So. What's the point of redoing stuff that's already done just to censor it? Lol
Really late here, but there's plenty of good reasons for redoing the scan, even if we wind up with the original censorship.
The biggest one is that Hammer Shock's version has pages that are missing from Japanese Unloved Manga's version.
Page 12 on Hammer's version is missing between pages 13 and 14 on JUM's version.
And Page 18 on Hammer's version is missing between pages 18 and 19 on JUM's version.
And to add insult to injury, Page 18 from Hammer's version is a double-page spread. Which, to me, makes this a cardinal sin for it being excluded in JUM's offering.
Also, JUM's translation is janky as all hell.
They seem to disregard the fact that Japanese tends to be like Yoda-speak - their preferred sentence structure is backwards from English. This is why some of the speech bubbles in the Underrated version seem like they'd go better if switched around.
If you don't know what I mean, read the following and compare it to some of the mixed up speech in Underrated's translation.
<voice="Yoda">
Troublesome the differences are! At many points oppose each other does Japanese and English. Left to right, then down does English read. However, first reading down, then right to left is the traditional Japanese way. Rely on much punctuation does English. But no punctuation does traditional Japanese have. Instead, sentence structure marked by special syllables as word suffixes. Where one sentence ends and another begins, hard to know it is. What to look for, one must know already.</voice>
Also, JUM's translation looks like it got fed into a large language model AI, and they copy-pasta'd the results without checking for important things like continuity of context.
What do I mean by "Continuity of Context"? Stay with me and follow along carefully - I promise this will make sense.
All machine translation engines, even the AI-powered ones, treat each individual line of text - that is, lines of text that have line-breaks between them - as completely separate statements with their own context.
Absolutely no context carries over from the previous line to the next one, or the one after that.
This is absolutely critical to avoid because Japanese hinges on context more than most other languages out there. There are so many homonyms and synonyms that if you walk in on a conversation, you'll actually have to ask what they're talking about. Otherwise, a perfectly normal and polite conversation could be taken as something... else.
And that's what's happening in JUM's translation. It looks like they inputted each speech bubble as it's own line of text, and the result is like as if each speech bubble is a different conversation you just walked-in on.
This is what causes panels to have a weird series of sentences that make it seem like someone is talking about two different things at the same time, and can even result in pronoun confusion (where suddenly a female character is being referred to as a "he", and vice versa).
But if you chain the sentences from individual speech bubbles together on a single line of text, and
only add line breaks where another character says/thinks something, then the translation is
a lot more coherent. These results tend to be a lot more accurate, and usually require very little in the way of editing to not only make sense, but also sound natural.
All of that said, I'm going to be reporting JUM's version of this chapter for the missing pages - that's actually a
big no-no in Mangadex's rules; all uploads must preserve the original composition of the raws. And I'll also be going back to check if there are any more missing pages in the rest of the chapters they translated.