Street Fighter V (Beta) by Tiberious at 11:37 PM EDT on October 23, 2015
So, now that this is on Steam, I'm wondering what format it uses for music.
The lone .pak file, StreetFighterVBeta-WindowsNoEditor.pak, is an archive containing pretty much everything.
There's already been a file list made of its contents, and I'm now seeing if the person that made the list will be willing to share the BGM directory's contents.
Would anyone be interested in looking into this format?
@u@ I kinda wanna know if anyone already did a .hca rip already cause... All the converted rips have the wrong loop points... So I was just gonna go in and loop em myself... but I would need to get the beta first and it would be easier if someone already had the files on hand.
EDIT: Looks like I can't extract the files anyway since UnrealPak is for x64 and I use x86... QuQ *cries*
I'm also doing a LOGG/OGG variation mock-up of this set since it supports multiple channels and less space consumption (although it's probably not perfect when using oggenc from 6-channel .wav). Either way, it will have both stereo and 5.1 folders. .ADX might be better though, since after applying channel mixed effects to listen quality different, the converted .wav sounds more clearer.
EDIT: Nevermind, had to revise cmd line properties. Should sound good now. The only issue is uncertainty of how 5.1 .ogg consistency will be.
Got the loop point data... I can't believe I didn't try this for the Street Fighter x Tekken ones... It would have worked... So yeah... Use hca_decorder or whatever it is called then put the wavs in Wavosaur. The loop points are kept in the converted wav files.
Edit: Actually, no. Give me a minute to look through these. For some reason the main menu one had the incorrect loop points, at least in that channel. Reconverting then checking all the wav files.
Edit 2: Ok so during the conversion, the loop points get messed up. Like those that loop still have loop points and ones that don't still do not have loop points, but the ones with loop points have them in the wrong places.
... is v1.12 the most recent version of hca_decoder?
Thanks. Imma test use this and see if any dif. If it does the same thing Imma try and figure out why.
Edit.
Ok. So these incorrect loops are def coded into these files, but get this... It may be pure coincidence, but out of the three songs I've tried this on so far, if u remove the last 325 samples or so and then divide the incorrect loop point by two, you will get a perfect loop. It's worked on three of the songs so far. So yeah...
I mean, I never played the beta so idk if these loops are correct... But they SOUND correct. So idk... maybe I'm on to something, maybe I'm not... You tell me.
You can use VGMToolbox's .POS file creator to make .pos files with the correct loop points. See the screen shot on this page: here. Put the .pos files in the same folder as the converted .wav files and vgmstream can play them back with the looping intact.
Yeah I can do a pos, but I can also just change the wav file's loop points in wavosaur @u@...
Also, u kind of have to delete the last 325 samples of each song. Thing is with these, the last 325 samples is either a repeat of the first samples of the correct loop or some kind of... Imma just say it's silence, I don't know the terms. This leads to cliping so instead of checking to see which does what, it's just easier to remove the last 325 samples and just use the inloops÷2 to achive a perfect loop.
Updated: I was doing this for listening and music moding purposes, but if anyone would like the wav files in stereo with the correct loops I can upload them once I mix all the channels and stuff... But I won't have the time till later today. I'm about to be afk for a while.
Hi everyone ! I converted to mp3 almost all tracks from the beta, so, I brought you the opportunity to those who not want to bother with the particular format.
Ok so as soon as I get home I'm going to fix up the tracks with their correct loop points as wav files...
So I ran into an issue... and well "me and my needing everything to be perfect" self have just realized that not all of the files need the last 325 samples removed... there's this... silence in some of them and for the first ones I did, 325 samples was the constant amount from the ending needed to be take from them... so after giving a listen to this one post edit... I noticed it still clipped... zoomed in to see it had WAY MORE than 325 samples of silence on the end... so yeah... Imma... have to... redo... all of these... *cries* (not doing em right now tho... I got a commission to finish and a dog to walk. @u@)
But the loop points still look to be the incorrect loop points which are coded in to them divided by 2. (In sample form obvs)
Loop points are after the loop string from the original .hca file. You can then use $V*1024 in consideration of that given value for the converted .wav. Loop start (offset range 0x2C~0x2F) is first string following up, and then loop end (offset range 0x30~0x33) is the string after. V in $V corresponds sample length in big endian 32-bit.
Using this, I did not run into any issues in looping playback (this is assuming start/end samples are supposed to be unchanged). The only issue was Karin's theme, but that's likely due to an audio cue mistake on the cymbal being louder from loop end than from loop start (in other words, an error on the production's end).
Going to continue finishing character themes and have an ogg/logg knock-off set.
so you're saying... THAT I HAVE BEEN CHANGING THESE THING'S LOOP POINTS JUST BECAUSE OF... whatever you said... @u@ yeah u said all that stuff but all my brain goes is "AAAAAHHHHHHHH! I'M TOO DUMB FOR THIS ****" cause I am... I was just... trying to contribute for once QuQ...
But yeah, I have no idea what you just said... like at all... I just want to listen to this music and have it loop correctly... help pls?
(And here I was about to upload "The Sounds of Fighting -Necalli Sample-.zip"
that's what I named the sample I was about to upload... but nvm now...
I have a question now that I think about it... Would it be possible to add HCA support to vgmstream? Like so one could just run these through it and BAM!?
And about the one theme that doesn't sound right looped, but has a perfect loop point none the less... That's a simple fix if the loop ends in a simliar way that it starts. Just take a few samples of the end and move the starting loop forward by the same amount.
Whether the EOF or beginning audio has unnecessary silence I have no idea if HCA_decoder could cause that problem.
Basically, with the original .hca file and your converted .wav file, open up the .hca file in hex editor.
There should be a decimal string saying "loop" you want the next string (has size 4) in Big Endian.
Here's a picture using Hex Workshop if it helps:
http://i.imgur.com/3qA6eyc.png
Once you got the Loop Start and Loop End values multiplied, apply that sample count to your application for your converted .wav. It should loop normally from there.
And I decided to look at the loop points of ur Oggs and... You got the same starting loops as I did with my "divide the incorrect starting loop by 2 method" @u@ which makes me happy... I didn't completely fail for once... tho my ending points... yeah... failed with most of them QuQ
plpl3000: The 5.1 channel .ogg portion is the only area I'm uncertain about, mainly because I'm not certain if the converted .ogg kept the same channel placements from the .wav counterpart. In case you were wondering, this was the format I went by (had this in notepad, ignore loop part):
Loop = $v*1024
SFV .HCA: 0 = Center 1 = Front Left 2 = Sub 3 = Rear Left 4 = Front Right 5 = Rear Right
NATURAL 5.1 .WAV: 1 = Front Left 2 = Front Right 3 = Center 4 = Sub 5 = Rear Left 6 = Rear Right
...where the numbers in the natural section represent their respective channel numbers, and the numbers of the SFV .hca files represent the last number of their respective file names. It's also worth noting that the sub is pretty much blank in every 5.1 .hca file.
I rearranged the .wav samples into 1 file containing 6 channels via SoundForge. I then used oggenc to prepare conversion of .wav to .ogg. I set the quality from batch file to q5 instead of q3 (q3 sounded fine until you use karaoke variation and hear a quality decrease, so that felt unacceptable)... didn't use q10 due to file size and, after using Spek and hearing normal/karaoke variations from the .wav and the converted .ogg, it sounded similar to each other.
Anyways, the .wav that was edited should have 6 channels corresponding to their necessary channels. I then used the batch of the multi-channeled .wav to convert to .ogg. The .ogg now contains 6-channels.
The question I'm not sure about how to test or fix, is whether the 6-channels on the newly converted .ogg actually function as intended. If you were to open up with SoundForge for example, the channels have completely moved around... but this comes with the question whether SoundForge is even able to decipher proper channels of multi-channeled .oggs or not (since, it does not allow saving multi-channeled .ogg's).
EDIT: hcs: "according to the Vorbis spec, the order is:
front left center front right rear left rear right LFE"
Going by this, the channel arrangement from the converted .ogg matches the channel order listed above.
Wow a lot of the real loop points are off... I mean, I understand why as they are all multiples of thr same number for whatever reason, but... dang... So far I've had to move a few of the loop points over a few hunder or so samples because it didn't sound right.
I hope these loops are only off because it's just a beta cause damn. I can excuse the loop of the Game Over theme of SFIV being off because every other loop was on point, but damn Capcom, It doesn't take much to fix your loop points before putting them into the beta, does it?
@hcs: Thanks for the source. If that's the case, then the conversion to 5.1 surround sound went just fine. I probably should edit above post for .wav properties listing (probably add your quote too for ogg specs).
@plpl3000: $V*1024 doesn't necessarily have to be used per se. If you looked into calculation processes for .POS loops in VGMToolbox for example, you might see a tab where you put calculation information. This is where it would go to.
The .ogg process was done manually without the use of VGMToolbox, since I don't recall seeing anything in VGMToolbox that has a .logg file maker (it would be nice, but could take up time).
For the method I did (might be more time consuming than other ways via script), all I did was use Winamp, a calculator, and Hex Workshop (other hex editors or an online hex converter can work too). Take the loop start/end strings from the desired .hca file (check Hex Workshop image above) and copy the 32-bit big endian value and multiply that by 1024.
Once you have that, edit the comments section of the specific .ogg file with the values you just calculated. You might come across a bug where you have to make one comment, press OK, and then edit the info again to add the 2nd comment, since it may not like you to add two comments all at once.
@WrappedInBlack: Felt like only two audio files sounded odd, but the most obvious one for me was Karin's theme, more so that the cymbal (correct me if I'm bad at instrument names) is louder near near the end of file than it is from loop start. Balrog's (or Vega's) theme also loops in a weird way, but it doesn't seem to be that bad.
Custom .ogg/.logg set for the 1.08 beta. Includes 5.1 and Stereo sub-category folders.
Note: The 5.1 surround sound files are converted to another Stereo folder since in_vgmstream does not remap the 6 channels from 5.1 surround sound to play accordingly with Winamp using .logg files. File types .ogg, however, can play accurately in 5.1 surround sound folder IF you set in_vorbis settings to remap 6 channels... which in_vgmstream does not support, since .logg uses in_vgmstream.
Do you have the original HCAs? I don't have access to a PC or anything right now, but when I get my PC working I was hoping to update "The Sounds of Fighting" and I would really appreciate it if u could post them here.