(cache)What I Saw and Heard on SeaDex, and My Personal Thoughts

What I Saw and Heard on SeaDex, and My Personal Thoughts

(Just wanted to record my own thoughts. I originally planned to put this in a spoiler tag, but that seems to mess with some features, so I’ll just leave it as is.)

I never expected to post this so early. I have a certain (foreign) someone to thank for giving me the motivation. Because of that, the content is extremely, extremely, extremely incomplete, though the main figures (I wanted to talk about) are all here.

For those who don't understand the following and have no interest in it, please just treat it as me talking to myself and feel free to move along. Since this is a personal reflection posted exclusively on Tianxue, any form of reposting, quoting, or mentioning is strictly forbidden—not that there is much worth reposting anyway. If you see this anywhere else, it does not reflect my intentions.

This post mainly discusses encoders (or encoding groups) and related individuals recommended by SeaDex. It does not include old acquaintances everyone knows well, or those who stopped encoding long ago, such as Beatrice-Raws or VCB-Studio—though the former two don't encode very frequently nowadays anyway. Due to the tone of this post, I won't mention my current personal favorite encoder, but since they haven't encoded many projects, it shouldn't matter much (probably). However, I will recommend a few individuals/groups that I personally think are quite good.

Additionally, I want to make it specifically clear: being recommended by SeaDex does not mean having any sort of relationship with them.

(Ah... just a heads-up, the wording in some sections below might be a bit harsh. I didn't plan to be this blunt, but thanks to a certain someone, if you are a fan of theirs, I apologize in advance.)

Thoughts on SeaDex Itself

First, speaking of SeaDex, we definitely have to talk about the site itself. Personally, I am very grateful for this website (releases.moe). Many project entries include comparison screenshots, which saves a lot of trouble. At the same time, they moved some encodes that originally only existed on specific PTs (Private Trackers) over to Nyaa. It was actually because of this that I managed to discover my favorite encoder.

However, I still have to say that their choice of file naming using the TVDB format is completely terrible. This is especially true for OVAs, movies, and specials, which they name as S00EXX. For example, a certain OVA might be named S00E11. When you see this, you get confused: Since when was there a Season 0? This OVA is episode 11, does that mean there are 10 previous episodes of OVAs I haven't watched? No, there are no "first 10 episodes" of OVAs at all. There is only one episode; the rest are just OPs, EDs, and things like that. To top it off, TVDB lists specials from different seasons all under the S00 prefix sequentially. I mean, how are you supposed to differentiate which season they belong to? Do we have to match timestamps one by one? (And I'm not the only one complaining about this; I've seen encoders who use English subtitles get confused by it too.)

Furthermore, some anime have different title suffixes for different seasons, like Nanoha, Railgun, Maria†Holic, etc. By insisting on the TVDB naming convention, Nanoha A's becomes Nanoha S02EXX. Where did their self-proclaimed spirit of "respecting the original material" go? (I actually have seen people name files this way.) Not to mention, the title suffixes for series like Nanoha and Railgun carry their own specific meanings. Thankfully, not many encoders (or groups) use such terrible naming. Most only add it extra in the release title or filename, and many don't use it at all. The few who do use it will usually clarify that the episode numbers correspond to TVDB. Of course, there are also those who fully adopt SeaDex's suggested naming convention without giving any heads-up—including a highly recommended encoder mentioned below, though he basically only releases the main feature, so maybe it doesn't matter as much.

Additionally, while I agree that past encode versions should be included in comparisons, I deeply detest their behavior of criticizing encoders (or groups) from over a decade ago. To put it bluntly, if they were to encode those difficult titles using strictly the tools of that era (encoder libraries, script versions, etc.), they might not necessarily handle them any better than the people back then. Conversely, if top-tier encodes from the past were processed using today's tools, I don't think the final output would be inferior to any of the individuals listed below.

Well... if you just want to look at comparison screenshots or find out which version SeaDex recommends, you can get all the necessary information simply by visiting their website.

As for their Discord... yeah... I wouldn't really recommend going there... especially if you are quite sensitive about "progressive" matters...

A side note here: Certain people, particularly in the West (not all), always assume that the wild, diverse imagination and creativity in Japanese ACGN align with their "ideologies." But I think for the vast majority, that is not the case. Pure fetishes are being mistaken for "correctness."

That being said, I deeply respect other people's choices and ideologies, and I hope for mutual respect rather than being forced to accept things. Frustratingly, reality is quite the opposite. Most of them not only want to force your thoughts to align with theirs, but they even want to turn around and influence the industry itself, wanting to change everything into their "ideal" state (real non-diversity). I should say, there have already been specific cases where settings were directly influenced and changed for a game character I really like (gnashing my teeth). No wonder the father of Dragon Quest expressed his dissatisfaction, and the original editor of Dragon Ball said bluntly: "It's like an evil god, acting in the name of goodwill, interfering with creative freedom."

Anyway, after spending some time in SeaDex's Discord, I noticed a few clues... Oh, right, on the release pages they mirror from PTs to Nyaa, they sometimes put up posters promoting their ideology. I'm really curious (like Chitanda): Did this get the consent of the original encoders whose work was mirrored? Or has the original encoder already been "represented" without choice? In any case, at least I haven't seen my favorite encoder in this Discord, and I believe he is someone willing to understand and respect others mutually.

As for those who don't want mutual understanding and respect, but only want the world to turn into their "ideal" shape, whether they are in that Discord or not (hopefully not)—

Somebody who lives in paradise but yearns for hell deserves exclusively veritable hell that doesn't involve innocent others.

(I didn't expect talking about SeaDex alone would take up so much space, way too many tangents... anyway, there isn't much content below).

Before starting my personal tier list for encoders (or groups), let me state that unlike others, I don't judge purely based on video quality. Instead, I look at whether the attitude they present matches their level of video quality (quite peculiar, right?). Let’s begin. (Don't complain about the formatting; I purposely kept it simple).

Tier A

motbob (MTBB, Okay-Subs, Okay-Raws...)

The core figure/cult leader of SeaDex. He is like an upgraded encoder (?) but downgraded translator version of ssnake (Snake-da, sorry for mentioning you in a place like this >_<).
Some people might find it strange: "Isn't MTBB one of the top English fansubbers? Even if he doesn't do the primary translation for most projects, he has done translation checking, right? He even criticized the official English translation of the Re:Zero light novels recently for having severe issues. No matter what, he shouldn't be bad, right? Furthermore, how can you determine a translation downgrade using Chinese vs. English?" Setting aside that almost all projects handled by Snake-da are translated and checked by himself, reading the original work before translating is the absolute bare minimum, right? Who cares what the official translation looks like. Or perhaps he is very passionate and expects competent translations for original works that have subtitles... (Based on my personal experience following subtitles from FLsnow and the .subbers project), before translating a work, Snake-da conducts all kinds of textual research, including but not limited to interviewing the original author, looking into the presentation direction of the anime production staff, and even buying all kinds of merchandise to study and analyze character backgrounds and the author's creative mindset. Oh... right, there is one more thing that Sir Motbob might (ha...) not be able to do—attending various events held by the production team for works he particularly loves, to listen to the creators talk about every single step in person. Uh... sorry, got sidetracked again. Saying all this is just to express that both Motbob-da and Snake-da are rare all-rounders in the fansub scene, truly impressive.
On the encoding side, he is very strong in everything from source analysis and filtering scripts to encoding parameters. However, I am quite curious: what is the basis for someone whose encodes were mediocre five or six years ago to say "dehaloing cel is the devil"? (If anyone knows, feel free to share). Well... right, I dug up and checked out his deleted or missing past work from certain places on Nyaa and Tokyo Toshokan. Quite a few were just remuxes of other encoding groups. As for how his own encodes turned out—since they've all been deleted, I think it goes without saying. But who doesn't have a past? That page of history has already turned, so there's no need to obsess over it. Still, fundamental principles must be understood. Therefore, I am even more curious: some project recommendations on older SeaDex sheets used to list a dehaloed rip as the "best," so why have they now completely switched to recommending BDRemuxes or versions that retain the halos? I respect other people's philosophies, but changing something so important back and forth isn't great, is it? Could it be a change made by cult followers just to cater to their leader, Motbob? Or did the maintenance staff of SeaDex change?
All his works encoded with x265 are recommended (roughly starting from the second half of 2021). If you are interested, you can collect them (note that some file sizes are massive). As for his x264 ones... not even the SeaDex people (dared to) recommend them...
Personally, I actually admire him quite a bit. I guess he probably got mocked for his video quality in the past, seized the opportunity during the rise of x265, worked hard to improve his encoding skills, and his actual strength basically matches his status as a grandmaster. It’s hard not to put him here. As for whether his past pages will be flipped through by future newcomers? Who knows.

sam (Should only have this release name, usually uploaded by others?)

Very strong in source processing, filtering scripts, and encoding parameters.
Massive File Size Creator 1.0. A slice-of-life anime episode with no complex visuals can easily be encoded to at least 2GB or more. Perhaps this is what they call "transparent"... It reminds me of a post I saw on Doom9 where an x264 developer, back when x264 10-bit was just starting out, found out that tp7 from ANE encoded a slice-of-life episode to over 1GB. The developer told him, "Don't do that, tp7." At the time, this was mockingly talked about by detractors. I wonder what those mockers would think if they saw Sam's x265 file sizes today. Oh... almost forgot, someone actually defended his file sizes in the comment section of a release page: "Using modern technology, we have entered an age of high quality anime encodes. The level of quality makes the file size possible. And the file size makes the quality possible." I thought the biggest selling point of x265 was smaller file sizes with better quality? How did it turn into "this file size makes this quality possible"? In that case, you might as well just release the processed BDMV directly; wouldn't that be more transparent?
He encodes quite a lot of anime, and a remarkably high percentage of them feature a female lead as the main protagonist. Funnily enough, among my favorite works with a female lead, barely any have been encoded by him. Combined with the fact that he encodes some titles leaning towards the Otome genre, it once made people suspect that despite the name "Sam," this person might actually be a lady. However, looking at the situation with SeaDex (as mentioned in the previous section), it might be what I think, or it might not be. Regardless, personal respect must be given.
Basically, all his works are recommended (even though the file sizes are really, really massive). He only uses x265 for encoding (presumably). In the early days, many releases bearing Sam's name were actually just taking other people's encodes and remuxing English subtitles into them before releasing (the selected videos were always the best versions, so it didn't really matter if you grabbed them). As for when he started encoding his own stuff, the release pages have detailed descriptions. Find them yourself if you want to download them; I'm too lazy to dig through them. Also, a quick reminder: his encodes very frequently get a v2 or even v3, so remember to download the latest version.

derpie

Very strong in source processing, filtering scripts, and encoding parameters.
He has encoded very few projects (so far). The information on his release pages is very detailed. He keeps a very low profile personally and wants people to report any issues on the SeaDex Discord, though he is rarely seen in the DC himself.
He can be considered one of the more liked encoders among those listed here.

Tsukudakobashi (Pizza, Half-Baked)

Very strong in source processing, filtering scripts, and encoding parameters.
Massive File Size Creator 2.0. Honestly, some of the self-encoded parts from their group (Pizza) weren't encoded by him, but they are six of one and half a dozen of the other, so there's no need to differentiate.
One of the most arrogant encoders I have seen from past to present (there are two; the second one is below). Yes, even more arrogant than the ones mentioned above and past predecessors, and loves to trash other people's encodes (especially older codecs). The video quality is indeed good, and he has encoded a fair amount of works, so placing him here is not an issue.
The guy who gave me the motivation to write this post. I completely reject everything about him as a person except for his encoding (not because of his arrogance). I know this is a very severe statement, but his ideology is indeed one of the very few cases that I absolutely cannot agree with. I don't know how many retards (sorry, I have to use this word) like him know absolutely nothing about the true side of what they worship. If he knows exactly what he is worshiping and still feels proud of it, then that is pure evil.
Some projects have abnormally massive file sizes, and some early works were not self-encoded. Keep an eye out if you plan to collect his releases.

(Originally, there was another tier above this, but since this is a spur-of-the-moment post, I didn't want to over-categorize... Also, the big chunk below won't have any weird tones, haha).

Serendipity

Massive File Size Creator 3.0 (Sam's level).
Has encoded very few projects (so far). I originally wanted to share some personal views on their post-processing (pp), but forget it. Since I didn't finely divide the tiers anyway, as long as the video quality is good, it's fine.

Kawatare (The release name for high-bitrate/crf encodes is Tasokare)

Massive File Size Creator 4.0 (Sam's level).
Originally wanted to say a few things about their post-processing, but forgot about it too.
Likely a Chinese encoder. His projects always have Chinese subtitles muxed in and set as default. I have some concerns here—the Chinese subtitle community isn't like the foreign subtitle community; you can't just take a fansub group's subtitles, mux them, and re-release them at will (open-source subtitles are rare too). Did he get permission from the fansub groups? I hope anyone in the know can share.
Although the file sizes are huge, he mainly encodes movies, so they are still highly recommended to collect.
By the way, the names of both him and Serendipity above are quite interesting. The former is Japanese, the latter is English. Look them up if you are interested.

Vodes

His newer works are also leaning toward the massive file size creator category.
Some of his encodes give off a feeling of being incompletely processed (even though they are already the best available version). His past encodes seemed to have some issues (halos), but they seem to have been completely deleted. I didn't bother looking for them specifically. The ones currently existing on Nyaa should all be safe to grab, except for Hunter x Hunter, which uses SOFCJ-Raws as the video base to remux subtitles.

Ignore

Massive File Size Creator 5.0 (Sam's level).
Very similar to derpie. End of story. I know this is a bit perfunctory, but that's basically it.

(That concludes Tier A. Actually, FLE, smol, cappybara, etc., should also be put in here, but I want to wrap things up, so I'll leave them for interested individuals to discover or for others to supplement).

Tier B

YURASUKA

Beatrice-Raws 2.0
I'm not saying he has any relationship with Beatrice-Raws, and his encoding style is definitely different. It’s simply that his current positioning is very much like Beatrice-Raws in the past—a high volume of encodes. Of course, no matter how much a single person encodes, they can't match a group in its prime. Regardless, I am very grateful for his dedication.
Some people might wonder: isn't he often criticized by SeaDex (standards) for having insufficient bitrate or over-compressing? It is precisely because he can still be frequently recommended despite not quite meeting SeaDex's standards that you know how well he handles other aspects. If his bitrate were as high as the ones listed above, his video quality would probably be on par with derpie.
Also, let me clarify: I didn't put YURASUKA here because he did a bunch of delowpass. In my view, delowpass is just a shortcut. If the source material doesn't have other non-lowpassed sources, how do you even delowpass? Therefore, only by comparing works without delowpass can you judge one's ability to fix video artifacts. Purely in terms of post-processing, I would put him in Tier A.
Highly recommended encoder (specifically his works from 2022 onward; his bitrate before that was truly low). If you are used to collecting works from Beatrice-Raws or VCB-Studio, you can add him to your backup list. Perhaps VCB enthusiasts might feel YURASUKA's recent file sizes are slightly large—though VCB's file sizes have also grown recently—but YURASUKA basically only encodes the main feature, without live-action extras, CDs, or other bonus content. The overall size should be similar to VCB. For someone like me who doesn't want to collect giant files for every single anime, he is an excellent choice.

H-Enc

VCB-Studio 2.0
Since there is a Beatrice-Raws 2.0, of course, we must have a VCB-Studio 2.0 as well. Though his style isn't quite the same as VCB, and his file sizes are a notch larger... Furthermore, his style before and after shifted significantly; I don't know if it's different people encoding (I haven't gone to H-Enc's TG channel to ask). He has a large volume of encodes recently.
A Chinese encoder (or group), probably has no relationship with VCB... right? Maybe has a relationship with ANK. Does this name hint that they specifically encode H-anime?
The BD extras are basically complete (I didn't check carefully to see if there are live-action extras). Recommended for people who love collecting complete extras, though it might put some pressure on VCB loyalists...

(Thus concludes Tier B. Yes, there are only two in Tier B. Likewise, there are plenty of others that could be included, but I need to wrap it up, so that's it).

A quick summary: I highly recommend YURASUKA and H-Enc. The advantages of the former have been made clear above, while the latter is a more convenient download option for certain people.

(Another reminder: my personal tier list criteria is based on whether the video quality matches their attitude, it does not mean the pure encoding quality below fits this ranking).

Tier blabla (blabla → Z)

Aergia

The second of the most arrogant encoders I've seen from past to present. Appears to be like 304 from NMM (Senior 304, sorry for mentioning you in a place like this >_<).
This guy is very much like Senior 304—not many works to his name, yet loves to act like a boss lecturing others on their views (or doubts). However, unlike Senior 304 who clearly expresses his insights, this gentleman will straight up insult you to your face and make you experience his foul mouth. He usually starts with "Go learn some video knowledge, you idiot..." Oh... a quick side note here: I have never interacted with the following individuals; I have only witnessed their arrogance from a bystander's perspective.
The only project he fully encoded is Soul Eater (counting only BD or DVD). It is indeed encoded very well, after all, he did scenefilter.
Because he has well-encoded videos, he wasn't ranked down to the next tier by me. And because he became this arrogant after encoding just one project, he is placed right here. Perhaps what he is actually better at is splicing various versions of BDMV to make a BDRemux.

Reza27 (Reza)

I considered listing him among the most arrogant encoders, but a few of his projects weren't encoded that well (unlike Tsukudakobashi and Aergia), so forget it.
Kara no Kyoukai (Garden of Sinners) is probably his proudest work. He wrote a massive essay on the release page talking about how his encode is far better than everyone else's. Setting aside the fact that those weird, thick-yet-faint lines that appear randomly on the BD weren't processed at all (let's assume it's an "artistic choice"), the banding and aliasing were indeed handled more delicately than ANE. However, what was shocking is that despite using x265 10-bit and having a file size at least 100GB larger than other rips, the details in certain places were actually preserved worse than ANE and Beatrice-Raws, and many parts could only be said to be on par with version A and version B...
Hey, man, you used x265 to encode this, and it's 100GB larger; that extra size should give your rip an absolute advantage in detail over other versions... I know he made some excuses, saying things like: "I minimized detail loss as much as possible, but naturally, some still occurs," "There are still some scenes that cannot be completely fixed because of how bad the source is," "There is not now, and never will be, a method to encode a perfect KnK," blabla. Nobody asked you to encode a perfect KnK. You were the one who said your version is "far better than the BD or any other encode." At the very least, it should be noticeably better across the board. Looking at it overall, his KnK might be the best (excluding the condition that the bizarre lines weren't fixed at all), but the margin of improvement over ANE is nowhere near worth that extra 100+ GB, even though I don't consider myself someone who skimps on file size for works I particularly love. Furthermore, I believe those inexplicable lines should have been fixed (not to mention this is x265 10-bit vs. x264 8-bit).
His encode of Oreimo is also only better than ANE in the literal sense (the file size is also larger). I don't know if SeaDex is a blessing or a curse for him—while saying his encode isn't great, they also helped him find an excuse by pointing out the source's interlacing.
tp7, look at you. You were so arrogant back then (though you undoubtedly had the right to be), but look at what kind of characters the successors are today... They don't give people the sense of amazement your encodes did back in the day, yet each one acts more arrogant than you. Though, for someone like you who lost interest in encoding long ago and now works at a certain institution in Russia, you probably have even less interest in these "trash" in your eyes.
Actually, Reza was almost ranked in the next tier by me. Before typing this post, I saw that his encode of Hanbun no Tsuki ga Noboru Sora was pretty decent, so he stays here.
The main reason I have a bad impression of him comes from certain remarks of his. He loves to show off how strong he is (which only comes off as arrogant), or tells some meme jokes that aren't funny at all (like mimicking the encoder of Judas). Though the latter is actually fine; at other times, he seems to rarely insult people randomly. Friends who know Reza's character can share. If he is just a guy trying to act tough, I will move him to the tier above.

Tier Zzzz...

herkz (commie)

A consistent clown from the old Nyaa days until now.
Has been thrown into the dark room (banned) multiple times. It seems the frequency has been lower since the new Nyaa (? By the way, personally, I think the old Nyaa had much more complete features than the current one. If I remember correctly, you could even directly search for grey torrents).
Some people might say a clown is just for entertainment, what is there to say? Well, everyone should understand the principle that shared joy is double joy.
No recommended encodes. For the few perhaps 720p BDs done by his group that people might collect, it's not my place to say anything.
His group is praised by many English users as a "premier" fansub group (although they are frequently called out for translation errors). He has probably translated a lot of works (tbc), but... that's exactly why it's unbelievable. How can someone spout so much nonsense that is so far removed from the facts... (And he especially loves to farm engagement in the comment sections of popular release pages).
Given that his clown deeds are far too abundant to list comprehensively, and I am already tired by the time I typed this far, those interested can look at the comment section of the 4K version of the Violet Evergarden Movie released by Beatrice-Raws on Nyaa. Fairly speaking, what he said wasn't entirely wrong, but the part regarding the production staff completely contradicts the information I have come across. Especially since this guy could actually say words like, "Because of..., so I think... cannot be counted as..." Who does "he thinks" he is? Can that be used as a basis for the anime production process?
For background information, I recommend checking out Snake-da's explanations on the Anime Subtitle Forum (bbs.acgrip.com). He attended the event and listened to the production team's explanations firsthand. The .subbers project release page for Violet Evergarden on DMHY also has some explanations. The above might be scattered (but I highly recommend them personally), or you can just go directly to violet-evergarden.jp/special/dolby/ to check (though information there is more limited).