VMAF 97 - crf 15 preset 5

VMAF (Video Multi-Method Assessment Fusion) is like a video critic that watches your content and scores how close it feels to the original, on a scale of 0–100. Imagine you're showing off your video to an audience:
95–100: The audience applauds — the quality is so excellent, they can’t tell it’s not the original.
90–94: A few keen eyes notice tiny imperfections, but most still cheer — very good quality.
80–89: Some viewers nod, but others squint at the visible compression — good but not flawless.
Below 80: Murmurs of distraction as artifacts steal the spotlight — fair to poor quality.
If you want standing ovations, aim for 95+. For casual enjoyment, 90+ will do.





___________________

-Encoding settings-
Ab-av1
crf-search -i "c:\file.mkv" --pix-format yuv420p10le --svt fast-decode=1 --svt tune=0 --preset 5 --min-samples 20 --min-vmaf 99
VMAF 97.07 predicted video stream size 743.44 MiB (14%) taking 42 minutes

Handbrake
Video.svt-av1 10 bit fast decode crf 15 Preset 5
Audio.5.1 AC3 passstrough
___________________

-File info-
Video: AV01 1920x960 23.976fps [V: av1 main, yuv420p10le, 1920x960 [default]]
Audio: DD+ 48000Hz 6ch 768kbps [A: Surround [eng] (eac3 atmos, 48000 Hz, 5.1(side), 768 kb/s) [default]]
Subtitle: Advanced SubStation Alpha [S: English [eng] (ass)]
Subtitle: Advanced SubStation Alpha [S: English [SDH] [eng] (ass)]
Subtitle: Advanced SubStation Alpha [S: Arabic [ara] (ass)]
Subtitle: Advanced SubStation Alpha [S: Bulgarian [bul] (ass)]
Subtitle: Advanced SubStation Alpha [S: Chinese Simplified [chi] (ass)]
Subtitle: Advanced SubStation Alpha [S: Chinese Traditional [chi] (ass)]
Subtitle: Advanced SubStation Alpha [S: Chinese Yue [chi] (ass)]
Subtitle: Advanced SubStation Alpha [S: Czech [cze] (ass)]
Subtitle: Advanced SubStation Alpha [S: Danish [dan] (ass)]
Subtitle: Advanced SubStation Alpha [S: Dutch [dut] (ass)]
Subtitle: Advanced SubStation Alpha [S: Estonian [est] (ass)]
Subtitle: Advanced SubStation Alpha [S: European Spanish [spa] (ass)]
Subtitle: Advanced SubStation Alpha [S: Finnish [fin] (ass)]
Subtitle: Advanced SubStation Alpha [S: French [fre] (ass)]
Subtitle: Advanced SubStation Alpha [S: French Canadian [fre] (ass)]
Subtitle: Advanced SubStation Alpha [S: German [ger] (ass)]
Subtitle: Advanced SubStation Alpha [S: Greek [gre] (ass)]
Subtitle: Advanced SubStation Alpha [S: Hebrew [heb] (ass)]
Subtitle: Advanced SubStation Alpha [S: Hindi [hin] (ass)]
Subtitle: Advanced SubStation Alpha [S: Hungarian [hun] (ass)]
Subtitle: Advanced SubStation Alpha [S: Indonesian [ind] (ass)]
Subtitle: Advanced SubStation Alpha [S: Italian [ita] (ass)]
Subtitle: Advanced SubStation Alpha [S: Japanese [jpn] (ass)]
Subtitle: Advanced SubStation Alpha [S: Korean [kor] (ass)]
Subtitle: Advanced SubStation Alpha [S: Latin America Spanish [spa] (ass)]
Subtitle: Advanced SubStation Alpha [S: Latvian [lav] (ass)]
Subtitle: Advanced SubStation Alpha [S: Lithuanian [lit] (ass)]
Subtitle: Advanced SubStation Alpha [S: Malay [may] (ass)]
Subtitle: Advanced SubStation Alpha [S: Norwegian [nor] (ass)]
Subtitle: Advanced SubStation Alpha [S: Polish [pol] (ass)]
Subtitle: Advanced SubStation Alpha [S: Portuguese [por] (ass)]
Subtitle: Advanced SubStation Alpha [S: Portuguese Brazilian [por] (ass)]
Subtitle: Advanced SubStation Alpha [S: Russian [rus] (ass)]
Subtitle: Advanced SubStation Alpha [S: Slovak [slo] (ass)]
Subtitle: Advanced SubStation Alpha [S: Slovenian [slv] (ass)]
Subtitle: Advanced SubStation Alpha [S: Swedish [swe] (ass)]
Subtitle: Advanced SubStation Alpha [S: Tamil [tam] (ass)]
Subtitle: Advanced SubStation Alpha [S: Telugu [tel] (ass)]
Subtitle: Advanced SubStation Alpha [S: Thai [tha] (ass)]
Subtitle: Advanced SubStation Alpha [S: Turkish [tur] (ass)]
Subtitle: Advanced SubStation Alpha [S: Ukrainian [ukr] (ass)]
Subtitle: Advanced SubStation Alpha [S: Vietnamese [vie] (ass)]
Subtitle: UTF-8 [S: No subtitles]

General
Unique ID                      : 68899225737765859000956159629160909384 (0x33D582195A327E4D9518ADC22C948A48)
Complete name                  : Shrinking.s02e03.web.apple.sdr.tv.1080p.av1.5.1.e-ac3.vmaf97-Rosy.mkv
Format                         : Matroska
Format version                 : Version 4
File size                      : 791 MiB
Duration                       : 34 min 21 s
Overall bit rate               : 3 218 kb/s
Frame rate                     : 23.976 FPS
Encoded date                   : 2024-10-23 16:34:09 UTC
Writing application            : HandBrake 1.8.2 2024081000
Writing library                : Lavf61.1.100
ErrorDetectionType             : Per level 1

Video
ID                             : 1
Format                         : AV1
Format/Info                    : AOMedia Video 1
Format profile                 : Main@L4.0
Codec ID                       : V_AV1
Duration                       : 34 min 21 s
Bit rate                       : 2 386 kb/s
Width                          : 1 920 pixels
Height                         : 960 pixels
Display aspect ratio           : 2.000
Frame rate mode                : Constant
Frame rate                     : 23.976 (24000/1001) FPS
Color space                    : YUV
Chroma subsampling             : 4:2:0 (Type 0)
Bit depth                      : 10 bits
Bits/(Pixel*Frame)             : 0.054
Stream size                    : 586 MiB (74%)
Default                        : Yes
Forced                         : No
Color range                    : Limited
Color primaries                : BT.709
Transfer characteristics       : BT.709
Matrix coefficients            : BT.709

Audio
ID                             : 2
Format                         : E-AC-3 JOC
Format/Info                    : Enhanced AC-3 with Joint Object Coding
Commercial name                : Dolby Digital Plus with Dolby Atmos
Codec ID                       : A_EAC3
Duration                       : 34 min 21 s
Bit rate mode                  : Constant
Bit rate                       : 768 kb/s
Channel(s)                     : 6 channels
Channel layout                 : L R C LFE Ls Rs
Sampling rate                  : 48.0 kHz
Frame rate                     : 31.250 FPS (1536 SPF)
Compression mode               : Lossy
Stream size                    : 189 MiB (24%)
Title                          : Surround
Language                       : English
Service kind                   : Complete Main
Default                        : Yes
Forced                         : No
Complexity index               : 16
Number of dynamic objects      : 15
Bed channel count              : 1 channel
Bed channel configuration      : LFE
Dialog Normalization           : -26 dB
compr                          : -0.28 dB
dialnorm_Average               : -26 dB
dialnorm_Minimum               : -26 dB
dialnorm_Maximum               : -26 dB
___________________

-Source file info-
Video: HVC1 3840x1920 23.976fps [V: English [eng] (hevc main 10 L5.0, yuv420p10le, 3840x1920) [default]]
Audio: DD+ 48000Hz 6ch 768kbps [A: English [eng] (eac3 atmos, 48000 Hz, 5.1(side), 768 kb/s) [default]]
Subtitle: UTF-8 [S: English [eng] (subrip)]
Subtitle: UTF-8 [S: English [SDH] [eng] (subrip)]
Subtitle: UTF-8 [S: Arabic [ara] (subrip)]
Subtitle: UTF-8 [S: Bulgarian [bul] (subrip)]
Subtitle: UTF-8 [S: Chinese Simplified [chi] (subrip)]
Subtitle: UTF-8 [S: Chinese Traditional [chi] (subrip)]
Subtitle: UTF-8 [S: Chinese Yue [chi] (subrip)]
Subtitle: UTF-8 [S: Czech [cze] (subrip)]
Subtitle: UTF-8 [S: Danish [dan] (subrip)]
Subtitle: UTF-8 [S: Dutch [dut] (subrip)]
Subtitle: UTF-8 [S: Estonian [est] (subrip)]
Subtitle: UTF-8 [S: European Spanish [spa] (subrip)]
Subtitle: UTF-8 [S: Finnish [fin] (subrip)]
Subtitle: UTF-8 [S: French [fre] (subrip)]
Subtitle: UTF-8 [S: French Canadian [fre] (subrip)]
Subtitle: UTF-8 [S: German [ger] (subrip)]
Subtitle: UTF-8 [S: Greek [gre] (subrip)]
Subtitle: UTF-8 [S: Hebrew [heb] (subrip)]
Subtitle: UTF-8 [S: Hindi [hin] (subrip)]
Subtitle: UTF-8 [S: Hungarian [hun] (subrip)]
Subtitle: UTF-8 [S: Indonesian [ind] (subrip)]
Subtitle: UTF-8 [S: Italian [ita] (subrip)]
Subtitle: UTF-8 [S: Japanese [jpn] (subrip)]
Subtitle: UTF-8 [S: Korean [kor] (subrip)]
Subtitle: UTF-8 [S: Latin America Spanish [spa] (subrip)]
Subtitle: UTF-8 [S: Latvian [lav] (subrip)]
Subtitle: UTF-8 [S: Lithuanian [lit] (subrip)]
Subtitle: UTF-8 [S: Malay [may] (subrip)]
Subtitle: UTF-8 [S: Norwegian [nor] (subrip)]
Subtitle: UTF-8 [S: Polish [pol] (subrip)]
Subtitle: UTF-8 [S: Portuguese [por] (subrip)]
Subtitle: UTF-8 [S: Portuguese Brazilian [por] (subrip)]
Subtitle: UTF-8 [S: Russian [rus] (subrip)]
Subtitle: UTF-8 [S: Slovak [slo] (subrip)]
Subtitle: UTF-8 [S: Slovenian [slv] (subrip)]
Subtitle: UTF-8 [S: Swedish [swe] (subrip)]
Subtitle: UTF-8 [S: Tamil [tam] (subrip)]
Subtitle: UTF-8 [S: Telugu [tel] (subrip)]
Subtitle: UTF-8 [S: Thai [tha] (subrip)]
Subtitle: UTF-8 [S: Turkish [tur] (subrip)]
Subtitle: UTF-8 [S: Ukrainian [ukr] (subrip)]
Subtitle: UTF-8 [S: Vietnamese [vie] (subrip)]
Subtitle: UTF-8 [S: No subtitles]

General
Unique ID                      : 299600148922176704468824706029101381807 (0xE164E3CAEDF870C63212439D0F85F8AF)
Complete name                  : shrinking.s02e03.2160p.web.h265-successfulcrab.mkv
Format                         : Matroska
Format version                 : Version 4
File size                      : 5.17 GiB
Duration                       : 34 min 21 s
Overall bit rate               : 21.5 Mb/s
Frame rate                     : 23.976 FPS
Writing application            : mkvmerge v84.0 ('Sleeper') 64-bit
Writing library                : libebml v1.4.5 + libmatroska v1.7.1

Video
ID                             : 1
Format                         : HEVC
Format/Info                    : High Efficiency Video Coding
Format profile                 : Main 10@L5@High
Codec ID                       : V_MPEGH/ISO/HEVC
Duration                       : 34 min 21 s
Bit rate                       : 20.8 Mb/s
Width                          : 3 840 pixels
Height                         : 1 920 pixels
Display aspect ratio           : 2.000
Frame rate mode                : Constant
Frame rate                     : 23.976 FPS
Color space                    : YUV
Chroma subsampling             : 4:2:0
Bit depth                      : 10 bits
Bits/(Pixel*Frame)             : 0.117
Stream size                    : 4.98 GiB (96%)
Language                       : English
Default                        : Yes
Forced                         : No
Color range                    : Limited
Color primaries                : BT.709
Transfer characteristics       : BT.709
Matrix coefficients            : BT.709

Audio
ID                             : 2
Format                         : E-AC-3 JOC
Format/Info                    : Enhanced AC-3 with Joint Object Coding
Commercial name                : Dolby Digital Plus with Dolby Atmos
Codec ID                       : A_EAC3
Duration                       : 34 min 21 s
Bit rate mode                  : Constant
Bit rate                       : 768 kb/s
Channel(s)                     : 6 channels
Channel layout                 : L R C LFE Ls Rs
Sampling rate                  : 48.0 kHz
Frame rate                     : 31.250 FPS (1536 SPF)
Compression mode               : Lossy
Stream size                    : 189 MiB (4%)
Language                       : English
Service kind                   : Complete Main
Default                        : Yes
Forced                         : No
Complexity index               : 16
Number of dynamic objects      : 15
Bed channel count              : 1 channel
Bed channel configuration      : LFE
Dialog Normalization           : -26 dB
compr                          : -0.28 dB
dialnorm_Average               : -26 dB
dialnorm_Minimum               : -26 dB
dialnorm_Maximum               : -26 dB
___________________

-Tools used-
Ab-av1
Av1 guide v4 (check nfo file)
FFBitrateViewer
FFmpeg
FFmetrics
Freefilesync
Handbrake
Jackett
Mkvtoolknix
Qbittorent
Scenerulez
Shutterencoder
VVV-virtualvolumeview
Xnconvert
Xnview mp


===================

AV1 Short recipe

===================
Handbrake
crf    20-25
preset 5-7
fast-decode


===================

AV1 guide v4

===================

Step 1: Check VMAF Results I recommend "ab-av1"(1) but you can also check with ffmpeg, ffmetrics or shutterencoder.

1. 2160p Source:
  AV1:
  crf-search -i "c:\file.mkv" --pix-format yuv420p10le --svt fast-decode=1 --svt tune=0 --preset 5 --min-samples 20 --min-vmaf 98

  1080p Downscale (AV1):
  crf-search -i "c:\file.mkv" --pix-format yuv420p10le --svt fast-decode=1 --svt tune=0 --preset 5 --min-samples 20 --min-vmaf 96 --vfilter "scale=-1:1080"

  720p Downscale (AV1):
  crf-search -i "c:\file.mkv" --pix-format yuv420p10le --svt fast-decode=1 --svt tune=0 --preset 5 --min-samples 20 --min-vmaf 94 --vfilter "scale=-1:720"

2. x265 Encoding:
  2160p Source:
  crf-search -i "c:\file.mkv" --encoder libx265 --pix-format yuv420p10le --preset medium --min-samples 20 --enc x265-params=fastdecode=1 --min-vmaf 98

  1080p Downscale (x265):
  crf-search -i "c:\file.mkv" --encoder libx265 --pix-format yuv420p10le --preset medium --min-samples 20 --enc x265-params=fastdecode=1 --min-vmaf 96 --vfilter "scale=-1:1080"

  720p Downscale (x265):
  crf-search -i "c:\file.mkv" --encoder libx265 --pix-format yuv420p10le --preset medium --min-samples 20 --enc x265-params=fastdecode=1 --min-vmaf 94 --vfilter "scale=-1:720"

===================

Step 2: Encode Using HandBrake

Summary:
- Container: MKV

Dimensions:
- Cropping: None
- Anamorphic: None
- Borders: None

Video:
- Encoder: AV1 (10-bit SVT)
- Framerate: Same as source
- Preset: 5
- Fast Decode: ✓
- Encoder Profile: Main
- Encoder Level: Auto
- Quality: CRF determined from "ab-av1"(1)

Audio:
- DTS to Opus Conversion:
 - 7.1 → Opus 512 kbps
 - 5.1 → AC3  passtrough or opus 384 kbps
 - 2.0 → AAC  160 kbps or opus 128 kbps

Subtitles:
- Original
- Subtitle Translator
- Whisper OpenAI
- No subtitles (if preferred)

===================

Step 3: only if you want

Open mkvtoolnix and make global tags, with xml file from this:


<?xml version="1.0" encoding="ISO-8859-1"?>
<!DOCTYPE Tags SYSTEM "matroskatags.dtd">
<Tags>
 <!-- movie -->
 <Tag>
   <Targets>
     <TargetTypeValue>50</TargetTypeValue>
   </Targets>
   <Simple>
     <Name>TITLE</Name>
     <String>Dune</String>
   </Simple>
   <Simple>
     <Name>DIRECTOR</Name>
     <String>David Lynch</String>
   </Simple>
   <Simple>
     <Name>DATE_RELEASED</Name>
     <String>1984</String>
   </Simple>
   <Simple>
     <Name>COMMENT</Name>
     <String>Great sci-fi movie</String>
   </Simple>
 </Tag>
</Tags>


===================

General VMAF Estimates by CRF after several encodes:

1. 2160p Remux (Movie):
  VMAF 96-98 | CRF 15 | Preset 5
2. 2160p Remux (Animation):
  VMAF 98-99 | CRF 15 | Preset 5
3. 1080p Remux (Movie):
  VMAF 96-97 | CRF 15 | Preset 5
4. 1080p Remux (Animation):
  VMAF 97-98 | CRF 15 | Preset 5

5. 2160p Web-DL (Movie):
  VMAF 96-98 | CRF 25 | Preset 5
6. 2160p Web-DL (Animation):
  VMAF 98-99 | CRF 25 | Preset 5
7. 1080p Web-DL (Movie):
  VMAF 96-97 | CRF 25 | Preset 5
8. 1080p Web-DL (Animation):
  VMAF 97-98 | CRF 25 | Preset 5

===================

References:
1. Dolby, 2023. What are Dolby Vision profiles and levels? [Online] Available at: https://professionalsupport.dolby.com/s/article/What-is-Dolby-Vision-Profile?language=en_US (Accessed 18 August 2024).
2. Ozer, J., 2017. Video Encoding by the Numbers: Eliminate the Guesswork from your Streaming Video. Doceo Publishing. [Available on Amazon](https://www.amazon.com.br/dp/0998453005) (Accessed 18 August 2024).
3. Sayood, K., 2017. Introduction to Data Compression. 5th Edition.
4. Ozer, J., 2021. What is CBR, VBR, CRF, Capped-CRF? Rate Control Modes Explained. Available at: https://ottverse.com/what-is-cbr-vbr-crf-capped-crf-rate-control-explained/ (Accessed 05 August 2024).

===================

→ Why encode?

We often have too many files and insufficient infrastructure to store or manage all of them efficiently. Encoding helps save storage space, making it easier to share files faster while still maintaining an acceptable quality level. By encoding, you can reduce file sizes significantly without noticeable quality loss.

→ Why AV1?

AV1 is royalty-free, meaning it doesn’t come with licensing fees like many proprietary formats do. As a free and open standard, it has better long-term compatibility and sustainability compared to proprietary formats. AV1 also offers improved compression efficiency, meaning it can achieve better quality at lower bitrates, further optimizing storage and bandwidth usage.

→ Why MKV?

MKV (Matroska) is also royalty-free, making it a versatile and future-proof container format. Like AV1, being an open format allows it to stand the test of time in terms of compatibility, especially when compared to proprietary formats. MKV supports a wide range of codecs, making it a popular choice for high-quality video encoding.

→ What is VMAF?

VMAF (Video Multi-Method Assessment Fusion) is a video quality metric developed by Netflix. It combines different metrics and algorithms to predict perceived video quality from a viewer's perspective, providing a score between 0 and 100. It’s more advanced than older metrics like SSIM (Structural Similarity Index) and PSNR (Peak Signal-to-Noise Ratio), which often don't reflect human perception as accurately.

→ Why not SSIM, PSNR, SSIMULACRA, or even just judging by eye?

- SSIM focuses primarily on structural distortions but may not always correlate well with human perception.
- PSNR measures the difference between the original and compressed images in terms of pixel-by-pixel accuracy, but it tends to favor sharpness over subjective quality.
- SSIMULACRA is another quality metric focused on visual fidelity but is less widely adopted.
- Eye-pleasing judgment is highly subjective and inconsistent, especially when encoding content at scale, making it unreliable for technical purposes.

VMAF is considered superior because it more closely approximates how humans perceive video quality, making it the preferred choice for fine-tuning encoding processes.

→ Why VMAF?

There are several approaches you can take to encode video, depending on your priorities:

- CBR (Constant Bitrate): Keeps file sizes consistent by maintaining a fixed bitrate, but can result in unnecessary quality loss or bloated file sizes.
- VBR (Variable Bitrate): Adjusts bitrate dynamically to optimize size and quality, but requires more complex processing.
- CRF (Constant Rate Factor): Ensures constant quality without needing multiple encoding passes, offering a good balance between quality and size.
- CRF Capped: Similar to CRF, but with a file size limit constraint, ensuring the file doesn't exceed a certain size while maintaining decent quality.
- VMAF-Tuned Encoding: Uses CRF but optimizes for a specific VMAF score, ensuring that the encoded video meets a particular quality threshold according to the VMAF metric.
- VMAF Capped: This option targets a specific VMAF score for video quality while imposing a file size limit. It allows you to maintain a high-quality standard (measured by VMAF) but ensures the final file size stays within defined boundaries, striking a balance between quality and size constraints.

VMAF-guided encoding, whether capped or uncapped, allows you to ensure that your files achieve a consistent, high-quality viewing experience while optimizing for storage space and bandwidth.

===================

Do you have any recommendations? Is there something you disagree with? Feel free to share your thoughts in the comments section.