kitsunebi77

kitsunebi77's downloads & discussion

Recommended Posts

17 minutes ago, kitsunebi77 said:

I got bored and uploaded something. It's another issue of Famitsu, full of this and that as they are wont to be. Please pour out a 40 for the Chinese message board I found this on a year ago which has since gone the way of pogs and pudding pops.

This is a nice one! What else are you holding out on us? 🙂

I am curious about the backstory on this Chinese message board. Was it actually hosted on a Chinese ISP or just predominantly Chinese users? How did they get into gaming? Just kind of curious considering the state censorship and control in their culture.

Share this post


Link to post
Share on other sites
1 minute ago, Ethereal Dragonz said:

This is a nice one! What else are you holding out on us? 🙂

I am curious about the backstory on this Chinese message board. Was it actually hosted on a Chinese ISP or just predominantly Chinese users? How did they get into gaming? Just kind of curious considering the state censorship and control in their culture.

Dunno.  I think it was a .com, so no telling.  Users could have been from Hong Kong, Taiwan...who knows.  All I know is I had to figure out how to register and write everything in Mandarin.  Every download link was hidden until I posted a message to that issue's thread.  Got really good at typing "thank you so much" in Mandarin.  And the shitty download service was tricky as well...often took days to download a file.  They mostly had Chinese mags, which I didn't download.

Share this post


Link to post
Share on other sites
On 10/7/2018 at 5:03 PM, kitsunebi77 said:

large.6634687_Famitsu0279(April221994)a.jpgunliscensed.jpg

Ha, I found a post in this thread from when I was messing around with that site over a year ago.  The mag on the right was one of the ones available to download.  Not affiliated with Famitsu in any way, just plagiarism, pure and simple.  If they can't even come up with an original cover, how much effort are they gonna put into their coverage?  One of the reasons I didn't download the Chinese stuff (although the main reason was because it was such a pain in the ass to do so.)

Share this post


Link to post
Share on other sites
11 minutes ago, kitsunebi77 said:

Ha, I found a post in this thread from when I was messing around with that site over a year ago.  The mag on the right was one of the ones available to download.  Not affiliated with Famitsu in any way, just plagiarism, pure and simple.  If they can't even come up with an original cover, how much effort are they gonna put into their coverage?  One of the reasons I didn't download the Chinese stuff (although the main reason was because it was such a pain in the ass to do so.)

Not gonna lie, that's a great cover. Might as well use the best. 😀

Share this post


Link to post
Share on other sites
On 11/17/2019 at 4:41 AM, kitsunebi77 said:

I got bored and uploaded something. It's another issue of Famitsu, full of this and that as they are wont to be. Please pour out a 40 for the Chinese message board I found this on a year ago which has since gone the way of pogs and pudding pops.

https://archive.org/details/famitsu0288june241994

large.1391015391_Famitsu0288(June241994).jpg

 

I was gonna try an OCR but the PDF is too washed out, likely from JPEG compression.

Share this post


Link to post
Share on other sites

OK I managed to get an OCR of a page from the Gameboy Gallery Strategy Guide. It had a 90% transcription success rate... not too shabby though certainly requiring a bit of work to read, especially with the machine translation. I used https://www.newocr.com/

  • Thanks 1

Share this post


Link to post
Share on other sites
2 hours ago, vgmax said:

OK I managed to get an OCR of a page from the Gameboy Gallery Strategy Guide. It had a 90% transcription success rate... not too shabby though certainly requiring a bit of work to read, especially with the machine translation. I used https://www.newocr.com/

Well, that's a bit of a crock then!! You're likely to get something like "I like Princess Peaches star count" turned into "I like Princess Peaches star c<OCR removes the o>unt" then which will cause no end of confusion for the young Gameboy kiddies 😂 

Share this post


Link to post
Share on other sites
11 hours ago, vgmax said:

 

I was gonna try an OCR but the PDF is too washed out, likely from JPEG compression.

That Famitsu is a CBR, so if you downloaded the PDF you downloaded the wrong file (the PDF is a SUPER-compressed file automatically generated by the Internet Archive).  The Famitsu CBR is actually a very good scan, but its resolution is far from 600dpi (hence why it's not a 4GB file), so OCR is likely going to be difficult.

The reason the Game Boy Gallery page may have partially worked is because it's a small book for children, so the text is printed quite large on the page.  So even though the scan quality on the GB guide is worse than the Famitsu issue, since the text is so much larger, it was easier to OCR.  Since it's printed in a style unlike all magazines, however, it isn't a good choice to demonstrate how successful a magazine OCR would be.

Share this post


Link to post
Share on other sites

Well, right off the bat this highlights one of the problems OCRing magazines vs books: books aren't printed in columns.

I select the text on the page, but it copies the two columns as if they're the same block of text.  So the first line of text, rather than continuing onto the second line in the first column, continues onto the first line of the second column.

As you can imagine, this causes any direct copy and paste into a translator to be complete rubbish.

There are also many inconsistencies caused by the OCRd text simply being incorrect, but that might be solved if only the source file were a higher resolution.

Share this post


Link to post
Share on other sites

LOL.  Here's what google translate's site makes of the OCRd text of the caption on the dinosaur pic (well, and the pic next to it, whatever the hell it is):

This dinosaur is all about porin
Jurassic Bar
Do you dream of rafiq?
The development of
Do n’t worry about software
I don't mind
 

...Yyyyyeeaaaaahhhhhhhh.......

 

(btw, The dinosaur pic SHOULD say something like "This dinosaur is made entirely of polygons.  Are graphics equal to Jurassic Park no longer a dream?")

Share this post


Link to post
Share on other sites

I kinda want to see how modern commercial Japanese OCR solutions deal with this kind of data. A quick search shows the leading products are Panasonic's 読取革命, NTT's eTypist and Sourcenext's 本格読取 (which uses Panasonic's engine). Prices range from about $50 to about $100, and they all promise modern intelligent OCR (i.e. one that uses some linguistic AI in addition to pure optics) as well as intelligent understanding of layout. The first two have demo versions available for download, but they're Windows only so I'll have to wait until I'm home to try them.

In the meantime, I remembered app Google Translate has a manual OCR scanning mode where you can take a photo and help the engine by painting over sections of text. I tried it on the same Famitsu review from the other thread, super zoomed-in on my laptop screen. The OCR results are flawless. Sure, this is a very short text and it's fairly clear, but still, it's impressive.

 

IMG_1190.thumb.png.b875496d4a67d83502804b6ce343b00e.pngIMG_1191.thumb.png.bc83e5212c1ceea646ec89334fce93c7.png

  • Like 1
  • Thanks 1

Share this post


Link to post
Share on other sites

I use the Google Translate photo mode for translating screens from video games I'm playing. It is great.

I have a selection of OCR if we were going to do some sort of proper testing I would be happy to take part. None is Japanese, the software that came with my Fujitsu scanner is English licensed for the Western market but it does support Japanese though not sure how well. 

Also, @kitsunebi77 you can try to find PDF software that lets you select text using a marquee rather than by letter/word, which will allow you to select columns of text. Whether or not you still have to flow the text into one line/paragraph I'm not sure.

Edited by gingerbeardman

Share this post


Link to post
Share on other sites
9 hours ago, gingerbeardman said:

Also, @kitsunebi77 you can try to find PDF software that lets you select text using a marquee rather than by letter/word, which will allow you to select columns of text. Whether or not you still have to flow the text into one line/paragraph I'm not sure.

Actually, I'm just sort of playing the cynical devil's advocate here to keep the conversation going.  I honestly have no vested interest in OCR or machine translation so far as they apply to video game mags, since I don't read video game mags.  I read lots of comics, but even if I read a lot of manga (I don't), almost everything under the sun gets translated by someone anyway (officially or unofficially), so I have no need to jump through a million hoops OCRing and machine translating.  Which is a good thing, since even if both of those steps produced flawless results, the actual process of doing them is too tedious for me to want to deal with, quite frankly.

Share this post


Link to post
Share on other sites

Regarding OCR, I use ABBYY Finereader. It's pretty accurate if you have clean high-res scans, escept for similar things like つ & っ. Google OCR might also work, haven't tried that for scans. If you have something you want me to try it with, lemme know.

Share this post


Link to post
Share on other sites

I also have ABBYY, both a version bundled with my Fujitsu scanner and another FineReader standalone.

And I have Acrobat (very good), Prizmo, OCRKit, fScanX, ExactScan, VueScan and iText.

Happy to take part in an organised test.

Edited by gingerbeardman

Share this post


Link to post
Share on other sites

I think OCR for clear Japanese text is pretty much a standard for all apps nowadays. It's not like we're trying to OCR scratchy handwriting. I just installed ABBYY 14, maybe they fixed the tsu issue by now lol. I think that one's impossible though unless it is programmed to recognize some bits of the language itself, like when the sokuon is used.

Share this post


Link to post
Share on other sites
14 hours ago, vgmax said:

Organized test? Organized how?

We could put together a corpus of pages of varying styles, layouts, sources and run them through all the available software and then date and grade the results?

Just thinking out loud here. 

Share this post


Link to post
Share on other sites

Dungeon Master II guide.  Apparently this game was released first in Japan, despite being a Western-developed game.  Not really sure what the story is there, but I'm also not going to spend any time figuring it out, since I never was one of the Dungeon Master faithful.

https://archive.org/details/dungeonmasteriiofficialguide

large.1761354334_DungeonMasterIISkullkeepOfficialGuide.jpg

  • Like 1

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.