover, the high quality I frame, even with additional bits allocated, is not a good reference frame because the noisy temporal differences cannot be motion predicted well. With the I frame as a reference, the film grains in the predictive frames would be poorly encoded, and create a quality disparity between I and predictive frames, as illustrated inFIG. 3 . Thus allocating more bits to the I frame creates a “Dirty Window” for future film grain elimination in P and B frames. - Accordingly, it is desirable to provide better solutions for eliminating film grain comparable to the “Red Lady” video, and to avoid creating “Dirty Window” I frames.
- Embodiments of the present invention provide a system that enables improved video quality and compression efficiency during encoding by detecting video segments having film grain approaching the “Red Lady” problem and then optimizing the bit allocation between intra- and inter-predicted pictures using bit allocation variation between I, P and B type frames.
- To optimize the bit allocation when a video clip is identified as a “Red Lady” like clip, embodiments of the present invention encode smaller I frames and allocate more bits on P and B frames. Since allocating more bits to the I frame when the “Red Lady” film grain problem occurs does not enable better prediction for encoding in the P and B frames, additional bits to the I frame are not necessary. Thus, allocating extra bits allocated to the P and B frames and not using the extra bits in the I frame enables reduction of frame grain when the “Red Lady” like film grain problem occurs, and the “Dirty Window” I frame issue will no longer be a consideration.
- To identify the film grain level to determine when the optimization of bit allocation away from I frames to P and B frames should occur, a temporal analysis of motion-prediction data available is provided. For the temporal analysis, measurements of plotted frame-by-frame temporal differences (ME scores) of the received videos are determined. From the ME scores, two key indicators are identified: (1) The average temporal difference in frames with an intermediate motion level (i.e., greater than ME score of 20), is higher than frames of non-noisy video with intermediate motion; and (2) The fluctuation of the temporal differences between frames in a group is very small, unlike the non-noise video with natural motions which have higher motion differences without the noise. These two indicators are set to identify when a special bit allocation ratio between the I, P and B frames should be applied so that there will tend to be less difference between frame types for film-grain content.
- The system according to embodiments of the present invention uses a preprocessing filter that analyzes video frames prior to the encoder. The preprocessing filter computes the temporal difference score and stores it in a queue of data provided with the frames to the encoder. The encoder analyzes the temporal difference scores. If it is detected that the average of all temporal differences is higher than a threshold and the variance of them are smaller than a threshold, it means the video contains significant film grain or noise. Based on the level of film grain or noise detected, the encoder allocates bits to I, P and B frames dynamically.
- Further details of the present invention are explained with the help of the attached drawings in which:
-
FIG. 1 shows a frame from the “Red Lady” video; -
FIG. 2 illustrates that for the “Red Lady” video frames, the difference between two consecutive frames is mostly noise; -
FIG. 3 illustrates that with an I frame as a reference, the film grains in the predictive frames would be poorly encoded, and create a quality disparity between I and predictive frames; -
FIGS. 4A-4F show the measured and plotted frame-by-frame temporal differences (ME scores) of various video clips; -
FIG. 5 is a diagram of components for implementing embodiments of the present invention in an encoding system; and -
FIG. 6 is a flow chart illustrating steps for implementing encoding with film grain according to embodiments of the present invention. - For embodiments of the present invention, if a clip can be identified as a “Red Lady” like clip, good quality can be achieved by encoding smaller I frames and allocating more bits on P and B frames.
- To help understand how a determination of when a clip is a “Red Lady” type clip, several different clips of data are analyzed.
FIGS. 4A-4F show the measured and plotted frame-by-frame temporal differences (ME scores) of various video clips. InFIGS. 4A-4F , the X axis is a frame index and the Y axis shows a scaled ME score ranging from 0 to 100. - From the measurements in
FIGS. 4A-4F , two key indicators of noisy video are found: (1) The average temporal difference is at an intermediate level (greater than ME score of 20), and the ME score is higher than the one of non-noisy video with intermediate motion; and (2) The fluctuation of the temporal differences is very small, not like the non-noise video with natural motions. These two indicators are, thus, used to identify “Red Lady” like video frames from any streaming video. -
FIGS. 4A-4C illustrate the ME score levels for video with noise and limited or no motion. InFIG. 4C , the Red Lady video has a constant noise level ME score of just over 20. The “Sweep” video noise levels ofFIG. 4B are very high with an ME noise level of near 100. The “Zone Plate” noise with no motion and a set noise level has an ME score of just over 20 and can be used as a reference for ME levels. Note that in the “Blacksmith” frame ofFIG. 4D , a normal video that does not need special treatment during encoding using embodiments of the present invention that the temporal average noise difference is far below an ME level of 20 found with the “Red Lady” video ofFIG. 4C . - Unlike the “Blacksmith” of
FIG. 4D , the “Sprinkler Lady” ofFIG. 4E meets both the two key factors of (1) an ME score over 20 and (2) the fluctuation of the temporal differences is small even taking into account motion. The final video of “Basketball” inFIG. 4F has a relatively high ME score, but it is less than 20, and the motion in the video accounts for ME levels that on average may push the total ME score over 20. However with the criteria of (1) an average temporal difference with a ME score greater than 20 and (2) the fluctuation of the temporal differences very small, the video ofFIG. 4F does not require embodiments of the present invention to be used during encoding. -
FIG. 5 is a diagram of one embodiment of components for implementing embodiments of the present invention in an encoding system. InFIG. 5 , thepreprocessing filter 500 computes the temporal difference score and stores it in thequeue 502. The encoding inencoder 506 will be delayed by theframe buffer 504 until temporal difference scores of N frames are collected in thequeue 502. Theencoder 506 analyzes the temporal difference scores of N frames. If it is detected that the average of all temporal differences is higher than a threshold and the variance of them are smaller than a threshold, it means the video contains significant film grain or noise. Based on the level of film grain or noise detected, theencoder 506 allocates bits to I, P and B dynamically according to embodiments of the present invention described herein. Generally, if the level of film grain or noise is high, the encoder allocates more bits to P and B frames than for other content. -
FIG. 6 is a flow chart illustrating steps for implementing encoding when high film grain is detected according to embodiments of the present invention. First, instep 600 the video clips are received, such as the “Red Lady” clip illustrated inFIG. 4C . Next, instep 601, the temporal difference score, or ME score, is determined for the video frames and the result for each frame stored in a queue to provide to the encoder. Next, in step 602 a determination is made to decide if film grain noise is high enough to constitute “Red Lady” type film grain that requires application of embodiments of the present invention. For thestep 602 determination, if the average of a group of temporal differences is higher than a threshold and the variance is smaller than a threshold, the film grain noise is indicated to be significant for the frames of the video clip. - Once the determination is made in 602, next in
step 603 the determination is reviewed. If film grain noise for the clip is determined to be significant, then the program proceeds to step 604. If the film grain noise is determined to be insignificant, the program proceeds to step 605. Instep 604 when high film grain noise is detected encoding is performed by allocating enough bits so that the I frame at the beginning has little additional bits and the P and B frames have additional bits for encoding. Instep 605 when film grain noise is not detected as high, a normal bit allocation is performed by the encoder. - The results of applying the algorithm shown in
FIG. 6 will reduce the “dirty window” effect. The algorithm also makes some high texture clips, such as “sprinkler” ofFIG. 4E , look sharper. The algorithm does not change the quality of other non-noisy clips that do not rise to the detected level of the “Red Lady” video. - For reference, Appendix A below shows an example of coding in “C” to implement the algorithm illustrated by
FIG. 6 . - For components shown, like the
pre-processing filter 500 and theencoder 506, each component according to embodiments of the present invention can include a processor and memory to enable operation. The memory of each device stores code that is executable by the processor to enable the processor to perform the processes described herein. Further the memory can be used to provide data storage with the data accessible by the processor to store or retrieve when performing operations. - Although the present invention has been described above with particularity, this was merely to teach one of ordinary skill in the art how to make and use the invention. Many additional modifications will fall within the scope of the invention as that scope is defined by the following claims.
-
APPENDIX A // // code for Red Lady film grain detection #define RC_DYNAMIC_QMUL_ME_SCORE_MEAN_THRESHOLD_MIN 20 #define RC_DYNAMIC_QMUL_ME_SCORE_MEAN_THRESHOLD_MAX 75 #define RC_DYNAMIC_QMUL_ME_SCORE_VARIANCE_THRESHOLD 2 uint8_t get_qmul_b_increase_from_lookahead ( hlenc_fbnode_t* fbnode, uint8_t first_pass_enc_id) { img_par_t *imgpar = get_imgpar_by_fbnode(fbnode); uint32_t vfid = GET_DB_INDEX( fbnode->ext_vcap_pi ); uint32_t offset_tf = 0, offset_bf = 0; scene_change_info_t scene_change_info; uint8_t scene_id_first = 0; int8_t me_score[RC_2PASS_LOOKAHEAD_DISTANCE_DYNAMIC_QMUL] = {0}; uint32_t me_score_cnt = 0; uint32_t i = 0, sum = 0, var = 0; int8_t mean = 0; uint8_t b_increase = RC_DEFAULT_B_INCREASE_FOR_QMUL; //get offset of the me_score for top and bottom field offset_tf = (uint32_t)(&(((ext_vcap_package_t *)0)->sc_info_frm_top)); offset_bf = (uint32_t)(&(((ext_vcap_package_t *)0)->sc_info_bot)); //get the first scene_info get_epi_data(MAKE_DB_ID(first_pass_enc_id, (vfid & 0xFF)), (uint32_t)&scene_change_info, offset_tf, sizeof(scene_change_info_t)); scene_id_first = scene_change_info.scene_id; //the me_score for the first frame is always big because it is calculated between the new scene and old scene // we only want to use the me_score of the new scene do { vfid++; get_epi_data(MAKE_DB_ID(first_pass_enc_id, (vfid & 0xFF)), (uint32_t)&scene_change_info, offset_tf, sizeof(scene_change_info_t)); //todo: check scene change instead of scene id? if ( scene_id_first != scene_change_info.scene_id ) //if ( scene_change_info.sc_here ) { orc_printf(″scene_id_first %d scene_change_info.scene_id %d me_score_cnt %d″,scene_id_first,scene_change_info.scene_id,me_score_cnt); break; } me_score[me_score_cnt++] = ( scene_change_info.me_score>>3 ); if ( imgpar->pic_is_field ) { //get me score for bottom field get_epi_data(MAKE_DB_ID(first_pass_enc_id, (vfid & 0xFF)), (uint32_t)&scene_change_info, offset bf, sizeof(scene_change_info_t)); me_score[me_score_cnt++] = ( scene_change_info.me_score >>3 ); } } while (me_score_cnt < RC_2PASS_LOOKAHEAD_DISTANCE_DYNAMIC_QMUL ); //get the mean and variance for (i = 0; i < me_score_cnt; i ++) { sum += me_score[i]; //orc_printf(″me_score[%d] %d″,i,me_score[i]); } mean = sum/me_score_cnt; for (i = 0; i < me_score_cnt; i ++) { var += (me score[i] - mean)*(me_score[i] - mean); } var /= me_score_cnt; //this is to check redlady like noisey video. // When the average me_score is higher than a threshold but me_scores have a very small fluctuation, it means the temporal prediction error // was primarily casued by a low level of noise not the actual natural motion. In this case, we want to spend more bits on P and B frames so we encode less skip blocks. if (mean > RC_DYNAMIC_QMUL_ME_SCORE_MEAN_THRESHOLD_MIN && mean < RC_DYNAMIC_QMUL_ME_SCORE_MEAN_THRESHOLD_MAX && var <= RC_DYNAMIC_QMUL_ME_SCORE_VARIANCE_THRESHOLD) { b_increase = 25; } //orc_printf(″ me_score_mean %d var %d b_increase %d″,mean, var, b_increase); return b_increase; }
Claims (3)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US17/889,645 US20220394279A1 (en) | 2015-01-05 | 2022-08-17 | Method of encoding video with film grain |
Applications Claiming Priority (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US201562099672P | 2015-01-05 | 2015-01-05 | |
US14/962,814 US11451798B2 (en) | 2015-01-05 | 2015-12-08 | Method of encoding video with film grain |
US17/889,645 US20220394279A1 (en) | 2015-01-05 | 2022-08-17 | Method of encoding video with film grain |
Related Parent Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/962,814 Continuation US11451798B2 (en) | 2015-01-05 | 2015-12-08 | Method of encoding video with film grain |
Publications (1)
Publication Number | Publication Date |
---|---|
US20220394279A1 true US20220394279A1 (en) | 2022-12-08 |
Family
ID=55070691
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/962,814 Active 2037-08-08 US11451798B2 (en) | 2015-01-05 | 2015-12-08 | Method of encoding video with film grain |
US17/889,645 Abandoned US20220394279A1 (en) | 2015-01-05 | 2022-08-17 | Method of encoding video with film grain |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US14/962,814 Active 2037-08-08 US11451798B2 (en) | 2015-01-05 | 2015-12-08 | Method of encoding video with film grain |
Country Status (7)
Country | Link |
---|---|
US (2) | US11451798B2 (en) |
EP (1) | EP3041236B1 (en) |
BR (1) | BR112017014421B1 (en) |
CA (1) | CA2972989C (en) |
GB (1) | GB2533858B (en) |
MX (1) | MX368521B (en) |
WO (1) | WO2016111799A1 (en) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US10454987B2 (en) | 2016-10-28 | 2019-10-22 | Google Llc | Bitrate optimization for multi-representation encoding using playback statistics |
FR3073999B1 (en) | 2017-11-23 | 2024-02-09 | Ateme | INTELLIGENT COMPRESSION OF SEEDED VIDEO CONTENT |
US11930222B1 (en) * | 2021-03-30 | 2024-03-12 | Amazon Technologies, Inc. | Encoding video for film grain synthesis |
CA3156314A1 (en) * | 2021-04-19 | 2022-10-19 | Comcast Cable Communications, Llc | Methods, systems, and apparatuses for adaptive processing of video content with film grain |
US20230059035A1 (en) * | 2021-08-23 | 2023-02-23 | Netflix, Inc. | Efficient encoding of film grain noise |
US11622123B1 (en) * | 2021-09-02 | 2023-04-04 | Amazon Technologies, Inc. | Film grain preservation |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030053543A1 (en) * | 2001-07-24 | 2003-03-20 | Sasken Communication Technologies Limited | Motion estimation technique for digital video encoding applications |
US20040130680A1 (en) * | 2002-03-13 | 2004-07-08 | Samuel Zhou | Systems and methods for digitally re-mastering or otherwise modifying motion pictures or other image sequences data |
US20060140267A1 (en) * | 2004-12-28 | 2006-06-29 | Yong He | Method and apparatus for providing intra coding frame bit budget |
US20140112386A1 (en) * | 2012-10-22 | 2014-04-24 | General Instrument Corporation | Algorithms for determining bitrate for a statistical multiplexing system to account for signal complexity including film mode and gop structural changes |
US9813706B1 (en) * | 2013-12-02 | 2017-11-07 | Google Inc. | Video content analysis and/or processing using encoding logs |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7924922B2 (en) | 2005-03-07 | 2011-04-12 | Hewlett-Packard Development Company, L.P. | Color registration in a digital video |
US8243804B2 (en) | 2005-12-01 | 2012-08-14 | Lsi Corporation | Hierarchical motion estimation for images with varying horizontal and/or vertical dimensions |
US8213500B2 (en) | 2006-12-21 | 2012-07-03 | Sharp Laboratories Of America, Inc. | Methods and systems for processing film grain noise |
EP2048888A1 (en) | 2007-10-12 | 2009-04-15 | Thomson Licensing | Method and decoder for decoding an image frame of an image frame sequence |
-
2015
- 2015-12-08 US US14/962,814 patent/US11451798B2/en active Active
- 2015-12-09 WO PCT/US2015/064703 patent/WO2016111799A1/en active Application Filing
- 2015-12-09 BR BR112017014421-2A patent/BR112017014421B1/en active IP Right Grant
- 2015-12-09 MX MX2017008847A patent/MX368521B/en active IP Right Grant
- 2015-12-09 CA CA2972989A patent/CA2972989C/en active Active
- 2015-12-15 GB GB1522138.5A patent/GB2533858B/en active Active
- 2015-12-15 EP EP15200232.5A patent/EP3041236B1/en active Active
-
2022
- 2022-08-17 US US17/889,645 patent/US20220394279A1/en not_active Abandoned
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030053543A1 (en) * | 2001-07-24 | 2003-03-20 | Sasken Communication Technologies Limited | Motion estimation technique for digital video encoding applications |
US20040130680A1 (en) * | 2002-03-13 | 2004-07-08 | Samuel Zhou | Systems and methods for digitally re-mastering or otherwise modifying motion pictures or other image sequences data |
US20060140267A1 (en) * | 2004-12-28 | 2006-06-29 | Yong He | Method and apparatus for providing intra coding frame bit budget |
US20140112386A1 (en) * | 2012-10-22 | 2014-04-24 | General Instrument Corporation | Algorithms for determining bitrate for a statistical multiplexing system to account for signal complexity including film mode and gop structural changes |
US9813706B1 (en) * | 2013-12-02 | 2017-11-07 | Google Inc. | Video content analysis and/or processing using encoding logs |
Also Published As
Publication number | Publication date |
---|---|
MX2017008847A (en) | 2017-10-24 |
BR112017014421A8 (en) | 2023-02-07 |
BR112017014421A2 (en) | 2018-04-10 |
BR112017014421B1 (en) | 2023-05-09 |
EP3041236A1 (en) | 2016-07-06 |
US11451798B2 (en) | 2022-09-20 |
EP3041236B1 (en) | 2021-02-17 |
GB2533858A (en) | 2016-07-06 |
US20160198165A1 (en) | 2016-07-07 |
WO2016111799A1 (en) | 2016-07-14 |
CA2972989C (en) | 2019-11-12 |
GB201522138D0 (en) | 2016-01-27 |
MX368521B (en) | 2019-10-07 |
CA2972989A1 (en) | 2016-07-14 |
GB2533858B (en) | 2017-08-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20220394279A1 (en) | Method of encoding video with film grain | |
US9426475B2 (en) | Scene change detection using sum of variance and estimated picture encoding cost | |
US10225550B2 (en) | Controlling a video content system by computing a frame quality score | |
US9118912B2 (en) | Object-aware video encoding strategies | |
US9706203B2 (en) | Low latency video encoder | |
US9565440B2 (en) | Quantization parameter adjustment based on sum of variance and estimated picture encoding cost | |
US8452120B2 (en) | Image processing device and method | |
US20120288204A1 (en) | Image processing apparatus and image processing method | |
US9955158B2 (en) | Video encoder with adjustable intra-refresh rate | |
JP4366571B2 (en) | Video encoding apparatus and method | |
US20110317912A1 (en) | Method, apparatus and computer-readable medium coding and decoding depth image using color image | |
US11321390B2 (en) | Video data generation apparatus, video data generation method, and program | |
US20150163484A1 (en) | Variable bitrate encoding for multiple video streams | |
US20140177701A1 (en) | System and method for time budget achievement in real-time video encoding | |
KR20120019662A (en) | Pre-processing method of a moving picture to be transmitted and system transmitting a moving picture using thereof | |
US10136134B2 (en) | Method to encode video with close multiple scene changes | |
US20240121414A1 (en) | Hybrid inter-frame coding using an autoregressive model | |
KR20130121913A (en) | Method and apparatus for measuring quality of video based on frame loss pattern | |
KR102049644B1 (en) | Hierarchical Noise Reduction Method for Efficient Coding of Video with Noise | |
US20170061216A1 (en) | Event sensing method and system using dct coefficient in surveillance camera |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AS | Assignment |
Owner name: ARRIS ENTERPRISES, INC., GEORGIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:ZHOU, WEI;PERERA, JANE;MICHELSEN, WAYNE D.;AND OTHERS;SIGNING DATES FROM 20160105 TO 20160115;REEL/FRAME:060831/0825 Owner name: ARRIS ENTERPRISES LLC, GEORGIA Free format text: CHANGE OF NAME;ASSIGNOR:ARRIS ENTERPRISES, INC.;REEL/FRAME:061202/0962 Effective date: 20151221 |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NON FINAL ACTION MAILED |
|
STCB | Information on status: application discontinuation |
Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION |