Kumo Desu ga, Nani ka? | So I’m a Spider, So What? - MyAnimeList.net |
To achieve higher precision audio sync I had to encode the english audio from AAC to OPUS (read the comments, hertz). OP and ED subtitles of Ember’s release were added to the EXTRA’s. Also fixed chapters for EP 14. Episodes contain a lot of fonts and heavy subtitle styling, so consider using mpv for smooth playback.
Summary:
Track-Type | Title | Attributes |
---|---|---|
video | Japanese | x265 HEVC 8bit 23.976FPS 1920x1080 |
audio | Japanese | 48.0kHz stereo OPUS |
audio | English | 48.0kHz stereo OPUS (default) |
subtitles | Dialogue | ass |
subtitles | Sign&Songs | ass (default) |
chapters |
BDMV - video, audio (jpn)
Golumpa - audio (eng)
Ember - subtitles (english, sign&songs), chapters
Screenshot 1 | Screenshot 2 | Screenshot 3 | Screenshot 4 |
---|---|---|---|
ffprobe version 4.4 Copyright (c) 2007-2021 the FFmpeg developers
built with gcc 11.1.0 (GCC)
configuration: ...
libavutil 56. 70.100 / 56. 70.100
libavcodec 58.134.100 / 58.134.100
libavformat 58. 76.100 / 58. 76.100
libavdevice 58. 13.100 / 58. 13.100
libavfilter 7.110.100 / 7.110.100
libavresample 4. 0. 0 / 4. 0. 0
libswscale 5. 9.100 / 5. 9.100
libswresample 3. 9.100 / 3. 9.100
libpostproc 55. 9.100 / 55. 9.100
Input #0, matroska,webm, from 'So I'm a Spider, So What (2021) - S01E01 - Reincarnation, in Another World.mkv':
Metadata:
ENCODER : Lavf58.76.100
Duration: 00:23:43.45, start: -0.007000, bitrate: 2425 kb/s
Chapters:
Chapter #0:0: start 0.000000, end 125.120000
Metadata:
title : Intro
Chapter #0:1: start 125.120000, end 215.190000
Metadata:
title : OP
Chapter #0:2: start 215.190000, end 652.630000
Metadata:
title : Part A
Chapter #0:3: start 652.630000, end 1326.320000
Metadata:
title : Part B
Chapter #0:4: start 1326.320000, end 1416.390000
Metadata:
title : ED
Chapter #0:5: start 1416.390000, end 1421.502000
Metadata:
title : PV
Stream #0:0(jpn): Video: hevc (Main), yuv420p(tv, bt709, progressive), 1920x1080 [SAR 1:1 DAR 16:9], 23.98 fps, 23.98 tbr, 1k tbn, 23.98 tbc (default)
Metadata:
title : [BDMV] Japanese
ENCODER : Lavc59.10.100 libx265
BPS : 2108795
DURATION : 00:23:43.435708333
NUMBER_OF_FRAMES: 34128
NUMBER_OF_BYTES : 375216686
_STATISTICS_WRITING_APP: mkvpropedit v62.0.0 ('Apollo') 64-bit
_STATISTICS_WRITING_DATE_UTC: 2021-11-06 15:26:51
_STATISTICS_TAGS: BPS DURATION NUMBER_OF_FRAMES NUMBER_OF_BYTES
Stream #0:1(jpn): Audio: opus, 48000 Hz, stereo, fltp
Metadata:
title : [BDMV] Japanese
ENCODER : Lavc59.10.100 libopus
BPS : 93212
DURATION : 00:23:43.421000000
NUMBER_OF_FRAMES: 71172
NUMBER_OF_BYTES : 16585040
_STATISTICS_WRITING_APP: mkvpropedit v62.0.0 ('Apollo') 64-bit
_STATISTICS_WRITING_DATE_UTC: 2021-11-06 15:26:51
_STATISTICS_TAGS: BPS DURATION NUMBER_OF_FRAMES NUMBER_OF_BYTES
Stream #0:2(eng): Audio: opus, 48000 Hz, stereo, fltp (default)
Metadata:
title : [Golumpa] English
ENCODER : Lavc58.134.100 libopus
BPS : 91652
DURATION : 00:23:42.461000000
NUMBER_OF_FRAMES: 71124
NUMBER_OF_BYTES : 16296515
_STATISTICS_WRITING_APP: mkvpropedit v62.0.0 ('Apollo') 64-bit
_STATISTICS_WRITING_DATE_UTC: 2021-11-06 15:26:51
_STATISTICS_TAGS: BPS DURATION NUMBER_OF_FRAMES NUMBER_OF_BYTES
Stream #0:3(eng): Subtitle: ass
Metadata:
title : [Ember] Dialogue
ENCODER : Lavc58.134.100 ass
BPS : 13151
DURATION : 00:23:40.960000000
NUMBER_OF_FRAMES: 7958
NUMBER_OF_BYTES : 2336007
_STATISTICS_WRITING_APP: mkvpropedit v62.0.0 ('Apollo') 64-bit
_STATISTICS_WRITING_DATE_UTC: 2021-11-06 15:26:51
_STATISTICS_TAGS: BPS DURATION NUMBER_OF_FRAMES NUMBER_OF_BYTES
Stream #0:4(eng): Subtitle: ass (default)
Metadata:
title : [Ember] Sign&Songs
ENCODER : Lavc58.134.100 ass
BPS : 14233
DURATION : 00:21:36.310000000
NUMBER_OF_FRAMES: 7563
NUMBER_OF_BYTES : 2306338
_STATISTICS_WRITING_APP: mkvpropedit v62.0.0 ('Apollo') 64-bit
_STATISTICS_WRITING_DATE_UTC: 2021-11-06 15:26:51
_STATISTICS_TAGS: BPS DURATION NUMBER_OF_FRAMES NUMBER_OF_BYTES
Stream #0:5: Attachment: ttf
Metadata:
filename : CorporateLogoBold.ttf
mimetype : application/x-truetype-font
Stream #0:6: Attachment: ttf
Metadata:
filename : DJB Chalk It Up.ttf
mimetype : application/x-truetype-font
Stream #0:7: Attachment: ttf
Metadata:
filename : Fairymuffinroundpop-1GG5e.ttf
mimetype : application/x-truetype-font
Stream #0:8: Attachment: ttf
Metadata:
filename : FBMillerDisplayBold.ttf
mimetype : application/x-truetype-font
Stream #0:9: Attachment: ttf
Metadata:
filename : Figure writing.ttf
mimetype : application/x-truetype-font
Stream #0:10: Attachment: otf
Metadata:
filename : FOT-NewRodin ProN B.otf
mimetype : application/vnd.ms-opentype
Stream #0:11: Attachment: otf
Metadata:
filename : FOT-OedKtr Std E.otf
mimetype : application/vnd.ms-opentype
Stream #0:12: Attachment: otf
Metadata:
filename : GANDHISANS-BOLD.OTF
mimetype : application/vnd.ms-opentype
Stream #0:13: Attachment: otf
Metadata:
filename : GANDHISANS-BOLDITALIC.OTF
mimetype : application/vnd.ms-opentype
Stream #0:14: Attachment: ttf
Metadata:
filename : rune.ttf
mimetype : application/x-truetype-font
Stream #0:15: Attachment: ttf
Metadata:
filename : seguisym.ttf
mimetype : application/x-truetype-font
Stream #0:16: Attachment: ttf
Metadata:
filename : solid-speech-bubbles-paul-tos.ttf
mimetype : application/x-truetype-font
Stream #0:17: Attachment: otf
Metadata:
filename : 1579-AVENIRLTSTD-MEDIUM.OTF
mimetype : application/vnd.ms-opentype
Stream #0:18: Attachment: ttf
Metadata:
filename : Another Danger - Demo.otf
mimetype : application/x-truetype-font
Stream #0:19: Attachment: ttf
Metadata:
filename : ArchivoNarrow-Bold.ttf
mimetype : application/x-truetype-font
Stream #0:20: Attachment: ttf
Metadata:
filename : arial.ttf
mimetype : application/x-truetype-font
Stream #0:21: Attachment: ttf
Metadata:
filename : Avenir 85 Heavy.ttf
mimetype : application/x-truetype-font
Stream #0:22: Attachment: ttf
Metadata:
filename : Caramella_Regular.ttf
mimetype : application/x-truetype-font
If you find anything that seems out of the ordinary, just post a comment or in my discord and I’ll take a look at it.
Looks like NTFS doesn’t like the question-mark my folder name, sorry about that. Will keep an eye on that in the future.
Comments - 21
friendevil22
Thanks!
Mr-EfTi
💕ArigAtouuuu💕gozAimAsuuuu💕
DrWolfnstine13
Thanks!
Simplistic
What?
ThatWaxEllo (uploader)
@Simplistic
ffmpeg’s delaying is not as accurate, if you just copy the stream. So instead of encoding aac to aac, I went with the highest possible codec, while keeping the file size acceptable. Tips for better execution are appreciated.
Simplistic
Use MKV container delays, they seem to be accurate enough, but the downside is you lose the delays if you ever demux later. I know there’s a way to losslessly delay streams, but I haven’t really learned how to do so myself yet.
ThatWaxEllo (uploader)
@Simplistic
That does actually sound like a very good idea. Thank you a lot!
Simplistic
Use container delays when losslessly shifting the audio doesn’t work out for you. I think eac3to can do it during the conversion process so that only helps if you start with a lossless track, but I might be wrong (I recommend it for all kinds of audio encoding really).
EDIT: If you go with the container delay option I recommend noting the delays used and listing them somewhere so anyone who needs them later doesn’t have to find out for themselves.
Otaskai
thanks a lot
Xota
Any reason why it’s 8 bit instead of a 10 bit HEVC encode? Thanks!
Simplistic
Actually after some experimenting if the delay is negative, it’s permanently deleted from the audio track (which I guess is what this warning is trying to say). But, they’re still not accurately, meaning the results end up being slightly off but small enough that it’s not noticeable. If it’s positive it’s not deleted, meaning positive delays will be lost during remuxing but not negative ones. Not sure how I never noticed this. Delays made with eac3to are permanent and seem to be autisticly accurate, though.
Image example
Well as we know making cuts usually isn’t accurate, so I’m not too surprised.
ThatWaxEllo (uploader)
@Simplistic
Thank you very much for the valuable findings. Crazy to think there isn’t any non-destructive accurate way to do this. !More attention to audio instead of H.266 and what not! Looking at your screenshots, I would actually say that MKV container-delays don’t seem to be good enough especially when it comes to audio. Of course this depends on the file. Thank you again for the follow-up and explaining it to me. :)
@Xota
The source video BDMV is 8-bit, so there was nothing I could do. Mediainfo - privatebin.at
Simplistic
All BDMVs are, but we usually encode 10-bit HEVC anyway, It helps with something but I forget what.
Yeah, I don’t really understand audio enough to say how you’d make the cuts accurately other than eac3to or audacity (losslessly anyway), but I don’t know how to get eac3to to take opus files as source files due to outdated codecs. Maybe someone with a big brain could help us with this. Also, I think it’s good enough, considering I’m only just now noticing that It’s slightly off.
RoxReaper
Here is an explanation from Daiz
Daiz being, alongside other fansubbers, one of the earlier adopters of hi10(x264).
fuee
this release looks bad i think? Am i wrong or just retarded?
ThatWaxEllo (uploader)
@Simplistic/@RoxReaper
Seems like a weird practice trying to create ‘better’ colors out of nothing, but thanks to the post, it actually makes sense. Thank you, noted.
@fuee
Could you please be a bit more clear. If you are referring to the drawing quality maybe it relates to this comment, posted on the BDMV release.
Simplistic
Yeah, I figured it was banding.
Simplistic
Actually, if this is to be believed, then it should be accurate? idk though there’s a 70% chance I’m misunderstanding it. All I know is it sounds fine to my ears and eyes.
hertz
Epic!
ParagonOxy
You converted a lossy audio source to another lossy audio format?
ThatWaxEllo (uploader)
@ParagonOxy
Please read the comments/description. Should be obvious why I didn’t convert to lossless.