Ys IV: The Dawn of Ys English Dub Patch Released!

[Burnt Lasagna]

@ParanoiaDragon
You can just overwrite the patched one. You can thank NightWolve for giving the patcher this ability! 

@Drakon
You can hear the clipping here,

Here's a recording.  Starts getting insane at about 1:18-ish.  Click to download - 16.36 MB

Not that big a deal, since it only has to happen with this one clip. I guess you can decide for yourself though.

[ParanoiaDragon]

Hmm, when I run Patcher.cmd, I get the prompt, push any key, & it says this has already been patched!

[Bernie]

If you kept the backup files, just swap those back into your image and redo the patch.  That should take care of it. 

[Drakon]

Or you can just turborip it all over again and repeat.  I honestly have no clue if I even used the toc fixer right.  I just opened the isos in it and clicked repair on both of them.

[NightWolve]

OK, my disc definitely says "R2F" but Turborip rips and name the files "R1F". I'll PM you a link to the TOC file.

Thanks! Well, it confirms I don't have to add that TOC or do anything. I thought this might be a 3rd Ys IV unique TOC, but it looks like it just shares the same TOC as one of the others,  R1F.

Magic Engine does not identify the patched disc.

Yeah, if you want, you can run my TocFixer, included with the patch, to repair the image after patching. Choosing either of the two Ys IV IDs, and clicking to repair the image will allow it to be identified by emulators. It's not necessary, but it's a plus. Real hardware won't care, and the emulators will just identify it as a PCE CD and will load it all the same if you don't bother.

Thank you so much Nightwolve, Burnt Lasagna, Sparky, Sadler, and everyone else for your work on all of these different projects!

Err, you're welcome!

Hmm, when I run Patcher.cmd, I get the prompt, push any key, & it says this has already been patched!

You'll have to delete the backup folder, then Patcher.cmd will run again and repatch. I didn't completely modify the original patching code (cubanraul's), just did enough to combat the early patching difficulties. It will likely skip the waves and won't identify them, but since they're replaced, it don't matter (You could run TocFixer to repair the image and that would cause it to completely redo everything though).

Err, yeah, as Bernie said, that works too, restore the files from the backup folder (and delete it afterwards). Come to think of it, there should be a simple "restore.bat" to handle that.

[Drakon]

Lol that clipping is horrible.  Guess I'm reburning this.

[NightWolve]

Here's a recording.  Starts getting insane at about 1:18-ish.  Click to download - 16.36 MB

What causes something like this?  Why doesn't it do this in emulators?  Is it a bug with the real hardware?

 It has to do with the ADPCM chip in the CD unit. ADPCM uses delta's, and it adds them on every sample then outputs the sample so you hear it. The if the sum adds up past what a 12bit value can hold (either positive or negative), this particular ADPCM controller chip doesn't not correct the 'overflow' like other later or more expensive chips do (they'll saturate it; keep it from overflowing). On the TGCD the overflow value wraps back around, causing all samples after that point to be incorrect in the stream (until it starts a new stream). Some emulators take care of overflow even though the original hardware doesn't; but it's not a given that all emulators will do this (mednafen will probably allow overflow). It's not a problem with the TGCD per se, but rather the encoder you use for the ADPCM. It needs to take care for this. If the encoder doesn't have this feature, then you have to lower the volume of the source material (wave file) so that it doesn't produce this in the encode - which is an el cheapo way of doing this.

 I assume this is what happened (or they got the byte offset wrong when inserting it back into the data track).

Ah, thanks for the technical explanation Tom! Yeah, in testing, we learned that you had to lower the volume level close to or lower than what the originals were, and I did observe that one mistake in the stream would cause a complete, garbled disaster thereafter and it'd take a while to correct itself, if ever! That ADPCM chip is "way too damn sensitive" was my feeling, but your explanation sheds light on things better. My ADPCM Finder and batch file builder is pretty solid, so I think all the offset data is pretty good.

Anyhow, here is the SOX commandline I chose to convert to VOX:

Code: [Select]

sox -V -G "sample.wav" -r 16000 -e oki-adpcm "sample.vox"
-G option: "Automatically invoke the gain effect to guard against clipping."

SOX Options regarding clipping:

−−norm[=dB-level]
Automatically invoke the gain effect to guard against clipping and to normalise the audio. E.g.
sox −−norm infile −b 16 outfile rate 44100 dither −s
is shorthand for
sox infile −b 16 outfile gain −h rate 44100 gain −nh dither −s
Optionally, the audio can be normalized to a given level (usually) below 0 dBFS:
sox −−norm=−3 infile outfile
See also −V, −G, and the gain effect.


−G, −−guard
Automatically invoke the gain effect to guard against clipping. E.g.
sox −G infile −b 16 outfile rate 44100 dither −s
is shorthand for
sox infile −b 16 outfile gain −h rate 44100 gain −rh dither −s
See also −V, −−norm, and the gain effect.

Given these SOX command line options and your intimate knowledge on the specifics of NEC's ADPCM, is there any smarter command line that would help prevent this ? It seems you knew about SOX long before I ever did, I found out about it from Charles McDonald's NEC documentation work myself. Right now, Mike just keeps lowering the volume level, but from vestcoat's feedback, it would appear it's being done a tad much, though he says it's fine in real NEC hardware, so maybe it's OK. Just sounds like from what you said, there is a "smarter" way to catch and correct this problem instead of continually lowering the volume level overall. SOX warned you about what it thought would clip, but things have to be a lot lower when it comes to that OKI ADPCM chip, so SOX's warnings stopped being helpful at some point, though they did teach me about the concept of clipping, as did the way the game responded when I tested it on real hardware.

[Burnt Lasagna]

Mike just keeps lowering the volume level, but from vestcoat's feedback, it would appear it's being done a tad much, though he says it's fine in real NEC hardware, so maybe it's OK.

To make clear, I'm not exactly lowering the volume when I fix the clips.

Most of the new ADPCM needed a high-pass or a bass-cut filter over them to get them to play right in mednafen and real hardware. A majority wouldn't even play at all if I didn't mess with the clips cutoff frequency before hand, with said filter.

Yeah, so most of the time I just mess with the high-pass filter settings in audacity till it plays and sounds nicely in the game. I don't usually lower volume, though that is the case some times.

[Drakon]

Hey, whatever works.

To make clear, I'm not exactly lowering the volume when I fix the clips.

Most of the new ADPCM needed a high-pass or a bass-cut filter over them to get them to play right in mednafen and real hardware. A majority wouldn't even play at all if I didn't mess with the clips cutoff frequency before hand, with said filter.

Yeah, so most of the time I just mess with the high-pass filter settings in audacity till it plays and sounds nicely in the game. I don't usually lower volume, though that is the case some times.

[Bernie]

So....  Whats next on the translation/dubbing scene? 

[NightWolve]

To make clear, I'm not exactly lowering the volume when I fix the clips.

Most of the new ADPCM needed a high-pass or a bass-cut filter over them to get them to play right in mednafen and real hardware. A majority wouldn't even play at all if I didn't mess with the clips cutoff frequency before hand, with said filter.

Yeah, so most of the time I just mess with the high-pass filter settings in audacity till it plays and sounds nicely in the game. I don't usually lower volume, though that is the case some times.

I thought we had shared info that they needed to be lowered down to that +/- .4 range which is where most of the Japanese clips were at on average ? I dunno much about these filters, but are you saying if you had just say lowered the volume level to the .3 range, they wouldn't play unless you messed with other fancy features of Audacity (cutoff, high-bass/bass-cut filters, etc.) ??

[Burnt Lasagna]

The clip does need to be around the +/- .4 range, but it might also need the filter mentioned, depending on how it was recorded.

I dunno much about these filters, but are you saying if you had just say lowered the volume level to the .3 range, they wouldn't play unless you messed with other fancy features of Audacity (cutoff, high-bass/bass-cut filters, etc.) ??

I found all this out during my initial test play through. I believe it was the scene where Slano shows up that I ran into this issue. The clip would play, but no matter how low I set the volume it would always clip.

I'm not exactly an expert on this, but basically the recordings for Slano had to much bass for the PCE to handle. Doing a High-pass, or as it's sometimes called, a Bass-cut lets you bring the clip down to something the PCE can actually run.  This makes the clip sound a tad grainer, but it was needed in order for the clip to even work correctly.

I was never able to get an exact setting that worked with every clip, but it was what audacity had as default that worked for most. 

You can read about a High-pass filter here,
http://en.wikipedia.org/wiki/High-pass_filter

So....  Whats next on the translation/dubbing scene? 

I don't have any plans for another dubbing project as of now.
Though I would be up to doing one again! Just need the right game...

[Sadler]

I don't have any plans for another dubbing project as of now.
Though I would be up to doing one again! Just need the right game...

Snatcher! Please Snatcher! The LoX's would be good too, but how hard would it be to leverage the SegaCD Snatcher?

[Duo_R]

You know what would be cool, Policenauts voice acting. Same idea as Ys IV, the translation is already there but all the audio is in Japanese. That would be an amazing and historic project. Of course whole different system but that would be awesome. I would for sure do some of those roles.

[ParanoiaDragon]

The question with Snatcher, would be, go the easy route & use the already existing dubbing(assuming it's been, or can be, completely ripped), or start from scratch with new actor's.  I'd actually prefer an RPG of some sort, but it'd still be nice to play Snatcher, as I've never played the Sega Cd version. 

There's also the Valis games, I believe someone else mentioned this.  Though, briefly the possibility of Sunsoft releasing them dubbed or remade or something, but, that looks to be fruitless since they vanished off the face of the earth with most of Telenet's games.   But yeah, Valis 1 & 4 are options, & I'm sure we could do better then what was done in Valis 3!

Anything Falcom would be great, Sorcerian probably has the most text to translate.  Popful Mail, I wonder how doable it would be using WD's dub.  There's probably scenes that aren't in the Sega CD version that are in the Turbo version, IIRC.  As mentioned before, the Legend of Xanadu's, but 2 is already in the works.  Brandish would be cool, not sure how much text there would be to worry about though.

Obviously any of the Cosmic Fantasys, though, CF's 1 & 3 I believe are being worked on by somebody.  I know atleast 1 was already being translated, maybe by Dave Shadoff?

Double Dragon 2, probably very little text to translate, maybe none.  River City Ransom, could maybe use most of the text from the NES version(?), though I think alot of names n' stuff were changed.

Kaze Kiri or Legend of Iga Ninja would be cool.  I don't recall how much text is in those.  Probably not much.

Mystic Formula, I believe there's zero text to worry about, though, the game isn't the most sought out, as it's only average.

Shubibinman 3, pretty cool game, almost no text, & was originally planned both by TTI(as Shockman 2) & WD(as Cyber Twins) seperately BITD.