[WOR] Please Put Them On, Takamine-san S01 [BD AV1 1080p FLAC] multi subs | Haite Kudasai, Takamine-san
Download Torrent
Opens in your torrent client (e.g. qBittorrent)
Health
Excellent72/8
Info Hashc806ded742316b096d43779c03a117c3e2e28891
Peers Updated11 hours ago (2026-03-23 22:27:13)
Description
Offerto da Winds Of Re:Sub
Please Put Them On, Takamine-san
Video: JPBD Encode AV1
Audio: FLAC 2.0 Japanese
Sub: Italian (basati su Crunchyroll, con restyle, typesetting e vari fix), English (CR), Brazilian Portuguese (CR), Latin American Spanish (CR), Spanish (CR), French (CR), German (CR), Russian (CR)
All subs, except italian, were sourced from this release
Technical Notes:
Filtering Process
Although the BD source is high quality, the presence of grain made filtering and encoding particularly tricky, requiring tight integration between the two to achieve a good result at a reasonable file size.
The main advantage for this anime was its native resolution (Bilinear 875.25p; however, there are multiple “correct” resolutions near 875p, as well as various “correct” shifts), which was relatively easy to determine. This allowed us to fix several issues like dirty lines and, most importantly, boost the contrast of details that had been obscured by the original grain. After rescaling and performing a complete denoise and degrain (carefully avoiding detail loss), it was necessary to apply antialiasing; since the source is pre-sharpened, some aliasing persisted even after rescaling (additionally, some haloing was present and subsequently removed).
Given the heavy grain, I chose to completely ignore potential banding and focus instead on matching that coarse, dynamic grain without distorting the original look. Given the situation, no dither was applied.
After numerous tests and parameter combinations in both Vaporsynth and AV1, we managed to strike the right balance:
We added a coarse, highly dynamic grain.
We applied a subsequent denoise pass to remove very high frequencies, smoothing it out to resemble the original.
Finally, we left the management of the remaining high frequencies to AV1 to improve compressibility.
Final Encode Parameters (SVT-AV1-PSYEX 3.0.2-B)
--preset 2 --crf 7 --low-q-taper 1 --qm-min 13 --chroma-qm-min 8 --keyint 240 --qp-scale-compress-strength 2 --scm 0 --tune 0 --psy-rd 1.5 --complex-hvs 1 --noise-norm-strength 0 --hbd-mds 1 --spy-rd 2 --enable-restoration 0 --enable-cdef 0 --enable-tf 0 --fast-decode 1 --luminance-qp-bias 20 --variance-boost-strength 1 --variance-octile 7 --photon-noise 2
With possible variations to CRF (7 -> 8) and qp-bias (20 -> 10) when the episode reached excessively large sizes (>2.5GB).
These parameters were chosen starting from a baseline created by BlueSwordM, whom I thank, and were subsequently fine-tuned to find a way to preserve the grain as much as possible, maintaining its original size and dynamism.
For every codec, dynamic grain is the single biggest factor hindering compression. We chose AV1 after testing various settings in x265 that failed to give us the expected results; x265 actually added high-frequency digital noise that was not present in the filtered source, exceeding even the noise found when using photon-noise 4 in AV1.
CRF 7 was a bold choice, born from an experiment that went well beyond expectations. The jump from CRF 9 -> 8 -> 7 offers enormous advantages for grain preservation, in addition to avoiding any bit starvation while remaining within an acceptable file size for a high-quality release.
Luma-bias: This was useful because the grain was applied in dark areas, where AV1 tends to assign less importance.
qm-min 13: This was another crucial parameter for achieving our objective.
However, the parameters usually related to grain preservation were not of much use in this specific case. This is likely because those standard parameters are designed for higher CRFs and aim for psychovisual preservation, whereas our goal was the exact preservation of the original grain structure.
Comparison
For any questions related to AV1, you can find @Man3500 on the AV1 Weeb Discord Server.
For any questions related to filtering, you can find @PingWer74 on the JET Discord Server.
I will not respond to comments here.
Please Put Them On, Takamine-san
Video: JPBD Encode AV1
Audio: FLAC 2.0 Japanese
Sub: Italian (basati su Crunchyroll, con restyle, typesetting e vari fix), English (CR), Brazilian Portuguese (CR), Latin American Spanish (CR), Spanish (CR), French (CR), German (CR), Russian (CR)
All subs, except italian, were sourced from this release
Technical Notes:
Filtering Process
Although the BD source is high quality, the presence of grain made filtering and encoding particularly tricky, requiring tight integration between the two to achieve a good result at a reasonable file size.
The main advantage for this anime was its native resolution (Bilinear 875.25p; however, there are multiple “correct” resolutions near 875p, as well as various “correct” shifts), which was relatively easy to determine. This allowed us to fix several issues like dirty lines and, most importantly, boost the contrast of details that had been obscured by the original grain. After rescaling and performing a complete denoise and degrain (carefully avoiding detail loss), it was necessary to apply antialiasing; since the source is pre-sharpened, some aliasing persisted even after rescaling (additionally, some haloing was present and subsequently removed).
Given the heavy grain, I chose to completely ignore potential banding and focus instead on matching that coarse, dynamic grain without distorting the original look. Given the situation, no dither was applied.
After numerous tests and parameter combinations in both Vaporsynth and AV1, we managed to strike the right balance:
We added a coarse, highly dynamic grain.
We applied a subsequent denoise pass to remove very high frequencies, smoothing it out to resemble the original.
Finally, we left the management of the remaining high frequencies to AV1 to improve compressibility.
Final Encode Parameters (SVT-AV1-PSYEX 3.0.2-B)
--preset 2 --crf 7 --low-q-taper 1 --qm-min 13 --chroma-qm-min 8 --keyint 240 --qp-scale-compress-strength 2 --scm 0 --tune 0 --psy-rd 1.5 --complex-hvs 1 --noise-norm-strength 0 --hbd-mds 1 --spy-rd 2 --enable-restoration 0 --enable-cdef 0 --enable-tf 0 --fast-decode 1 --luminance-qp-bias 20 --variance-boost-strength 1 --variance-octile 7 --photon-noise 2
With possible variations to CRF (7 -> 8) and qp-bias (20 -> 10) when the episode reached excessively large sizes (>2.5GB).
These parameters were chosen starting from a baseline created by BlueSwordM, whom I thank, and were subsequently fine-tuned to find a way to preserve the grain as much as possible, maintaining its original size and dynamism.
For every codec, dynamic grain is the single biggest factor hindering compression. We chose AV1 after testing various settings in x265 that failed to give us the expected results; x265 actually added high-frequency digital noise that was not present in the filtered source, exceeding even the noise found when using photon-noise 4 in AV1.
CRF 7 was a bold choice, born from an experiment that went well beyond expectations. The jump from CRF 9 -> 8 -> 7 offers enormous advantages for grain preservation, in addition to avoiding any bit starvation while remaining within an acceptable file size for a high-quality release.
Luma-bias: This was useful because the grain was applied in dark areas, where AV1 tends to assign less importance.
qm-min 13: This was another crucial parameter for achieving our objective.
However, the parameters usually related to grain preservation were not of much use in this specific case. This is likely because those standard parameters are designed for higher CRFs and aim for psychovisual preservation, whereas our goal was the exact preservation of the original grain structure.
Comparison
For any questions related to AV1, you can find @Man3500 on the AV1 Weeb Discord Server.
For any questions related to filtering, you can find @PingWer74 on the JET Discord Server.
I will not respond to comments here.