Ys IV: The Dawn of Ys English Dub Project.

[NightWolve]

Ah, good to hear!

[Bernie]

Man, I am so stoked to see this being worked on like this.     Also, thanks for the videos NightWolve. 

[ParanoiaDragon]

The new actress for Karna is Mizura (internet name).
I would say she's definitely a great voice for Karna. I would go as far as saying that she's just as good, if not better, then the actress from JJ's 2004 attempt.

Awesome.  I do seem to recall really liking Karna's voice BITD.

[NightWolve]

Awesome.  I do seem to recall really liking Karna's voice BITD.

That's her , did you miss it? I wasn't sure I should upload it to youtube since she's getting overwritten and it might be misleading (though I made it clear about that), but it's a little bit of history so what the hell.

@BL, eh, I do have something to report... I'm making preparations for performing the S-Video mod on my TurboDuo and I had burned a CD-R of Ys IV with a fully patched image for later testing. So earlier today, I loaded the game up to check out the Dogi dub for the opening scene and the distortion bug was in full effect unfortunately... I suspect it's a preamp level issue, the best idea is to use Audacity to lower the volume to get it to match the original or something. Here's what happens: Dogi says, "Can you believe it, Adol?" and that part sounds just fine, but then the louder "HEH!" part right after causes a distortion and it seems to carry over for everything else that's said thereafter - it messes up ADPCM decoding/processing (you get loud static, obnoxious sounds thereafter for most of the rest of the dialogue)... So that's as best as I can describe it while it works fine in the emulator of course since I just fine... I wonder how your Dracula X dub behaves on real hardware as well. Whatever the case, we're doing something wrong. 

[ParanoiaDragon]

Yup, I remember seeing those scenes of Karna BITD.

So Dogi be clippin' eh?  That stinks, maybe we need some volunteers that can beta test on real hardware once everything is inserted!

[NightWolve]

Ah, so that's the proper term then, clipping ? I read about the issue in the SOX manual. I guess it's a matter of watching the decibel level, finding out what's the max limit and making sure no part of the audio exceeds it.

EDIT: Well, I get it and I guess it's perfectly visible - BL likely has it covered. The answer is likely to keep things between 0.5 and -0.5 or 1.0 and -1.0 at max... All Japanese samples I looked at are safely between .5 and -.5. The top wave is English dubbed Dogi, the bottom is his Jap counterpart. Big difference:



Easy to fix with Audacity then, needs a -5 decibel reduction which does the trick, so nothing to see here. I'll burn another CD-R, but I'm sure it's that simple of a fix.

[Burnt Lasagna]

Yeah JJ just sent in all of Chris's new lines for Dogi the other day, so I decided to do a couple of test inserts to see if it would do the distortion thing even when using SOX. The answer is (of course) yes.

This wont be that big an issue since (as NW described) you can easily circumvent this problem by just lower the volume to every clip to match the Japanese. I've been down this road in the past with Rondo of Blood. Also mednafen can pick up these distortions just like real hardware, so I will be using that to do a test play through after I'm done syncing in all the new dialogue. 

[spenoza]

But once the audio is already in digital form, the clipping has already taken effect. If the volume goes above the limit the waveform gets clipped. Reducing the volume on the resulting clipped waveform just results in a clipped waveform at a lower decibel level.

[Burnt Lasagna]

I'm not entirely sure what your getting at.
The clipping takes effect when the new English audio is inserted with a higher volume then the TG-CD can handle. Which makes sections of the clip produce load buzzing/static noise when played on the real hardware/mednafen.

Simply lowering the volume of the new English ADPCM to mimic the volume of the original before inserting fixes this. I've even tested this numerous times and it works.
(wiki page on audio clipping)
http://en.wikipedia.org/wiki/Clipping_%28audio%29

[spenoza]

As long as the original recording isn't clipped, then yes, volume manipulation can keep it from being further clipped. I was just noting that if the waveform is clipped at all in its original state, prior to insertion, that you can't fix that. Clipping is something you have to plan for right from the initial recording phase.

[NightWolve]

Right, there is permanent loss; it's an indicator that the actor is speaking too loudly or more likely that the microphone recording volume level needs to be turned down some what and the boost feature most definitely needs to be turned off (if on). What's lost in clipping at the loud points in the audio is negligible though, no ? I would guess. But I agree, it would be preferable to have a standard for the actors when it comes to the microphone, figuring out the best recording volume level, how close to be when speaking into it, etc. Instructions to each actor like: record a sample, open in Audacity, and see that it's within the -.5 to .5 range and if it's way past that, turn down the mic's recording volume level (either in the Windows Sound mixer panel or your soundcard's custom panel), try recording again, and/or a combination of moving the mic further away from your mouth as well, etc.

[ParanoiaDragon]

I'm we all have different mic's(& different room conditions in general), so I don't know that definitive numbers will work.  I'm assuming(& hoping) that everyone on the project has a dampner(can't for the life of me think of what it's called), that keeps certain letters, like the letter "p" from clipping, not to mentlon keeping the sound of one's breathing to an absolute minimum.

Does Audacity not show you clipping as you record?  I've been using my old Sound Forge, so I haven't fiddled with Audacity.

[spenoza]

What's lost in clipping at the loud points in the audio is negligible though, no ? I would guess.

It entirely depends. If the character makes a loud yell or cry, then the clipping could be pretty impactful on that one vocal expression. If there is a long monologue and only one loud peak is slightly clipped and nothing else, I don't think anyone would notice without looking closely at the waveform. Clipping does have a greater impact on music than on speech, in part because most music is mixed to reduce dynamics, and rather than noticing explicit clipped peaks, widespread clipping usually simply manifests itself as audio quality decline or listener fatigue. Now, if one of your voice actors has some long lines to deliver and has the mic volume up way too loud, you could get some problematic sections, but a clipped peak here or there isn't likely to provide any great problem. Hell, humans talk on phones quite a bit, and they compress the hell out of speech. Speech survives the transition to low-quality audio much better than music.

[NightWolve]

Some more fun, something I've been toying with as a possible intro to a new Ys IV patch (in the distant future)! Check it out!

http://www.ysutopia.net/downloads/ys4/DarmDemo.7z  (Free 7UnZipper if you ya need it right here)

That is an executable that simply launches a [splash] video recorded in Flash. It's very lightweight too, does the bare minimum as a win32 executable to create a window and an ActiveX instance of the Shockwave Flash control to load the video. I must say, I'm rather impressed with its compatibility after playing with this: I tested it on Windows 98SE and 2000 Virtual PC instances and it worked just fine, in addition to my PC's native Vista platform - as long as the user visited youtube and allowed for the installation of Flash, chances are most PCs will have it. Flash really is impressive in another more important sense: The recorded video (with whatever generic codec) comes out to about ~10 MB, but after encoding to Flash, it's only 746 KB!!! Adding an extra 10 MB to the patcher for this would be out of the question, but since I can get it down to less than a MB, it's somewhat negligible.

Anyway, the idea for this intro is to simply use Darm's line of "You don't stand a chance..." and lead in very quickly with an animated cut-scene that I recorded from the actual opening, him smirking and the 3 fireballs shooting off, etc. I thought it'd be pretty awesome if I could accomplish something like that. But if such a video is gonna start every time the patcher is run, it's gotta be short to avoid becoming annoying, and it turns out since this is all I wanted (as seen in the demo), the length seems just about right coming at under ~5 seconds. It's of course WIP, I'm waiting on clean audio from Justus to see about joining Darm's line with the rest of the audio. Right now, the exe plays a wave file first, before launching the Flash video, so it's far from ready; it's just to give an idea of what I wanna accomplish to Justus, etc.

http://www.ysutopia.net/downloads/ys4/darm.swf (This is the Flash video embedded in exe that is extracted and played, but it's missing the audio I wanna combine it with - clicking this in your browser should be viewable)

Well, I think it'll be a fun/cool idea if I can get the video just right and it looks like the technical issues of launching it before the patcher's welcome screen or within the patcher are mostly out of the way - it's sound engineering that's the problem (recycling Alan Oppenheimer's line and adding the background music to make it all flow with the audio of the video - really need Justus' help there).

[Duo_R]

Just curious if all the recordings were turned in for all the characters.