[HanaEncode] Kaoru Hana wa Rin to Saku - 07v2 (WEB Aliciä) / PT-BR / VOSTFR / 简繁日内封 / Sub Español / English-Sub / JPN / The Fragrant Flowers Bloom with Dignity / 花开凛凛满薰香 / 花開凜凜滿薰香 / 薰香花朵凛然绽放 / 薰香花朵凜然綻放 / 1080p / AV1 / (Weekly)

Subtitles

• Thanks to SubVision for their wonderful português brasileiro subtitle. It has a very lovely karaoke for OP and ED. You would not want to skip them if you're watching with their subtitle.
• Thanks to KHFR for their français subtitle. KHFR has a bustling Discord server specifically for this anime. You are welcome to join everyone here at https://discord.com/invite/wGTmFarxNY!
• Thanks to 拨雪寻春 for their 简繁日双语 subtitle. They put great effort into show, even meticulously recreated the title of the show in Chinese in ASS.
• Thanks to 北宇治字幕组 for their 简繁日双语 subtitle. They put a lot of care in song styling. Not only their OP and ED, but also their inserts fit the show very well.
• Thanks to DantalianSubs for their awesome español subtitle. You'll know this subtitle comes from DantalianSubs when you see carefully crafted dialogue text with a unique selection of fonts and colour, as well as fadein, fadeout when it's suitable. You are welcome to join their server at https://discord.com/invite/yg7kk6cSwK.
• Thanks to FLE for their polished English subtitle. The people doing typesetting there have spent a great amount of time refining every signs in the episode! Their subtitle also comes with a more than charming karaoke for the ED. You may even want to watch their ED karaoke even if your main language is not English. After you've enjoyed their subtitle, you're welcome to join their server at https://discord.gg/ZDswnwknHS.

Filtering and Encoding

This is a high effort filtering and encoding for Kaoru Hana wa Rin to Saku.
slow.pics compare: https://slow.pics/c/ftAw3VxA

For filtering, we first use vs-mlrt's anime-segmentation model to separate characters from background.
For characters, the Netflix source is pretty unbearable with terrible encoing artefacts. To hopefully remove the artefacts and return the character lines to what it should be, we use very strong vsdenoise.bm3d to denoise, followed by a medium strength AA. After that, we apply vsdehalo.edge_cleaner and vsdehalo.dehalo_sigma to remove as much halo as we can without damaging the white lustre in some cuts.
For background, we use vsdenoise.mc_degrain followed by a very conservative vsdenoise.bm3d to remove temporal noise. Then, we join with the character for the same AA. After that, we use strong vsdehalo.fine_dehalo for dehalo.

For encoding, we use Progression Boost Preset-Character-Boost-Butteraugli-Max to achieve a good quality across the whole episode.
We use 5fish's SVT-AV1-PSY v2.3.0-C --tune 3 for the encode.
Specifically for characters, we use Progression Boost's Character Boost feature to specifically boost the characters using ROI maps and --crf, and we use --complex-hvs 1 --psy-rd 1.5 --spy-rd 2 to hopefully retain all the fine lines. For high motion scenes, we use --filtering-noise-detection 4 to avoid artefacts.
For background details, we use --variance-boost-strength 3 --variance-octile 3 and we use strong --qp-scale-compress-strength 5.5 for better interframe consistency.

With Character Boost and various optimisations in Progression Boost, metric result is not very representative of the actually quality.
However, as a reference, this is the SSIMU2 score of episode 01 encode with --photon-noise removed compared against the filtered source:

• SSIMU2 Mean: 90.405
• SSIMU2 5th Percentile: 86.806

Thanks to all the people creating the tools, plugins and vsfunc's that make this possible, and thanks to everyone in the community for teaching me, pointing out issues, and giving me advice.

You are welcome to join AV1 weeb edition at https://discord.gg/83dRFDFDp7.

You are welcome to do anything with this encode, including remuxing and republishing it. However, we would ask you to leave a credit to HanaEncode if you can.

MediaInfo

English chapters recreated and included.

We recommend using MPV to playback our releases. Thank you for seeding after download.
Feedback is welcome via Marshmallow: https://marshmallow-qa.com/s8bbyicls38av1j.