Honzuki no Gekokujou: Shisho ni Naru Tame ni wa Shudan wo Erandeiraremasen (Ascendance of a Bookworm)

Series

SavI0r	Episodes	MyAnimeList	TVDB	Source
Season 1	14 (S01E01-S01E14)	Honzuki no Gekokujou: Shisho ni Naru Tame ni wa Shudan wo Erandeiraremasen	Ascendance of a Bookworm Season 1	sam, MTBB, Erai-Raws, 千夏字幕组&喵萌奶茶屋&VCB-Studio, Kitsunekko, Fuyu
OVA	1 (S00E01-S00E02)	Honzuki no Gekokujou: Shisho ni Naru Tame ni wa Shudan wo Erandeiraremasen OVA	Ascendance of a Bookworm OVAs	sam, MTBB, Erai-Raws, 千夏字幕组&喵萌奶茶屋&VCB-Studio, Fuyu
Season 2	12 (S01E15-S01E26)	Honzuki no Gekokujou: Shisho ni Naru Tame ni wa Shudan wo Erandeiraremasen 2nd Season	Ascendance of a Bookworm Season 1	sam, MTBB, Erai-Raws, Kitsunekko, Fuyu
Season 3	10 (S01E27-S01E36)	Honzuki no Gekokujou: Shisho ni Naru Tame ni wa Shudan wo Erandeiraremasen 3rd Season	Ascendance of a Bookworm Season 1	sam, Erai-Raws, Fuyu

Release

An up-to-date Dual Audio release encoded in AV1 and OPUS for optimum size and quality. This release is not considered complete as the next season is not available yet and recap episodes are not included. Files are named to support media servers like Plex using TVDB metadata.

Encoding Details:

Encoded with our SWAN system to target a SSIMULACRA2 score of 85.

Here is a non-exhaustive list of things we have discovered or developed since our first release:

• Developed batch multiplexing with heuristics-based track selection and modification (Translation: We automated muxing multiple sources together based on what’s in them)
• Switched from AOM to faster SVT
• Encode faster without film grain synthesis and instead add photon-noise with Grav1synth or Av1an
• Grav1synth corrupts SVT bitstreams with temporal filtering
• Developed scripts to generate keyframes and zones files from parsing Av1an scenes JSON to navigate scenes in vs-preview
• Manually optimized every scene in WIXOSS release
• BestSource and DGDecNV obsolete L-SMASH-Works, which caused issues with numerous encodes including the previous of this one
• Discovered variable framerate support for Av1an
• LastBreeze contributes to Cross Ange release with Descaling/Rescaling with Waifu2x
• Lumen produces French subtitle files with OCR for Cross Ange release
• Ironclad produces rescaling guide
• Trix writes insightful articles on the Codec Wiki analyzing SVT
• Switched to SVT-AV1-Psy fork using new features (Now it’s better too!)
• Julek makes objective image metric SSIMULACRA2 available in VapourSynth
• Developed the ssimulacra.py CLI tool (MIT licensed) to easily and quickly measure encode quality with SSIMULACRA2
• Developed Auto-Boost competitor: Normal-Boost (NB)
• Developed Triple Av1an Bypass (TAB) to completely overhaul and speed up NB by obsoleting the “fast-pass”
• SVT-AV1-Psy fork features ported to mainline
• Ironclad develops grav1an.py all-in-one encoding script with features like target-quality, inspiring SWAN
• Introducing the most advanced Av1an GUI: Av1ation-Station. It’s free, open source (MIT), and designed for novices and experts alike. All releases without NB/TAB or SWAN were encoded with it since its inception. Surprise!
• Gui-Yom develops TurboMetrics to measure SSIMULACRA2 10x faster with CUDA
• Kosaka develops SSIMULACRApy, one of a few forks of our ssimulacra.py tool
• Rescaling with ArtCNN 10x faster
• LastBreeze develops scripts for determining the CRF needed per “scene” using SSIMULACRA2, inspiring SWAN
• Developed Scene-Weighted Av1an Normalizer: SWAN (TLDR: Every individual scene is automatically optimized to hit a target SSIMULACRA2 score)
• Introducing our successor to ssimulacra.py: Media-Metrologist
• Lumen develops Vapoursynth-HIP plugin to measure SSIMULACRA2 and Butteraugli with hardware acceleration
• MA0MA0 develops Grav1synth fix for bitstream corruption

SVT-AV1-Psy 2.3.0-B with Av1an (--photon-noise 2)

--preset 4 --crf 28 --color-primaries 1 --transfer-characteristics 1 --matrix-coefficients 1 --keyint 0 --aq-mode 2 --tune 2 --kf-tf-strength 1 --enable-qm 1 --qm-min 0 --qm-max 15 --variance-octile 6 --variance-boost-strength 2 --enable-dlf 2 --qp-scale-compress-strength 3 --noise-norm-strength 3 --spy-rd 1

This project is a production with the goal of the smallest video size with moderate subjective compromises to image quality. This is not the hallmark of AV1, but close enough. This release should only contain changes to the video stream and audio tracks. The subtitle and attachment tracks are usually copied from the noted sources. Timings are not adjusted except in cases of extreme delay as a result of mixing sources. - SavI0r

Comparisons

The header above links to a comparison of S01E25 between the source, our previous AOM AV1, and our resulting AV1. Below are a few comparisons between them.

sam (HEVC)	Sav1or (AOM)	AV1 (SVT)

Source

Thanks to the following sources:

• sam - S01, S02, S03 Video, Audio, Chapter, S03 Subtitle
• MTBB - S01, S02 Subtitle
• Erai-raws - S01, S02, S03 AR, DE, ES, ES-LA, FR, IT, PT-BR, RU Subtitle
• 千夏字幕组&喵萌奶茶屋&VCB-Studio - S01 zh-Hans, zh-Hant Subtitle
• Kitsunekko - S01-S02 JA Subtitle
• Fuyu - VI Subtitle

Media Info