search

fraunhoferhhi / vvenc

VVenC, the Fraunhofer Versatile Video Encoder
https://www.hhi.fraunhofer.de/en/departments/vca/technologies-and-solutions/h266-vvc.html
BSD 3-Clause Clear License
942 stars 170 forks source link

QPA causes significant detail loss in static parts of video #369

Open joletb opened 6 months ago

joletb commented 6 months ago

While QPA gives quite good increases in fidelity in intra coding, it can cause static parts of a video to lose all detail in inter coding, even when there is a good reference frame available.

Furthermore, if we remove the first frame from our sample below, this detail loss is remedied and the encoder properly uses info from previous/next frames, signifying a possible issue with the GOP structure when used with QPA.

This issue does not happen at all in non-QPA encoding, but then detail preservation in general is quite a bit worse.

The following command line was used for the encoding: vvencffapp -c config/slow --y4m 1 -q 20 -t -1 --PrintFrameMSE 1 --PrintSequenceMSE 1 --ALF 1 -ip 0 -rs 11 --CCALF 1 --STA 0 --WppBitEqual -1 --MaxParallelFrames -1 --SAO 0 --qpa 1 --BIM 0 --MCTF 0 --InputFile - -b slow.266

Here are a couple of comparisons: https://slow.pics/c/koHLpziy Note also that this issue is significantly more prevalent in anime as there tend to be a lot of duplicate frames.

The encoded bitstreams and the video source used for encoding can be found here: https://drive.google.com/drive/folders/1OAsQxfd6y-Xl8u1syeXkZ9u22scQD0YG?usp=sharing

jungleboynx commented 6 months ago

I encoded the source in x264@CRF21 and h266@QP28 files but couldn't see any difference when playing back the files at the same time in single step mode: https://www.dropbox.com/scl/fo/j6r4po3naxgsa5z6vk961/h?rlkey=1a0y3i488qdqt226zcssalxfb&dl=0

The video is a bit odd with alternative bright and dark frames in places. The clouds seem to move on every frame but the anime moves on alternative frames - maybe this is normal.

joletb commented 6 months ago

Sure, there's no difference... if you don't count the texture and lineart loss. Although the main issue here is about QPA making the encoder mangle data instead of reusing it, which by the way, it does in your encode as well: https://slow.pics/c/dwOWsn5s

adamjw24 commented 6 months ago

Thanks, we'll have a look.

crhelmrich commented 5 months ago

Hi joletb,

While QPA gives quite good ... fidelity in intra coding, it can cause static parts of a video to lose all detail in inter coding

Can you clarify what you mean by 'static'? The video scene features quite some motion, but 'static' to me means not moving.

This issue does not happen at all in non-QPA encoding, but then detail preservation in general is quite a bit worse.

So, is (to your eyes) the QPA version of the video still better overall than the non-QPA version when played at normal speed?

I have to admit that, when comparing the 'qpa/noqpa' bitstreams you shared (thanks), I cannot reliably spot any differences. Are your observations based on watching the video during playback, or just on inspecting isolated frames (i.e. screenshots)?

Note that VVenC's QP adaptation makes use of spatial as well as temporal psycho-visual modelling. It's based on the XPSNR visual model, details of which you can read up on in this open-access paper: http://handle.itu.int/11.1002/pub/8153d78b-en

This means that VVenC's QPA is solely intended, and optimized, for videos at typical frame rates, NOT for still image coding. Or, in other words, screenshots of individual frames don't tell the whole (perceptual) story.

That said, I'll take another look at the issue with the first frame in the video.

Christian

joletb commented 5 months ago

The "static" part here refers to the to the duplicate parts of the building in the video, there is practically no motion between them there and as such there should be no detail loss. (If we're going this route, should the detail loss be extended to the duplicate parts to make the video more consistent?) This behavior can happen on scene changes as well, where, with enough of a low bitrate, can be very noticeable.

That being said, I don't think this is entirely a QPA issue as this does not happen when you remove the first frame of the sample. (Both the original sample and this one mentioned below are perceptually the same) I've included the trimmed sample and an encode of it in the original GDrive link as CSM_test_clip_no_duplicate_first_frame_yuv420p10.y4m and slow-qpa-q20-no-duplicate.266, respectively.

birdie-github commented 5 months ago

I'm an absolute newbie in video encoding but to me this issue looks similar to #212.

L4cache commented 5 months ago

I'm an absolute newbie in video encoding but to me this issue looks similar to #212.

It's not.

ZenKiyoshi commented 4 months ago

I have noticed this before. https://github.com/fraunhoferhhi/vvenc/issues/133

https://github.com/fraunhoferhhi/vvenc/issues/132

L4cache commented 2 months ago

Maybe the problem is the opposite: the subsequent frame of the "problematic" frame is actually getting too many quality.

Tested with the y4m clip provided by joletb ,arguments used: --preset faster -ip -1 -dr idr_no_radl --poc0idr 1 --sao 0 --bim 0 --mctf 0 -t 16 --wavefrontsynchro 1 --ctusize 128 --ifp 0 --ciip 3 -qpa 1 --qp 20 -fs 1 (skip first frame makes the problem more significant in this case)

With QPA:

POC     0 TId: 0 (  IDR_N_LP, I-SLICE, QP 10)           846416 bits [Y 55.6109 dB    U 60.6117 dB    V 60.1662 dB] [ET     0 ] [L0 ] [L1 ]
POC    16 TId: 1 (      STSA, B-SLICE, QP 22)           410800 bits [Y 47.9555 dB    U 52.5970 dB    V 51.4652 dB] [ET     0 ] [L0 0 ] [L1 0 ]
POC     8 TId: 2 (      STSA, B-SLICE, QP 22)           405784 bits [Y 48.2472 dB    U 52.3015 dB    V 51.2012 dB] [ET     0 ] [L0 0 ] [L1 16 ]
POC     4 TId: 3 (      STSA, B-SLICE, QP 23)           289112 bits [Y 47.6080 dB    U 51.9036 dB    V 50.8292 dB] [ET     0 ] [L0 0 ] [L1 8 ]
POC     2 TId: 4 (      STSA, B-SLICE, QP 24)           182560 bits [Y 46.9975 dB    U 51.3339 dB    V 50.5454 dB] [ET     0 ] [L0 0 ] [L1 4 ]
POC     1 TId: 5 (      STSA, B-SLICE, QP 28)           130064 bits [Y 45.6190 dB    U 49.5797 dB    V 49.1447 dB] [ET     0 ] [L0 0 2 ] [L1 2 4 ]
POC     3 TId: 5 (      STSA, B-SLICE, QP 23)           155544 bits [Y 47.5512 dB    U 52.3980 dB    V 51.5035 dB] [ET     0 ] [L0 2 ] [L1 4 ]
POC     6 TId: 4 (      STSA, B-SLICE, QP 26)           230000 bits [Y 45.8785 dB    U 49.9708 dB    V 48.9032 dB] [ET     0 ] [L0 4 ] [L1 8 ]
POC     5 TId: 5 (      STSA, B-SLICE, QP 23)            82296 bits [Y 47.0603 dB    U 52.7286 dB    V 51.3601 dB] [ET     0 ] [L0 4 ] [L1 6 ]
POC     7 TId: 5 (      STSA, B-SLICE, QP 24)           115256 bits [Y 46.0015 dB    U 51.5641 dB    V 49.9041 dB] [ET     0 ] [L0 6 ] [L1 8 ]
POC    12 TId: 3 (      STSA, B-SLICE, QP 24)           312096 bits [Y 46.9459 dB    U 51.1315 dB    V 50.4386 dB] [ET     0 ] [L0 8 ] [L1 16 ]
POC    10 TId: 4 (      STSA, B-SLICE, QP 26)           239440 bits [Y 45.9486 dB    U 50.0011 dB    V 49.0176 dB] [ET     0 ] [L0 8 ] [L1 12 ]
POC     9 TId: 5 (      STSA, B-SLICE, QP 23)            56528 bits [Y 47.5988 dB    U 52.5054 dB    V 51.3064 dB] [ET     0 ] [L0 8 ] [L1 10 ]
POC    11 TId: 5 (      STSA, B-SLICE, QP 23)           118464 bits [Y 46.0833 dB    U 51.5704 dB    V 50.3236 dB] [ET     0 ] [L0 10 ] [L1 12 ]
POC    14 TId: 4 (      STSA, B-SLICE, QP 27)           243656 bits [Y 44.8992 dB    U 48.9751 dB    V 48.1517 dB] [ET     0 ] [L0 12 ] [L1 16 ]
POC    13 TId: 5 (      STSA, B-SLICE, QP 21)           120608 bits [Y 46.9868 dB    U 52.5304 dB    V 51.4956 dB] [ET     0 ] [L0 12 ] [L1 14 ]
POC    15 TId: 5 (      STSA, B-SLICE, QP 22)           258928 bits [Y 45.0721 dB    U 50.4006 dB    V 49.3436 dB] [ET     0 ] [L0 14 ] [L1 16 ]
POC    24 TId: 2 (      STSA, B-SLICE, QP 21)           379304 bits [Y 48.2156 dB    U 52.6109 dB    V 51.6196 dB] [ET     0 ] [L0 16 ] [L1 16 ]
POC    20 TId: 3 (      STSA, B-SLICE, QP 22)           295616 bits [Y 46.6235 dB    U 51.1552 dB    V 50.1717 dB] [ET     0 ] [L0 16 ] [L1 24 ]
POC    18 TId: 4 (      STSA, B-SLICE, QP 25)           223184 bits [Y 44.9037 dB    U 49.5202 dB    V 48.3823 dB] [ET     0 ] [L0 16 ] [L1 20 ]
POC    17 TId: 5 (      STSA, B-SLICE, QP 20)            97368 bits [Y 47.8897 dB    U 53.2187 dB    V 52.1750 dB] [ET     0 ] [L0 16 ] [L1 18 ]
POC    19 TId: 5 (      STSA, B-SLICE, QP 19)           221616 bits [Y 47.2766 dB    U 52.3488 dB    V 51.0845 dB] [ET     0 ] [L0 18 ] [L1 20 ]
POC    22 TId: 4 (      STSA, B-SLICE, QP 24)           235288 bits [Y 45.3710 dB    U 50.1320 dB    V 49.1783 dB] [ET     0 ] [L0 20 ] [L1 24 ]
POC    21 TId: 5 (      STSA, B-SLICE, QP 19)           128992 bits [Y 47.6393 dB    U 52.9384 dB    V 51.8818 dB] [ET     0 ] [L0 20 ] [L1 22 ]
vvenc [info]: stats:  51.1% frame= 24/47 fps=   4.8 avg_fps=   4.8 bitrate=  5773.15 kbps avg_bitrate=  5773.15 kbps elapsed= 00h:00m:05s left= 00h:00m:05s
POC    23 TId: 5 (      STSA, B-SLICE, QP 19)           202016 bits [Y 47.1692 dB    U 52.5115 dB    V 51.5600 dB] [ET     0 ] [L0 22 ] [L1 24 ]
POC    28 TId: 3 (      STSA, B-SLICE, QP 23)           267192 bits [Y 46.9249 dB    U 51.4836 dB    V 50.4660 dB] [ET     0 ] [L0 24 ] [L1 24 ]
POC    26 TId: 4 (      STSA, B-SLICE, QP 24)           207704 bits [Y 45.1315 dB    U 49.8533 dB    V 49.1088 dB] [ET     0 ] [L0 24 ] [L1 28 ]
POC    25 TId: 5 (      STSA, B-SLICE, QP 19)            81592 bits [Y 48.3754 dB    U 53.3255 dB    V 52.1993 dB] [ET     0 ] [L0 24 ] [L1 26 ]
POC    27 TId: 5 (      STSA, B-SLICE, QP 19)           205680 bits [Y 47.4046 dB    U 52.6119 dB    V 51.5899 dB] [ET     0 ] [L0 26 ] [L1 28 ]
POC    30 TId: 4 (      STSA, B-SLICE, QP 24)           200624 bits [Y 45.5273 dB    U 50.2978 dB    V 49.5801 dB] [ET     0 ] [L0 28 ] [L1 28 ]
POC    29 TId: 5 (      STSA, B-SLICE, QP 20)            88024 bits [Y 47.5471 dB    U 52.8556 dB    V 51.8869 dB] [ET     0 ] [L0 28 ] [L1 30 ]
POC    31 TId: 5 (      STSA, B-SLICE, QP 18)           195600 bits [Y 47.9366 dB    U 53.1944 dB    V 52.3387 dB] [ET     0 ] [L0 30 ] [L1 30 ]
POC    32 TId: 0 (     TRAIL, B-SLICE, QP 18)           551912 bits [Y 49.2734 dB    U 54.3663 dB    V 53.7849 dB] [ET     0 ] [L0 0 ] [L1 0 ]
POC    40 TId: 2 (      STSA, B-SLICE, QP 18)           322016 bits [Y 49.5046 dB    U 54.2149 dB    V 53.9277 dB] [ET     0 ] [L0 32 ] [L1 32 ]
POC    36 TId: 3 (      STSA, B-SLICE, QP 20)           208840 bits [Y 47.9946 dB    U 53.5233 dB    V 52.7678 dB] [ET     0 ] [L0 32 ] [L1 40 ]
POC    34 TId: 4 (      STSA, B-SLICE, QP 22)           164744 bits [Y 46.2822 dB    U 51.9022 dB    V 51.0221 dB] [ET     0 ] [L0 32 ] [L1 36 ]
POC    33 TId: 5 (      STSA, B-SLICE, QP 19)            79096 bits [Y 49.8151 dB    U 54.9774 dB    V 54.3195 dB] [ET     0 ] [L0 32 ] [L1 34 ]
POC    35 TId: 5 (      STSA, B-SLICE, QP 19)           180712 bits [Y 48.7268 dB    U 54.6767 dB    V 53.8849 dB] [ET     0 ] [L0 34 ] [L1 36 ]
POC    38 TId: 4 (      STSA, B-SLICE, QP 22)           145384 bits [Y 46.8786 dB    U 51.6810 dB    V 51.2991 dB] [ET     0 ] [L0 36 ] [L1 40 ]
POC    37 TId: 5 (      STSA, B-SLICE, QP 19)           102520 bits [Y 49.0579 dB    U 55.4315 dB    V 54.8085 dB] [ET     0 ] [L0 36 ] [L1 38 ]
POC    39 TId: 5 (      STSA, B-SLICE, QP 19)           152176 bits [Y 49.0307 dB    U 54.7695 dB    V 54.3352 dB] [ET     0 ] [L0 38 ] [L1 40 ]
POC    44 TId: 3 (      STSA, B-SLICE, QP 23)           320504 bits [Y 46.1009 dB    U 50.4451 dB    V 50.0183 dB] [ET     0 ] [L0 40 32 ] [L1 40 32 ]
POC    42 TId: 4 (      STSA, B-SLICE, QP 24)           195072 bits [Y 45.2652 dB    U 49.9770 dB    V 49.3828 dB] [ET     0 ] [L0 40 ] [L1 44 ]
POC    41 TId: 5 (      STSA, B-SLICE, QP 19)            88920 bits [Y 50.0731 dB    U 55.1672 dB    V 55.0347 dB] [ET     0 ] [L0 40 ] [L1 42 ]
POC    43 TId: 5 (      STSA, B-SLICE, QP 22)           177168 bits [Y 47.3166 dB    U 51.6071 dB    V 51.0253 dB] [ET     0 ] [L0 42 ] [L1 44 ]
POC    46 TId: 4 (      STSA, B-SLICE, QP 27)           254904 bits [Y 42.9397 dB    U 49.1937 dB    V 48.3834 dB] [ET     0 ] [L0 44 ] [L1 44 ]
POC    45 TId: 5 (      STSA, B-SLICE, QP 22)           125168 bits [Y 47.0075 dB    U 51.6647 dB    V 50.9808 dB] [ET     0 ] [L0 44 40 ] [L1 46 44 ]

Without QPA:

POC     0 TId: 0 (  IDR_N_LP, I-SLICE, QP 17)           430544 bits [Y 53.2183 dB    U 58.1984 dB    V 57.8556 dB] [ET     0 ] [L0 ] [L1 ]
POC    16 TId: 1 (      STSA, B-SLICE, QP 20)           485752 bits [Y 49.8717 dB    U 53.5044 dB    V 52.5491 dB] [ET     0 ] [L0 0 ] [L1 0 ]
POC     8 TId: 2 (      STSA, B-SLICE, QP 21)           422848 bits [Y 49.3465 dB    U 52.7126 dB    V 51.6260 dB] [ET     0 ] [L0 0 ] [L1 16 ]
POC     4 TId: 3 (      STSA, B-SLICE, QP 24)           268520 bits [Y 47.9754 dB    U 51.9189 dB    V 50.8579 dB] [ET     0 ] [L0 0 ] [L1 8 ]
POC     2 TId: 4 (      STSA, B-SLICE, QP 25)           187520 bits [Y 47.7614 dB    U 51.2123 dB    V 50.4113 dB] [ET     0 ] [L0 0 ] [L1 4 ]
POC     1 TId: 5 (      STSA, B-SLICE, QP 27)           153736 bits [Y 46.9146 dB    U 49.9578 dB    V 49.5134 dB] [ET     0 ] [L0 0 2 ] [L1 2 4 ]
POC     3 TId: 5 (      STSA, B-SLICE, QP 27)            81448 bits [Y 46.4660 dB    U 50.8436 dB    V 50.1669 dB] [ET     0 ] [L0 2 ] [L1 4 ]
POC     6 TId: 4 (      STSA, B-SLICE, QP 25)           241920 bits [Y 46.9277 dB    U 50.4038 dB    V 49.4034 dB] [ET     0 ] [L0 4 ] [L1 8 ]
POC     5 TId: 5 (      STSA, B-SLICE, QP 27)            44304 bits [Y 46.6058 dB    U 51.5392 dB    V 50.3779 dB] [ET     0 ] [L0 4 ] [L1 6 ]
POC     7 TId: 5 (      STSA, B-SLICE, QP 27)            41664 bits [Y 45.7391 dB    U 50.3303 dB    V 48.9971 dB] [ET     0 ] [L0 6 ] [L1 8 ]
POC    12 TId: 3 (      STSA, B-SLICE, QP 24)           297776 bits [Y 47.5161 dB    U 51.2791 dB    V 50.5252 dB] [ET     0 ] [L0 8 ] [L1 16 ]
POC    10 TId: 4 (      STSA, B-SLICE, QP 25)           245984 bits [Y 46.8500 dB    U 50.3075 dB    V 49.3512 dB] [ET     0 ] [L0 8 ] [L1 12 ]
POC     9 TId: 5 (      STSA, B-SLICE, QP 27)            23920 bits [Y 47.7560 dB    U 52.1199 dB    V 50.9459 dB] [ET     0 ] [L0 8 ] [L1 10 ]
POC    11 TId: 5 (      STSA, B-SLICE, QP 27)            44000 bits [Y 45.7408 dB    U 50.4018 dB    V 49.3620 dB] [ET     0 ] [L0 10 ] [L1 12 ]
POC    14 TId: 4 (      STSA, B-SLICE, QP 25)           276512 bits [Y 46.6408 dB    U 50.0468 dB    V 49.3003 dB] [ET     0 ] [L0 12 ] [L1 16 ]
POC    13 TId: 5 (      STSA, B-SLICE, QP 27)            26000 bits [Y 46.5308 dB    U 51.1114 dB    V 50.2015 dB] [ET     0 ] [L0 12 ] [L1 14 ]
POC    15 TId: 5 (      STSA, B-SLICE, QP 27)           142112 bits [Y 45.6298 dB    U 49.6385 dB    V 48.7745 dB] [ET     0 ] [L0 14 ] [L1 16 ]
POC    24 TId: 2 (      STSA, B-SLICE, QP 21)           385904 bits [Y 49.7278 dB    U 53.2208 dB    V 52.3298 dB] [ET     0 ] [L0 16 ] [L1 16 ]
POC    20 TId: 3 (      STSA, B-SLICE, QP 24)           283608 bits [Y 47.7869 dB    U 51.5517 dB    V 50.7771 dB] [ET     0 ] [L0 16 ] [L1 24 ]
POC    18 TId: 4 (      STSA, B-SLICE, QP 25)           256120 bits [Y 46.8618 dB    U 50.5699 dB    V 49.6870 dB] [ET     0 ] [L0 16 ] [L1 20 ]
POC    17 TId: 5 (      STSA, B-SLICE, QP 27)            13472 bits [Y 48.7428 dB    U 52.7076 dB    V 51.8251 dB] [ET     0 ] [L0 16 ] [L1 18 ]
POC    19 TId: 5 (      STSA, B-SLICE, QP 27)             8648 bits [Y 46.7241 dB    U 50.5560 dB    V 49.6871 dB] [ET     0 ] [L0 18 ] [L1 20 ]
POC    22 TId: 4 (      STSA, B-SLICE, QP 25)           242240 bits [Y 46.9334 dB    U 50.6733 dB    V 49.8129 dB] [ET     0 ] [L0 20 ] [L1 24 ]
POC    21 TId: 5 (      STSA, B-SLICE, QP 27)             3352 bits [Y 47.6381 dB    U 51.4968 dB    V 50.7518 dB] [ET     0 ] [L0 20 ] [L1 22 ]
vvenc [info]: stats:  51.1% frame= 24/47 fps=   6.1 avg_fps=   6.1 bitrate=  4603.30 kbps avg_bitrate=  4603.30 kbps elapsed= 00h:00m:04s left= 00h:00m:04s
POC    23 TId: 5 (      STSA, B-SLICE, QP 27)             2800 bits [Y 46.8888 dB    U 50.7396 dB    V 49.8795 dB] [ET     0 ] [L0 22 ] [L1 24 ]
POC    28 TId: 3 (      STSA, B-SLICE, QP 24)           262632 bits [Y 48.2926 dB    U 51.9777 dB    V 51.1979 dB] [ET     0 ] [L0 24 ] [L1 24 ]
POC    26 TId: 4 (      STSA, B-SLICE, QP 25)           228776 bits [Y 47.0083 dB    U 50.8350 dB    V 50.0467 dB] [ET     0 ] [L0 24 ] [L1 28 ]
POC    25 TId: 5 (      STSA, B-SLICE, QP 27)             9640 bits [Y 49.2313 dB    U 52.8189 dB    V 51.9627 dB] [ET     0 ] [L0 24 ] [L1 26 ]
POC    27 TId: 5 (      STSA, B-SLICE, QP 27)             3832 bits [Y 46.9765 dB    U 50.8188 dB    V 50.0536 dB] [ET     0 ] [L0 26 ] [L1 28 ]
POC    30 TId: 4 (      STSA, B-SLICE, QP 25)           223616 bits [Y 47.3654 dB    U 51.0124 dB    V 50.4057 dB] [ET     0 ] [L0 28 ] [L1 28 ]
POC    29 TId: 5 (      STSA, B-SLICE, QP 27)            14384 bits [Y 47.8390 dB    U 51.5298 dB    V 51.0538 dB] [ET     0 ] [L0 28 ] [L1 30 ]
POC    31 TId: 5 (      STSA, B-SLICE, QP 27)             6264 bits [Y 47.2566 dB    U 51.0768 dB    V 50.4065 dB] [ET     0 ] [L0 30 ] [L1 30 ]
POC    32 TId: 0 (     TRAIL, B-SLICE, QP 19)           512360 bits [Y 50.7091 dB    U 55.4304 dB    V 54.8389 dB] [ET     0 ] [L0 0 ] [L1 0 ]
POC    40 TId: 2 (      STSA, B-SLICE, QP 21)           263560 bits [Y 50.0798 dB    U 54.2828 dB    V 53.9589 dB] [ET     0 ] [L0 32 ] [L1 32 ]
POC    36 TId: 3 (      STSA, B-SLICE, QP 24)           167304 bits [Y 48.3740 dB    U 53.4674 dB    V 52.8648 dB] [ET     0 ] [L0 32 ] [L1 40 ]
POC    34 TId: 4 (      STSA, B-SLICE, QP 25)           157424 bits [Y 47.3372 dB    U 52.2220 dB    V 51.7003 dB] [ET     0 ] [L0 32 ] [L1 36 ]
POC    33 TId: 5 (      STSA, B-SLICE, QP 27)             3480 bits [Y 50.1172 dB    U 55.0982 dB    V 54.5577 dB] [ET     0 ] [L0 32 ] [L1 34 ]
POC    35 TId: 5 (      STSA, B-SLICE, QP 27)             4512 bits [Y 47.2506 dB    U 52.2257 dB    V 51.7575 dB] [ET     0 ] [L0 34 ] [L1 36 ]
POC    38 TId: 4 (      STSA, B-SLICE, QP 25)           137048 bits [Y 47.8498 dB    U 52.4866 dB    V 52.0673 dB] [ET     0 ] [L0 36 ] [L1 40 ]
POC    37 TId: 5 (      STSA, B-SLICE, QP 27)             4584 bits [Y 48.1655 dB    U 53.5156 dB    V 52.9864 dB] [ET     0 ] [L0 36 ] [L1 38 ]
POC    39 TId: 5 (      STSA, B-SLICE, QP 27)             5896 bits [Y 47.6826 dB    U 52.5857 dB    V 52.1876 dB] [ET     0 ] [L0 38 ] [L1 40 ]
POC    44 TId: 3 (      STSA, B-SLICE, QP 24)           305800 bits [Y 47.8379 dB    U 51.2305 dB    V 50.9995 dB] [ET     1 ] [L0 40 32 ] [L1 40 32 ]
POC    42 TId: 4 (      STSA, B-SLICE, QP 25)           207312 bits [Y 47.3572 dB    U 51.1899 dB    V 50.6637 dB] [ET     0 ] [L0 40 ] [L1 44 ]
POC    41 TId: 5 (      STSA, B-SLICE, QP 27)            10056 bits [Y 49.4899 dB    U 53.6706 dB    V 53.4416 dB] [ET     0 ] [L0 40 ] [L1 42 ]
POC    43 TId: 5 (      STSA, B-SLICE, QP 27)            63456 bits [Y 46.8891 dB    U 50.4590 dB    V 49.9426 dB] [ET     0 ] [L0 42 ] [L1 44 ]
POC    46 TId: 4 (      STSA, B-SLICE, QP 25)           327392 bits [Y 46.3278 dB    U 50.5272 dB    V 49.9832 dB] [ET     0 ] [L0 44 ] [L1 44 ]
POC    45 TId: 5 (      STSA, B-SLICE, QP 27)            43880 bits [Y 47.1070 dB    U 50.5357 dB    V 50.1573 dB] [ET     0 ] [L0 44 40 ] [L1 46 44 ]

Notice how frame 17, 19, 21 and so on jumped from QP 27-ish to 19-ish between two samples while other frames are less significantly changed.

crhelmrich commented 1 month ago

Thanks, L4cache, for the detailed analysis! In case you have a hex editor available and know how to use it: could you please

  1. open the y4m input file you have used above in the hex editor and change the first few characters from YUV4MPEG2 C420p10 W1920 H1080 F24000:1001. . . to YUV4MPEG2 C420p10 W1920 H1080 F60000:1001. . .,
  2. save the result in a new file, QPA-encode that with the same settings and check the per-frame QP data?

This effectively changes the frame rate from 23.976 to 59.94 fps for testing and should help to isolate the problem. Thanks,

Christian

L4cache commented 1 month ago

To make things more clear, I'm using the original file CSM_test_clip_yuv420p10.y4m, not the later uploaded one, I didn't download it but I think it's functionally the same with -fs 1.

The 59.94 fps reports as follows

POC     0 TId: 0 (  IDR_N_LP, I-SLICE, QP 10)           846416 bits [Y 55.6109 dB    U 60.6117 dB    V 60.1662 dB] [ET     0 ] [L0 ] [L1 ]
POC    16 TId: 1 (      STSA, B-SLICE, QP 22)           417768 bits [Y 48.2431 dB    U 52.7842 dB    V 51.7518 dB] [ET     0 ] [L0 0 ] [L1 0 ]
POC     8 TId: 2 (      STSA, B-SLICE, QP 21)           454584 bits [Y 48.8063 dB    U 52.9376 dB    V 51.8077 dB] [ET     0 ] [L0 0 ] [L1 16 ]
POC     4 TId: 3 (      STSA, B-SLICE, QP 22)           335592 bits [Y 48.3010 dB    U 52.7306 dB    V 51.5995 dB] [ET     0 ] [L0 0 ] [L1 8 ]
POC     2 TId: 4 (      STSA, B-SLICE, QP 27)           160688 bits [Y 46.3129 dB    U 50.6753 dB    V 49.7035 dB] [ET     0 ] [L0 0 ] [L1 4 ]
POC     1 TId: 5 (      STSA, B-SLICE, QP 28)           131896 bits [Y 45.5920 dB    U 49.5847 dB    V 48.9226 dB] [ET     0 ] [L0 0 2 ] [L1 2 4 ]
POC     3 TId: 5 (      STSA, B-SLICE, QP 23)           140664 bits [Y 46.4539 dB    U 51.9440 dB    V 50.9965 dB] [ET     0 ] [L0 2 ] [L1 4 ]
POC     6 TId: 4 (      STSA, B-SLICE, QP 24)           269008 bits [Y 46.7617 dB    U 50.7232 dB    V 49.6592 dB] [ET     0 ] [L0 4 ] [L1 8 ]
POC     5 TId: 5 (      STSA, B-SLICE, QP 24)            64128 bits [Y 46.9640 dB    U 52.5232 dB    V 51.2525 dB] [ET     0 ] [L0 4 ] [L1 6 ]
POC     7 TId: 5 (      STSA, B-SLICE, QP 25)            62112 bits [Y 45.7728 dB    U 50.8016 dB    V 49.4913 dB] [ET     0 ] [L0 6 ] [L1 8 ]
POC    12 TId: 3 (      STSA, B-SLICE, QP 23)           358920 bits [Y 47.6277 dB    U 51.7516 dB    V 51.0343 dB] [ET     0 ] [L0 8 ] [L1 16 ]
POC    10 TId: 4 (      STSA, B-SLICE, QP 24)           275736 bits [Y 46.7379 dB    U 50.7964 dB    V 49.7902 dB] [ET     0 ] [L0 8 ] [L1 12 ]
POC     9 TId: 5 (      STSA, B-SLICE, QP 26)            32960 bits [Y 47.4694 dB    U 52.4524 dB    V 51.2320 dB] [ET     0 ] [L0 8 ] [L1 10 ]
POC    11 TId: 5 (      STSA, B-SLICE, QP 25)            60960 bits [Y 45.7882 dB    U 50.8931 dB    V 49.9163 dB] [ET     0 ] [L0 10 ] [L1 12 ]
POC    14 TId: 4 (      STSA, B-SLICE, QP 26)           266936 bits [Y 45.5909 dB    U 49.6675 dB    V 48.7914 dB] [ET     0 ] [L0 12 ] [L1 16 ]
POC    13 TId: 5 (      STSA, B-SLICE, QP 26)            34976 bits [Y 46.6737 dB    U 51.5714 dB    V 50.7319 dB] [ET     0 ] [L0 12 ] [L1 14 ]
POC    15 TId: 5 (      STSA, B-SLICE, QP 28)           122304 bits [Y 44.1318 dB    U 49.0740 dB    V 48.0811 dB] [ET     0 ] [L0 14 ] [L1 16 ]
POC    24 TId: 2 (      STSA, B-SLICE, QP 19)           433872 bits [Y 48.9951 dB    U 53.2815 dB    V 52.3878 dB] [ET     0 ] [L0 16 ] [L1 16 ]
POC    20 TId: 3 (      STSA, B-SLICE, QP 21)           335392 bits [Y 47.3971 dB    U 51.7798 dB    V 50.9777 dB] [ET     0 ] [L0 16 ] [L1 24 ]
POC    18 TId: 4 (      STSA, B-SLICE, QP 24)           252528 bits [Y 45.6840 dB    U 50.0415 dB    V 49.0722 dB] [ET     0 ] [L0 16 ] [L1 20 ]
POC    17 TId: 5 (      STSA, B-SLICE, QP 27)            17952 bits [Y 47.4932 dB    U 52.2680 dB    V 51.3135 dB] [ET     0 ] [L0 16 ] [L1 18 ]
POC    19 TId: 5 (      STSA, B-SLICE, QP 25)            13896 bits [Y 45.5879 dB    U 50.3274 dB    V 49.2577 dB] [ET     0 ] [L0 18 ] [L1 20 ]
POC    22 TId: 4 (      STSA, B-SLICE, QP 23)           264272 bits [Y 46.2046 dB    U 50.8005 dB    V 49.7762 dB] [ET     0 ] [L0 20 ] [L1 24 ]
POC    21 TId: 5 (      STSA, B-SLICE, QP 24)             8744 bits [Y 47.2945 dB    U 51.8468 dB    V 50.9791 dB] [ET     0 ] [L0 20 ] [L1 22 ]
POC    23 TId: 5 (      STSA, B-SLICE, QP 24)            10648 bits [Y 46.1858 dB    U 51.0102 dB    V 49.9896 dB] [ET     0 ] [L0 22 ] [L1 24 ]
POC    28 TId: 3 (      STSA, B-SLICE, QP 21)           310552 bits [Y 47.8112 dB    U 52.2543 dB    V 51.3842 dB] [ET     0 ] [L0 24 ] [L1 24 ]
POC    26 TId: 4 (      STSA, B-SLICE, QP 23)           242816 bits [Y 46.0092 dB    U 50.6642 dB    V 49.7180 dB] [ET     0 ] [L0 24 ] [L1 28 ]
POC    25 TId: 5 (      STSA, B-SLICE, QP 24)            21776 bits [Y 48.6274 dB    U 52.9685 dB    V 52.1410 dB] [ET     0 ] [L0 24 ] [L1 26 ]
POC    27 TId: 5 (      STSA, B-SLICE, QP 24)            16600 bits [Y 46.0051 dB    U 50.7933 dB    V 49.9122 dB] [ET     0 ] [L0 26 ] [L1 28 ]
POC    30 TId: 4 (      STSA, B-SLICE, QP 23)           226040 bits [Y 46.3681 dB    U 51.0292 dB    V 50.0967 dB] [ET     0 ] [L0 28 ] [L1 28 ]
POC    29 TId: 5 (      STSA, B-SLICE, QP 24)            28888 bits [Y 47.4146 dB    U 52.0181 dB    V 51.3057 dB] [ET     0 ] [L0 28 ] [L1 30 ]
POC    31 TId: 5 (      STSA, B-SLICE, QP 24)            11040 bits [Y 46.3451 dB    U 51.2104 dB    V 50.1856 dB] [ET     0 ] [L0 30 ] [L1 30 ]
POC    32 TId: 0 (     TRAIL, B-SLICE, QP 17)           623608 bits [Y 50.0547 dB    U 55.2409 dB    V 54.6955 dB] [ET     0 ] [L0 0 ] [L1 0 ]
POC    40 TId: 2 (      STSA, B-SLICE, QP 17)           367440 bits [Y 50.1177 dB    U 54.9260 dB    V 54.4741 dB] [ET     0 ] [L0 32 ] [L1 32 ]
POC    36 TId: 3 (      STSA, B-SLICE, QP 19)           238640 bits [Y 48.6593 dB    U 54.2452 dB    V 53.4073 dB] [ET     0 ] [L0 32 ] [L1 40 ]
POC    34 TId: 4 (      STSA, B-SLICE, QP 21)           191120 bits [Y 47.0548 dB    U 52.2748 dB    V 51.6382 dB] [ET     0 ] [L0 32 ] [L1 36 ]
POC    33 TId: 5 (      STSA, B-SLICE, QP 25)             9992 bits [Y 49.5208 dB    U 54.9424 dB    V 54.4474 dB] [ET     0 ] [L0 32 ] [L1 34 ]
POC    35 TId: 5 (      STSA, B-SLICE, QP 22)            14288 bits [Y 46.9742 dB    U 52.5735 dB    V 51.9073 dB] [ET     0 ] [L0 34 ] [L1 36 ]
POC    38 TId: 4 (      STSA, B-SLICE, QP 21)           165152 bits [Y 47.5891 dB    U 52.4326 dB    V 52.0483 dB] [ET     0 ] [L0 36 ] [L1 40 ]
POC    37 TId: 5 (      STSA, B-SLICE, QP 23)            13968 bits [Y 48.4364 dB    U 54.2124 dB    V 53.5265 dB] [ET     0 ] [L0 36 ] [L1 38 ]
POC    39 TId: 5 (      STSA, B-SLICE, QP 22)            17088 bits [Y 47.4661 dB    U 52.5431 dB    V 52.3188 dB] [ET     0 ] [L0 38 ] [L1 40 ]
POC    44 TId: 3 (      STSA, B-SLICE, QP 22)           349144 bits [Y 47.0165 dB    U 51.2898 dB    V 50.7982 dB] [ET     0 ] [L0 40 32 ] [L1 40 32 ]
POC    42 TId: 4 (      STSA, B-SLICE, QP 23)           218960 bits [Y 46.0866 dB    U 50.8193 dB    V 50.2444 dB] [ET     0 ] [L0 40 ] [L1 44 ]
POC    41 TId: 5 (      STSA, B-SLICE, QP 22)            24912 bits [Y 49.6042 dB    U 54.4916 dB    V 54.1490 dB] [ET     0 ] [L0 40 ] [L1 42 ]
POC    43 TId: 5 (      STSA, B-SLICE, QP 25)            85312 bits [Y 45.8683 dB    U 50.5941 dB    V 49.6878 dB] [ET     0 ] [L0 42 ] [L1 44 ]
POC    46 TId: 4 (      STSA, B-SLICE, QP 26)           296848 bits [Y 44.1138 dB    U 49.8165 dB    V 49.0596 dB] [ET     0 ] [L0 44 ] [L1 44 ]
POC    45 TId: 5 (      STSA, B-SLICE, QP 26)            53112 bits [Y 46.3283 dB    U 50.5932 dB    V 50.1579 dB] [ET     0 ] [L0 44 40 ] [L1 46 44 ]

But... it decodes with an error (both ffvvc and vvdec), in frame 32, it seems like the error is caused by -dr idr_no_radl, so here's the report with changed command-line: --preset faster -ip -1 --poc0idr 1 --sao 0 --bim 0 --mctf 0 -t 16 --wavefrontsynchro 1 --ctusize 128 --ifp 0 --ciip 3 -qpa 1 --qp 20 -fs 1 It moved frame 32 to second (encode order) and the rest is mostly the same except the bit cost and PSNR, and the decoding error solved!

POC     0 TId: 0 (  IDR_N_LP, I-SLICE, QP 10)           846416 bits [Y 55.6109 dB    U 60.6117 dB    V 60.1662 dB] [ET     0 ] [L0 ] [L1 ]
POC    32 TId: 0 (     TRAIL, B-SLICE, QP 17)           623368 bits [Y 50.0492 dB    U 55.2409 dB    V 54.6856 dB] [ET     0 ] [L0 0 ] [L1 0 ]
POC    16 TId: 1 (      STSA, B-SLICE, QP 22)           403768 bits [Y 48.2369 dB    U 52.8679 dB    V 51.7348 dB] [ET     0 ] [L0 0 ] [L1 32 ]
POC     8 TId: 2 (      STSA, B-SLICE, QP 21)           454344 bits [Y 48.8533 dB    U 52.9276 dB    V 51.7803 dB] [ET     0 ] [L0 0 ] [L1 16 ]
POC     4 TId: 3 (      STSA, B-SLICE, QP 22)           336512 bits [Y 48.3255 dB    U 52.6367 dB    V 51.5535 dB] [ET     0 ] [L0 0 ] [L1 8 ]
POC     2 TId: 4 (      STSA, B-SLICE, QP 27)           155808 bits [Y 46.4918 dB    U 50.5329 dB    V 49.7370 dB] [ET     0 ] [L0 0 ] [L1 4 ]
POC     1 TId: 5 (      STSA, B-SLICE, QP 28)           131320 bits [Y 45.6606 dB    U 49.4500 dB    V 49.0300 dB] [ET     0 ] [L0 0 2 ] [L1 2 4 ]
POC     3 TId: 5 (      STSA, B-SLICE, QP 23)           138672 bits [Y 46.4400 dB    U 51.8202 dB    V 50.9914 dB] [ET     0 ] [L0 2 ] [L1 4 ]
POC     6 TId: 4 (      STSA, B-SLICE, QP 24)           268872 bits [Y 46.7786 dB    U 50.6420 dB    V 49.5962 dB] [ET     0 ] [L0 4 ] [L1 8 ]
POC     5 TId: 5 (      STSA, B-SLICE, QP 24)            64744 bits [Y 46.9892 dB    U 52.5214 dB    V 51.2183 dB] [ET     0 ] [L0 4 ] [L1 6 ]
POC     7 TId: 5 (      STSA, B-SLICE, QP 25)            62448 bits [Y 45.7747 dB    U 50.7248 dB    V 49.4481 dB] [ET     0 ] [L0 6 ] [L1 8 ]
POC    12 TId: 3 (      STSA, B-SLICE, QP 23)           354992 bits [Y 47.6684 dB    U 51.8519 dB    V 50.9932 dB] [ET     0 ] [L0 8 ] [L1 16 ]
POC    10 TId: 4 (      STSA, B-SLICE, QP 24)           277160 bits [Y 46.8051 dB    U 50.8374 dB    V 49.8839 dB] [ET     0 ] [L0 8 ] [L1 12 ]
POC     9 TId: 5 (      STSA, B-SLICE, QP 26)            33168 bits [Y 47.5210 dB    U 52.4773 dB    V 51.2146 dB] [ET     0 ] [L0 8 ] [L1 10 ]
POC    11 TId: 5 (      STSA, B-SLICE, QP 25)            62048 bits [Y 45.8624 dB    U 50.9564 dB    V 49.9489 dB] [ET     0 ] [L0 10 ] [L1 12 ]
POC    14 TId: 4 (      STSA, B-SLICE, QP 26)           268592 bits [Y 45.6074 dB    U 49.8145 dB    V 48.8690 dB] [ET     0 ] [L0 12 ] [L1 16 ]
POC    13 TId: 5 (      STSA, B-SLICE, QP 26)            34712 bits [Y 46.6878 dB    U 51.6670 dB    V 50.7286 dB] [ET     0 ] [L0 12 ] [L1 14 ]
POC    15 TId: 5 (      STSA, B-SLICE, QP 28)           121872 bits [Y 44.1954 dB    U 49.2108 dB    V 48.1351 dB] [ET     0 ] [L0 14 ] [L1 16 ]
POC    24 TId: 2 (      STSA, B-SLICE, QP 19)           420400 bits [Y 49.0538 dB    U 53.3781 dB    V 52.2972 dB] [ET     0 ] [L0 16 ] [L1 32 ]
POC    20 TId: 3 (      STSA, B-SLICE, QP 21)           331152 bits [Y 47.4127 dB    U 51.9657 dB    V 50.8831 dB] [ET     0 ] [L0 16 ] [L1 24 ]
POC    18 TId: 4 (      STSA, B-SLICE, QP 24)           249848 bits [Y 45.6914 dB    U 50.1642 dB    V 49.0045 dB] [ET     0 ] [L0 16 ] [L1 20 ]
POC    17 TId: 5 (      STSA, B-SLICE, QP 27)            17632 bits [Y 47.4971 dB    U 52.3274 dB    V 51.2454 dB] [ET     0 ] [L0 16 ] [L1 18 ]
POC    19 TId: 5 (      STSA, B-SLICE, QP 25)            13544 bits [Y 45.5987 dB    U 50.3947 dB    V 49.1951 dB] [ET     0 ] [L0 18 ] [L1 20 ]
POC    22 TId: 4 (      STSA, B-SLICE, QP 23)           264912 bits [Y 46.2070 dB    U 50.9158 dB    V 49.8489 dB] [ET     0 ] [L0 20 ] [L1 24 ]
POC    21 TId: 5 (      STSA, B-SLICE, QP 24)             8888 bits [Y 47.3088 dB    U 52.0013 dB    V 50.9669 dB] [ET     0 ] [L0 20 ] [L1 22 ]
POC    23 TId: 5 (      STSA, B-SLICE, QP 24)            10800 bits [Y 46.1868 dB    U 51.0704 dB    V 49.9587 dB] [ET     0 ] [L0 22 ] [L1 24 ]
POC    28 TId: 3 (      STSA, B-SLICE, QP 21)           282144 bits [Y 48.0095 dB    U 52.3764 dB    V 51.4839 dB] [ET     0 ] [L0 24 ] [L1 32 ]
POC    26 TId: 4 (      STSA, B-SLICE, QP 23)           244040 bits [Y 46.2076 dB    U 50.6365 dB    V 49.7232 dB] [ET     0 ] [L0 24 ] [L1 28 ]
POC    25 TId: 5 (      STSA, B-SLICE, QP 24)            21808 bits [Y 48.7043 dB    U 53.0521 dB    V 52.0842 dB] [ET     0 ] [L0 24 ] [L1 26 ]
POC    27 TId: 5 (      STSA, B-SLICE, QP 24)            18592 bits [Y 46.2001 dB    U 50.7753 dB    V 49.8590 dB] [ET     0 ] [L0 26 ] [L1 28 ]
POC    30 TId: 4 (      STSA, B-SLICE, QP 23)           197728 bits [Y 47.0115 dB    U 51.1382 dB    V 50.2935 dB] [ET     0 ] [L0 28 ] [L1 32 ]
POC    29 TId: 5 (      STSA, B-SLICE, QP 24)            28688 bits [Y 47.6179 dB    U 52.0992 dB    V 51.3438 dB] [ET     0 ] [L0 28 ] [L1 30 ]
POC    31 TId: 5 (      STSA, B-SLICE, QP 24)             9632 bits [Y 46.9441 dB    U 51.2662 dB    V 50.4237 dB] [ET     0 ] [L0 30 ] [L1 32 ]
POC    40 TId: 2 (      STSA, B-SLICE, QP 17)           363944 bits [Y 50.1363 dB    U 54.8992 dB    V 54.5949 dB] [ET     0 ] [L0 32 ] [L1 32 ]
POC    36 TId: 3 (      STSA, B-SLICE, QP 19)           243640 bits [Y 48.6153 dB    U 54.3134 dB    V 53.5017 dB] [ET     0 ] [L0 32 ] [L1 40 ]
POC    34 TId: 4 (      STSA, B-SLICE, QP 21)           192176 bits [Y 47.0590 dB    U 52.2053 dB    V 51.6733 dB] [ET     0 ] [L0 32 ] [L1 36 ]
POC    33 TId: 5 (      STSA, B-SLICE, QP 25)             9944 bits [Y 49.5105 dB    U 54.9476 dB    V 54.4346 dB] [ET     0 ] [L0 32 ] [L1 34 ]
POC    35 TId: 5 (      STSA, B-SLICE, QP 22)            14368 bits [Y 47.0074 dB    U 52.5142 dB    V 52.0417 dB] [ET     0 ] [L0 34 ] [L1 36 ]
POC    38 TId: 4 (      STSA, B-SLICE, QP 21)           164712 bits [Y 47.6078 dB    U 52.3046 dB    V 52.0812 dB] [ET     0 ] [L0 36 ] [L1 40 ]
POC    37 TId: 5 (      STSA, B-SLICE, QP 23)            13880 bits [Y 48.3947 dB    U 54.2653 dB    V 53.5719 dB] [ET     0 ] [L0 36 ] [L1 38 ]
POC    39 TId: 5 (      STSA, B-SLICE, QP 22)            17608 bits [Y 47.4957 dB    U 52.4954 dB    V 52.3227 dB] [ET     0 ] [L0 38 ] [L1 40 ]
POC    44 TId: 3 (      STSA, B-SLICE, QP 22)           353328 bits [Y 47.0015 dB    U 51.2874 dB    V 50.6600 dB] [ET     0 ] [L0 40 32 ] [L1 40 32 ]
POC    42 TId: 4 (      STSA, B-SLICE, QP 23)           215440 bits [Y 46.0813 dB    U 50.7515 dB    V 50.1705 dB] [ET     0 ] [L0 40 ] [L1 44 ]
POC    41 TId: 5 (      STSA, B-SLICE, QP 22)            25456 bits [Y 49.5983 dB    U 54.4974 dB    V 54.2019 dB] [ET     0 ] [L0 40 ] [L1 42 ]
POC    43 TId: 5 (      STSA, B-SLICE, QP 25)            81712 bits [Y 45.8102 dB    U 50.3893 dB    V 49.5247 dB] [ET     0 ] [L0 42 ] [L1 44 ]
POC    46 TId: 4 (      STSA, B-SLICE, QP 26)           289600 bits [Y 44.2769 dB    U 49.8581 dB    V 49.1816 dB] [ET     0 ] [L0 44 ] [L1 44 ]
POC    45 TId: 5 (      STSA, B-SLICE, QP 26)            53120 bits [Y 46.3434 dB    U 50.5465 dB    V 49.8478 dB] [ET     0 ] [L0 44 40 ] [L1 46 44 ]
crhelmrich commented 1 month ago

Great, thanks! With those statistics, I can confirm that there are 2 separate issues here: one at frame 0 (reported by joletb), and one at frames 17, 19, 21, ... (reported by L4cache). The second one is relatively easy to fix and I'll start working on that soon. The first one is a bit harder to fix, so will take some time.

However, the fix to the second issue will require MCTF to be enabled (to have access to certain picture statistics). @joletb,

is there a particular reason why you chose to set --MCTF 0? This is an adaptive filter, which should be visually transparent even on noiseless content such as this anime clip. If it's picture blurring related, see also #406 to reduce the MCTF effect.

Christian

joletb commented 1 month ago

Great, thanks! With those statistics, I can confirm that there are 2 separate issues here: one at frame 0 (reported by joletb), and one at frames 17, 19, 21, ... (reported by L4cache). The second one is relatively easy to fix and I'll start working on that soon. The first one is a bit harder to fix, so will take some time.

However, the fix to the second issue will require MCTF to be enabled (to have access to certain picture statistics). @joletb,

is there a particular reason why you chose to set --MCTF 0? This is an adaptive filter, which should be visually transparent even on noiseless content such as this anime clip. If it's picture blurring related, see also #406 to reduce the MCTF effect.

Christian

Heya, Christian, great to hear progress is being made on this particular issue!

To address MCTF, it being set to 0 is mostly a remnant from my previous encoding tests. (Whose videos were far more noisy) Feel free to disregard it since like you said, it should have no real visual impact on this particular video.

Unfortunately, I am not able to produce any bitstreams until later next week, so @L4cache if you could help out a bit more with this, it'd be greatly appreciated!

joletb commented 2 weeks ago

Apologies for the delays, life has been getting busy.

Here are the stats for the MCTF encode, command line is: vvencffapp -c config/slow --y4m 1 -q 20 -t -1 --PrintFrameMSE 1 --PrintSequenceMSE 1 --ALF 1 -ip 0 -rs 11 --CCALF 1 --STA 0 --WppBitEqual -1 --MaxParallelFrames -1 --SAO 0 --qpa 1 --InputFile CSM_test_clip_yuv420p10.y4m -b slow-qpa-mctf-q20.266

POC    31 TId: 0 (IDR_W_RADL, I-SLICE, QP 18, TF 3)     438680 bits [Y 50.3954 dB    U 54.7653 dB    V 53.8979 dB] [Y MSE 9.4986  U MSE 3.4727  V MSE 4.2404] [ET    17 ] [L0 ] [L1 ]
POC    15 TId: 1 (      RADL, B-SLICE, QP 24, TF 2)     334888 bits [Y 47.0428 dB    U 51.3941 dB    V 50.1958 dB] [Y MSE 20.5550  U MSE 7.5472  V MSE 9.9453] [ET    14 ] [L0 31 ] [L1 31 ]
POC     7 TId: 2 (      RADL, B-SLICE, QP 24, TF 1)     254272 bits [Y 47.5767 dB    U 52.3244 dB    V 50.8937 dB] [Y MSE 18.1773  U MSE 6.0920  V MSE 8.4691] [ET    18 ] [L0 15 31 ] [L1 15 31 ]
POC     3 TId: 3 (      RADL, B-SLICE, QP 24, TF 0)     166080 bits [Y 47.5643 dB    U 52.3072 dB    V 51.1129 dB] [Y MSE 18.2294  U MSE 6.1161  V MSE 8.0520] [ET    26 ] [L0 7 15 ] [L1 7 15 ]
POC     1 TId: 4 (      RADL, B-SLICE, QP 17)           271056 bits [Y 51.3047 dB    U 57.1573 dB    V 56.1726 dB] [Y MSE 7.7043  U MSE 2.0020  V MSE 2.5115] [ET    25 ] [L0 3 7 ] [L1 3 7 ]
POC     0 TId: 5 (      RADL, B-SLICE, QP 18)             2280 bits [Y 51.3249 dB    U 57.2696 dB    V 56.3124 dB] [Y MSE 7.6686  U MSE 1.9509  V MSE 2.4320] [ET     1 ] [L0 1 3 ] [L1 1 3 ]
POC     2 TId: 5 (      RADL, B-SLICE, QP 28)           110584 bits [Y 46.5240 dB    U 50.4265 dB    V 49.7044 dB] [Y MSE 23.1633  U MSE 9.4308  V MSE 11.1368] [ET    21 ] [L0 1 3 ] [L1 3 7 ]
POC     5 TId: 4 (      RADL, B-SLICE, QP 25)           178192 bits [Y 47.1160 dB    U 52.1988 dB    V 50.6677 dB] [Y MSE 20.2114  U MSE 6.2708  V MSE 8.9214] [ET    27 ] [L0 3 7 ] [L1 7 15 ]
POC     4 TId: 5 (      RADL, B-SLICE, QP 23)           128520 bits [Y 48.1941 dB    U 53.2970 dB    V 52.1435 dB] [Y MSE 15.7685  U MSE 4.8696  V MSE 6.3512] [ET    15 ] [L0 3 5 ] [L1 5 7 ]
POC     6 TId: 5 (      RADL, B-SLICE, QP 23)            92952 bits [Y 47.4211 dB    U 53.1909 dB    V 51.6472 dB] [Y MSE 18.8405  U MSE 4.9901  V MSE 7.1200] [ET    12 ] [L0 5 3 ] [L1 7 15 ]
POC    11 TId: 3 (      RADL, B-SLICE, QP 26, TF 0)     207720 bits [Y 46.7905 dB    U 51.0987 dB    V 49.8808 dB] [Y MSE 21.7845  U MSE 8.0784  V MSE 10.6935] [ET    32 ] [L0 7 15 ] [L1 15 31 ]
POC     9 TId: 4 (      RADL, B-SLICE, QP 26)           199440 bits [Y 46.5932 dB    U 50.9319 dB    V 49.5454 dB] [Y MSE 22.7973  U MSE 8.3947  V MSE 11.5520] [ET    23 ] [L0 7 11 ] [L1 11 15 ]
POC     8 TId: 5 (      RADL, B-SLICE, QP 24)            78136 bits [Y 47.2833 dB    U 52.7852 dB    V 51.0158 dB] [Y MSE 19.4479  U MSE 5.4788  V MSE 8.2342] [ET    11 ] [L0 7 9 ] [L1 9 11 ]
POC    10 TId: 5 (      RADL, B-SLICE, QP 23)            94832 bits [Y 47.1757 dB    U 52.3468 dB    V 50.8039 dB] [Y MSE 19.9357  U MSE 6.0606  V MSE 8.6460] [ET    11 ] [L0 9 7 ] [L1 11 15 ]
POC    13 TId: 4 (      RADL, B-SLICE, QP 26)           194680 bits [Y 46.5341 dB    U 51.2063 dB    V 50.1098 dB] [Y MSE 23.1095  U MSE 7.8808  V MSE 10.1442] [ET    20 ] [L0 11 7 ] [L1 15 31 ]
POC    12 TId: 5 (      RADL, B-SLICE, QP 23)           101088 bits [Y 46.9884 dB    U 52.5017 dB    V 51.0637 dB] [Y MSE 20.8144  U MSE 5.8483  V MSE 8.1438] [ET    11 ] [L0 11 7 ] [L1 13 15 ]
POC    14 TId: 5 (      RADL, B-SLICE, QP 21)           146632 bits [Y 47.2896 dB    U 53.1361 dB    V 51.8816 dB] [Y MSE 19.4197  U MSE 5.0534  V MSE 6.7458] [ET    20 ] [L0 13 11 ] [L1 15 31 ]
POC    23 TId: 2 (      RADL, B-SLICE, QP 22, TF 1)     257640 bits [Y 47.7786 dB    U 52.2686 dB    V 51.1156 dB] [Y MSE 17.3517  U MSE 6.1708  V MSE 8.0471] [ET    16 ] [L0 15 31 ] [L1 31 15 ]
POC    19 TId: 3 (      RADL, B-SLICE, QP 25, TF 0)     186520 bits [Y 46.1958 dB    U 50.8995 dB    V 49.6530 dB] [Y MSE 24.9813  U MSE 8.4577  V MSE 11.2694] [ET    25 ] [L0 15 23 ] [L1 23 31 ]
POC    17 TId: 4 (      RADL, B-SLICE, QP 27)           181560 bits [Y 45.8104 dB    U 50.7842 dB    V 49.4205 dB] [Y MSE 27.2997  U MSE 8.6853  V MSE 11.8892] [ET    23 ] [L0 15 19 ] [L1 19 23 ]
POC    16 TId: 5 (      RADL, B-SLICE, QP 22)           187984 bits [Y 46.2393 dB    U 51.6784 dB    V 50.1270 dB] [Y MSE 24.7326  U MSE 7.0691  V MSE 10.1041] [ET    23 ] [L0 15 17 ] [L1 17 19 ]
POC    18 TId: 5 (      RADL, B-SLICE, QP 20)           189176 bits [Y 47.4962 dB    U 53.1370 dB    V 51.7673 dB] [Y MSE 18.5176  U MSE 5.0524  V MSE 6.9258] [ET    21 ] [L0 17 15 ] [L1 19 23 ]
POC    21 TId: 4 (      RADL, B-SLICE, QP 24)           177720 bits [Y 46.2248 dB    U 51.3321 dB    V 50.1169 dB] [Y MSE 24.8155  U MSE 7.6557  V MSE 10.1276] [ET    21 ] [L0 19 15 ] [L1 23 31 ]
POC    20 TId: 5 (      RADL, B-SLICE, QP 19)           168424 bits [Y 48.1651 dB    U 53.3149 dB    V 51.7849 dB] [Y MSE 15.8742  U MSE 4.8496  V MSE 6.8978] [ET    20 ] [L0 19 15 ] [L1 21 23 ]
POC    22 TId: 5 (      RADL, B-SLICE, QP 19)           176704 bits [Y 48.3106 dB    U 53.6533 dB    V 52.2536 dB] [Y MSE 15.3513  U MSE 4.4861  V MSE 6.1921] [ET    17 ] [L0 21 19 ] [L1 23 31 ]
POC    27 TId: 3 (      RADL, B-SLICE, QP 24, TF 0)     172248 bits [Y 46.7968 dB    U 51.4064 dB    V 50.2690 dB] [Y MSE 21.7530  U MSE 7.5260  V MSE 9.7791] [ET    23 ] [L0 23 15 ] [L1 31 23 ]
POC    25 TId: 4 (      RADL, B-SLICE, QP 25)           161048 bits [Y 47.0566 dB    U 51.6740 dB    V 50.5681 dB] [Y MSE 20.4899  U MSE 7.0763  V MSE 9.1284] [ET    24 ] [L0 23 15 ] [L1 27 31 ]
POC    24 TId: 5 (      RADL, B-SLICE, QP 19)           102304 bits [Y 48.4831 dB    U 53.5864 dB    V 52.3457 dB] [Y MSE 14.7534  U MSE 4.5557  V MSE 6.0622] [ET    16 ] [L0 23 15 ] [L1 25 27 ]
POC    26 TId: 5 (      RADL, B-SLICE, QP 19)           152000 bits [Y 48.5488 dB    U 53.5142 dB    V 52.3252 dB] [Y MSE 14.5319  U MSE 4.6321  V MSE 6.0909] [ET    21 ] [L0 25 23 ] [L1 27 31 ]
POC    29 TId: 4 (      RADL, B-SLICE, QP 25)           148416 bits [Y 47.1417 dB    U 51.5465 dB    V 50.4152 dB] [Y MSE 20.0923  U MSE 7.2870  V MSE 9.4553] [ET    21 ] [L0 27 15 ] [L1 31 27 ]
POC    28 TId: 5 (      RADL, B-SLICE, QP 19)           151784 bits [Y 48.6272 dB    U 53.6850 dB    V 52.2819 dB] [Y MSE 14.2717  U MSE 4.4535  V MSE 6.1519] [ET    13 ] [L0 27 15 ] [L1 29 31 ]
POC    30 TId: 5 (      RADL, B-SLICE, QP 20)           114600 bits [Y 48.8166 dB    U 53.7797 dB    V 52.5945 dB] [Y MSE 13.6630  U MSE 4.3574  V MSE 5.7247] [ET    10 ] [L0 29 27 ] [L1 31 29 ]
POC    47 TId: 1 (      STSA, B-SLICE, QP 24, TF 2)     257336 bits [Y 46.2034 dB    U 50.6985 dB    V 50.0403 dB] [Y MSE 24.9379  U MSE 8.8583  V MSE 10.3078] [ET    16 ] [L0 31 ] [L1 31 ]
POC    39 TId: 2 (      STSA, B-SLICE, QP 19, TF 1)     273200 bits [Y 48.6358 dB    U 53.9691 dB    V 53.4464 dB] [Y MSE 14.2436  U MSE 4.1715  V MSE 4.7050] [ET    14 ] [L0 31 47 ] [L1 47 31 ]
POC    35 TId: 3 (      STSA, B-SLICE, QP 22, TF 0)     159112 bits [Y 47.1581 dB    U 52.8945 dB    V 52.0230 dB] [Y MSE 20.0167  U MSE 5.3426  V MSE 6.5298] [ET    13 ] [L0 31 39 ] [L1 39 47 ]
POC    33 TId: 4 (      STSA, B-SLICE, QP 25)           126552 bits [Y 46.6110 dB    U 51.7645 dB    V 50.9545 dB] [Y MSE 22.7039  U MSE 6.9303  V MSE 8.3512] [ET    20 ] [L0 31 35 ] [L1 35 39 ]
POC    32 TId: 5 (      STSA, B-SLICE, QP 18)            45896 bits [Y 50.7515 dB    U 55.2753 dB    V 54.2179 dB] [Y MSE 8.7510  U MSE 3.0879  V MSE 3.9392] [ET     9 ] [L0 31 33 ] [L1 33 35 ]
POC    34 TId: 5 (      STSA, B-SLICE, QP 19)           186832 bits [Y 49.2955 dB    U 54.7744 dB    V 53.9358 dB] [Y MSE 12.2364  U MSE 3.4655  V MSE 4.2036] [ET    17 ] [L0 33 31 ] [L1 35 39 ]
POC    37 TId: 4 (      STSA, B-SLICE, QP 22)           120920 bits [Y 47.5435 dB    U 53.5297 dB    V 52.5316 dB] [Y MSE 18.3170  U MSE 4.6156  V MSE 5.8082] [ET    16 ] [L0 35 31 ] [L1 39 47 ]
POC    36 TId: 5 (      STSA, B-SLICE, QP 19)           139968 bits [Y 49.2605 dB    U 55.3072 dB    V 54.4835 dB] [Y MSE 12.3354  U MSE 3.0653  V MSE 3.7055] [ET    13 ] [L0 35 31 ] [L1 37 39 ]
POC    38 TId: 5 (      STSA, B-SLICE, QP 19)           127688 bits [Y 49.3207 dB    U 55.9173 dB    V 55.2597 dB] [Y MSE 12.1656  U MSE 2.6636  V MSE 3.0991] [ET    10 ] [L0 37 35 ] [L1 39 47 ]
POC    43 TId: 3 (      STSA, B-SLICE, QP 23, TF 0)     153840 bits [Y 46.2067 dB    U 50.9709 dB    V 50.0897 dB] [Y MSE 24.9188  U MSE 8.3198  V MSE 10.1913] [ET    14 ] [L0 39 31 ] [L1 47 39 ]
POC    41 TId: 4 (      STSA, B-SLICE, QP 22)           112736 bits [Y 48.1694 dB    U 53.0769 dB    V 52.8188 dB] [Y MSE 15.8585  U MSE 5.1228  V MSE 5.4365] [ET    18 ] [L0 39 31 ] [L1 43 47 ]
POC    40 TId: 5 (      STSA, B-SLICE, QP 19)            97136 bits [Y 49.7481 dB    U 55.6002 dB    V 55.1513 dB] [Y MSE 11.0253  U MSE 2.8654  V MSE 3.1774] [ET    16 ] [L0 39 31 ] [L1 41 43 ]
POC    42 TId: 5 (      STSA, B-SLICE, QP 19)           130032 bits [Y 49.8833 dB    U 55.4007 dB    V 55.0606 dB] [Y MSE 10.6873  U MSE 3.0001  V MSE 3.2445] [ET    11 ] [L0 41 39 ] [L1 43 47 ]
POC    45 TId: 4 (      STSA, B-SLICE, QP 25)           167808 bits [Y 45.7689 dB    U 49.9605 dB    V 49.3804 dB] [Y MSE 27.5621  U MSE 10.4989  V MSE 11.9994] [ET    36 ] [L0 43 39 ] [L1 47 43 ]
POC    44 TId: 5 (      STSA, B-SLICE, QP 22)           143808 bits [Y 48.1272 dB    U 52.6358 dB    V 51.8686 dB] [Y MSE 16.0134  U MSE 5.6705  V MSE 6.7661] [ET    14 ] [L0 43 39 ] [L1 45 47 ]
POC    46 TId: 5 (      STSA, B-SLICE, QP 22)           141184 bits [Y 47.8274 dB    U 52.0126 dB    V 51.3625 dB] [Y MSE 17.1576  U MSE 6.5454  V MSE 7.6023] [ET    14 ] [L0 45 43 ] [L1 47 45 ]

And here are the stats of the original command line in case they are necessary: vvencffapp -c config/slow --y4m 1 -q 20 -t -1 --PrintFrameMSE 1 --PrintSequenceMSE 1 --ALF 1 -ip 0 -rs 11 --CCALF 1 --STA 0 --WppBitEqual -1 --MaxParallelFrames -1 --SAO 0 --qpa 1 --BIM 0 --MCTF 0 --InputFile CSM_test_clip_yuv420p10.y4m -b slow-qpa-q20.266

POC    31 TId: 0 (IDR_W_RADL, I-SLICE, QP 17)           464216 bits [Y 50.9114 dB    U 55.2615 dB    V 54.3152 dB] [Y MSE 8.4346  U MSE 3.0978  V MSE 3.8519] [ET    18 ] [L0 ] [L1 ]
POC    15 TId: 1 (      RADL, B-SLICE, QP 22)           417448 bits [Y 48.7371 dB    U 52.8463 dB    V 51.8269 dB] [Y MSE 13.9154  U MSE 5.4022  V MSE 6.8314] [ET    18 ] [L0 31 ] [L1 31 ]
POC     7 TId: 2 (      RADL, B-SLICE, QP 22)           321344 bits [Y 49.1225 dB    U 53.4748 dB    V 52.0954 dB] [Y MSE 12.7335  U MSE 4.6744  V MSE 6.4218] [ET    24 ] [L0 15 31 ] [L1 15 31 ]
POC     3 TId: 3 (      RADL, B-SLICE, QP 22)           210712 bits [Y 48.9301 dB    U 53.5510 dB    V 52.4105 dB] [Y MSE 13.3104  U MSE 4.5930  V MSE 5.9724] [ET    28 ] [L0 7 15 ] [L1 7 15 ]
POC     1 TId: 4 (      RADL, B-SLICE, QP 17)           269752 bits [Y 51.4847 dB    U 57.1597 dB    V 56.3225 dB] [Y MSE 7.3915  U MSE 2.0009  V MSE 2.4263] [ET    28 ] [L0 3 7 ] [L1 3 7 ]
POC     0 TId: 5 (      RADL, B-SLICE, QP 18)             2184 bits [Y 51.5069 dB    U 57.2344 dB    V 56.3855 dB] [Y MSE 7.3537  U MSE 1.9668  V MSE 2.3914] [ET     4 ] [L0 1 3 ] [L1 1 3 ]
POC     2 TId: 5 (      RADL, B-SLICE, QP 28)           110584 bits [Y 46.7311 dB    U 50.4944 dB    V 49.8254 dB] [Y MSE 22.0845  U MSE 9.2844  V MSE 10.8308] [ET    21 ] [L0 1 3 ] [L1 3 7 ]
POC     5 TId: 4 (      RADL, B-SLICE, QP 25)           176160 bits [Y 47.2341 dB    U 52.2349 dB    V 50.6620 dB] [Y MSE 19.6696  U MSE 6.2188  V MSE 8.9331] [ET    24 ] [L0 3 7 ] [L1 7 15 ]
POC     4 TId: 5 (      RADL, B-SLICE, QP 23)           113112 bits [Y 48.4784 dB    U 53.5615 dB    V 52.5539 dB] [Y MSE 14.7693  U MSE 4.5820  V MSE 5.7785] [ET    11 ] [L0 3 5 ] [L1 5 7 ]
POC     6 TId: 5 (      RADL, B-SLICE, QP 23)            96216 bits [Y 47.6210 dB    U 53.3453 dB    V 51.7631 dB] [Y MSE 17.9930  U MSE 4.8158  V MSE 6.9325] [ET    10 ] [L0 5 3 ] [L1 7 15 ]
POC    11 TId: 3 (      RADL, B-SLICE, QP 24)           256272 bits [Y 48.1585 dB    U 52.2198 dB    V 51.1777 dB] [Y MSE 15.8984  U MSE 6.2405  V MSE 7.9328] [ET    28 ] [L0 7 15 ] [L1 15 31 ]
POC     9 TId: 4 (      RADL, B-SLICE, QP 26)           200704 bits [Y 46.8219 dB    U 51.1120 dB    V 49.5895 dB] [Y MSE 21.6275  U MSE 8.0537  V MSE 11.4353] [ET    28 ] [L0 7 11 ] [L1 11 15 ]
POC     8 TId: 5 (      RADL, B-SLICE, QP 24)            63480 bits [Y 47.9210 dB    U 53.1778 dB    V 51.5108 dB] [Y MSE 16.7919  U MSE 5.0052  V MSE 7.3471] [ET     9 ] [L0 7 9 ] [L1 9 11 ]
POC    10 TId: 5 (      RADL, B-SLICE, QP 23)            94728 bits [Y 47.4088 dB    U 52.4709 dB    V 50.9674 dB] [Y MSE 18.8939  U MSE 5.8899  V MSE 8.3264] [ET    12 ] [L0 9 7 ] [L1 11 15 ]
POC    13 TId: 4 (      RADL, B-SLICE, QP 26)           193680 bits [Y 46.7343 dB    U 51.1214 dB    V 50.1298 dB] [Y MSE 22.0683  U MSE 8.0364  V MSE 10.0978] [ET    21 ] [L0 11 7 ] [L1 15 31 ]
POC    12 TId: 5 (      RADL, B-SLICE, QP 23)            81992 bits [Y 47.6018 dB    U 52.8203 dB    V 51.6354 dB] [Y MSE 18.0727  U MSE 5.4347  V MSE 7.1394] [ET    17 ] [L0 11 7 ] [L1 13 15 ]
POC    14 TId: 5 (      RADL, B-SLICE, QP 21)           147096 bits [Y 47.5121 dB    U 53.2441 dB    V 51.9655 dB] [Y MSE 18.4498  U MSE 4.9294  V MSE 6.6168] [ET    16 ] [L0 13 11 ] [L1 15 31 ]
POC    23 TId: 2 (      RADL, B-SLICE, QP 20)           319880 bits [Y 48.9256 dB    U 53.2782 dB    V 52.2214 dB] [Y MSE 13.3243  U MSE 4.8908  V MSE 6.2382] [ET    18 ] [L0 15 31 ] [L1 31 15 ]
POC    19 TId: 3 (      RADL, B-SLICE, QP 23)           223208 bits [Y 47.3741 dB    U 51.9289 dB    V 50.6228 dB] [Y MSE 19.0453  U MSE 6.6729  V MSE 9.0141] [ET    25 ] [L0 15 23 ] [L1 23 31 ]
POC    17 TId: 4 (      RADL, B-SLICE, QP 27)           181760 bits [Y 45.9570 dB    U 50.8346 dB    V 49.6993 dB] [Y MSE 26.3937  U MSE 8.5851  V MSE 11.1500] [ET    26 ] [L0 15 19 ] [L1 19 23 ]
POC    16 TId: 5 (      RADL, B-SLICE, QP 22)           168360 bits [Y 46.5397 dB    U 51.8973 dB    V 50.4469 dB] [Y MSE 23.0797  U MSE 6.7215  V MSE 9.3866] [ET    22 ] [L0 15 17 ] [L1 17 19 ]
POC    18 TId: 5 (      RADL, B-SLICE, QP 20)           186752 bits [Y 47.6595 dB    U 53.1971 dB    V 51.9445 dB] [Y MSE 17.8342  U MSE 4.9829  V MSE 6.6489] [ET    21 ] [L0 17 15 ] [L1 19 23 ]
POC    21 TId: 4 (      RADL, B-SLICE, QP 24)           176928 bits [Y 46.3805 dB    U 51.3665 dB    V 50.0588 dB] [Y MSE 23.9415  U MSE 7.5954  V MSE 10.2640] [ET    24 ] [L0 19 15 ] [L1 23 31 ]
POC    20 TId: 5 (      RADL, B-SLICE, QP 19)           123688 bits [Y 48.4391 dB    U 53.4877 dB    V 52.0459 dB] [Y MSE 14.9036  U MSE 4.6605  V MSE 6.4954] [ET    18 ] [L0 19 15 ] [L1 21 23 ]
POC    22 TId: 5 (      RADL, B-SLICE, QP 19)           172832 bits [Y 48.4454 dB    U 53.6536 dB    V 52.4177 dB] [Y MSE 14.8818  U MSE 4.4858  V MSE 5.9625] [ET    16 ] [L0 21 19 ] [L1 23 31 ]
POC    27 TId: 3 (      RADL, B-SLICE, QP 22)           214872 bits [Y 47.9046 dB    U 52.3071 dB    V 51.2963 dB] [Y MSE 16.8554  U MSE 6.1163  V MSE 7.7191] [ET    31 ] [L0 23 15 ] [L1 31 23 ]
POC    25 TId: 4 (      RADL, B-SLICE, QP 25)           160984 bits [Y 47.2213 dB    U 51.6939 dB    V 50.4844 dB] [Y MSE 19.7276  U MSE 7.0438  V MSE 9.3060] [ET    17 ] [L0 23 15 ] [L1 27 31 ]
POC    24 TId: 5 (      RADL, B-SLICE, QP 19)            61480 bits [Y 49.1968 dB    U 53.9299 dB    V 52.7323 dB] [Y MSE 12.5175  U MSE 4.2093  V MSE 5.5459] [ET    14 ] [L0 23 15 ] [L1 25 27 ]
POC    26 TId: 5 (      RADL, B-SLICE, QP 19)           150584 bits [Y 48.7318 dB    U 53.5871 dB    V 52.2787 dB] [Y MSE 13.9323  U MSE 4.5551  V MSE 6.1565] [ET    18 ] [L0 25 23 ] [L1 27 31 ]
POC    29 TId: 4 (      RADL, B-SLICE, QP 25)           147096 bits [Y 47.1379 dB    U 51.5557 dB    V 50.4743 dB] [Y MSE 20.1097  U MSE 7.2717  V MSE 9.3276] [ET    17 ] [L0 27 15 ] [L1 31 27 ]
POC    28 TId: 5 (      RADL, B-SLICE, QP 19)           109416 bits [Y 48.9258 dB    U 53.8594 dB    V 52.5824 dB] [Y MSE 13.3236  U MSE 4.2782  V MSE 5.7406] [ET    16 ] [L0 27 15 ] [L1 29 31 ]
POC    30 TId: 5 (      RADL, B-SLICE, QP 20)           118320 bits [Y 48.9611 dB    U 53.8123 dB    V 52.5635 dB] [Y MSE 13.2156  U MSE 4.3248  V MSE 5.7657] [ET    15 ] [L0 29 27 ] [L1 31 29 ]
POC    47 TId: 1 (      STSA, B-SLICE, QP 22)           325400 bits [Y 47.8166 dB    U 52.6060 dB    V 51.8153 dB] [Y MSE 17.2005  U MSE 5.7096  V MSE 6.8496] [ET    16 ] [L0 31 ] [L1 31 ]
POC    39 TId: 2 (      STSA, B-SLICE, QP 18)           319144 bits [Y 49.5020 dB    U 54.7813 dB    V 54.3474 dB] [Y MSE 11.6681  U MSE 3.4600  V MSE 3.8235] [ET    16 ] [L0 31 47 ] [L1 47 31 ]
POC    35 TId: 3 (      STSA, B-SLICE, QP 20)           187056 bits [Y 48.1129 dB    U 53.9460 dB    V 53.0308 dB] [Y MSE 16.0659  U MSE 4.1937  V MSE 5.1775] [ET    16 ] [L0 31 39 ] [L1 39 47 ]
POC    33 TId: 4 (      STSA, B-SLICE, QP 25)           126528 bits [Y 46.6759 dB    U 51.8002 dB    V 50.9188 dB] [Y MSE 22.3670  U MSE 6.8735  V MSE 8.4202] [ET    18 ] [L0 31 35 ] [L1 35 39 ]
POC    32 TId: 5 (      STSA, B-SLICE, QP 18)            41296 bits [Y 51.0435 dB    U 55.4583 dB    V 54.4584 dB] [Y MSE 8.1818  U MSE 2.9606  V MSE 3.7270] [ET    11 ] [L0 31 33 ] [L1 33 35 ]
POC    34 TId: 5 (      STSA, B-SLICE, QP 19)           186176 bits [Y 49.4117 dB    U 54.7578 dB    V 54.0178 dB] [Y MSE 11.9133  U MSE 3.4787  V MSE 4.1250] [ET    16 ] [L0 33 31 ] [L1 35 39 ]
POC    37 TId: 4 (      STSA, B-SLICE, QP 22)           121208 bits [Y 47.7031 dB    U 53.6434 dB    V 52.6963 dB] [Y MSE 17.6561  U MSE 4.4964  V MSE 5.5920] [ET    18 ] [L0 35 31 ] [L1 39 47 ]
POC    36 TId: 5 (      STSA, B-SLICE, QP 19)           105128 bits [Y 49.4286 dB    U 55.4143 dB    V 54.6317 dB] [Y MSE 11.8669  U MSE 2.9907  V MSE 3.5812] [ET    16 ] [L0 35 31 ] [L1 37 39 ]
POC    38 TId: 5 (      STSA, B-SLICE, QP 19)           123008 bits [Y 49.4610 dB    U 56.0086 dB    V 55.2107 dB] [Y MSE 11.7787  U MSE 2.6082  V MSE 3.1342] [ET    14 ] [L0 37 35 ] [L1 39 47 ]
POC    43 TId: 3 (      STSA, B-SLICE, QP 22)           198232 bits [Y 47.4950 dB    U 52.2967 dB    V 51.4349 dB] [Y MSE 18.5226  U MSE 6.1310  V MSE 7.4766] [ET    19 ] [L0 39 31 ] [L1 47 39 ]
POC    41 TId: 4 (      STSA, B-SLICE, QP 22)           112728 bits [Y 48.3743 dB    U 53.1787 dB    V 52.8788 dB] [Y MSE 15.1277  U MSE 5.0041  V MSE 5.3620] [ET    16 ] [L0 39 31 ] [L1 43 47 ]
POC    40 TId: 5 (      STSA, B-SLICE, QP 19)            76760 bits [Y 50.1330 dB    U 55.9452 dB    V 55.4123 dB] [Y MSE 10.0901  U MSE 2.6466  V MSE 2.9921] [ET    10 ] [L0 39 31 ] [L1 41 43 ]
POC    42 TId: 5 (      STSA, B-SLICE, QP 19)           128344 bits [Y 50.1100 dB    U 55.4279 dB    V 55.2020 dB] [Y MSE 10.1437  U MSE 2.9813  V MSE 3.1405] [ET    12 ] [L0 41 39 ] [L1 43 47 ]
POC    45 TId: 4 (      STSA, B-SLICE, QP 25)           167864 bits [Y 45.8402 dB    U 49.8225 dB    V 49.4117 dB] [Y MSE 27.1131  U MSE 10.8382  V MSE 11.9131] [ET    33 ] [L0 43 39 ] [L1 47 43 ]
POC    44 TId: 5 (      STSA, B-SLICE, QP 22)           120128 bits [Y 48.3537 dB    U 52.7186 dB    V 51.8830 dB] [Y MSE 15.1995  U MSE 5.5634  V MSE 6.7437] [ET    15 ] [L0 43 39 ] [L1 45 47 ]
POC    46 TId: 5 (      STSA, B-SLICE, QP 22)           141816 bits [Y 47.9293 dB    U 51.9334 dB    V 51.4160 dB] [Y MSE 16.7599  U MSE 6.6659  V MSE 7.5092] [ET    12 ] [L0 45 43 ] [L1 47 45 ]

I've additionally uploaded the MCTF encode in the GDrive folder in the first post.