Previous Page | Next Page
- by MoldyPond at 2:36 PM EST on December 2, 2018
- 32khz ADPCM seems to be good enough. I mean, it's not winning any awards or anything but at least has the decency to not sound like compressed MP3s.
- by bnnm at 3:15 PM EST on December 2, 2018
- It's that kind of mentality that got us those opus, just saying. "Everybody does 48khz ADPCM but why bother, we can go 32khz, it's good enough for the kidz!". Tomorrow it'll be good-enoughs 32kbps codecs, etcetera.
- by MoldyPond at 3:46 PM EST on December 2, 2018
- Well seeing as how keeping all the music at 48khz like Smash 4 would've only added an extra GB or two at most to an already smallish file size, I still believe there was absolutely no excuse for lowering the quality so far. Why not 128kbps? I'm sure that would've only added like an extra 500MB.
Unfortunately audiophile grade quality in video games is practically non-existant, and coming from a company that doesn't even believe in high quality graphics it should come as no surprise either, ESPECIALLY if Namco's involved (as in cutting corners to save file space isn't anything new to them).
My one hope is that if Ultimate is anything like Sm4sh was for modding, we can just mod in higher quality music and that there will be a fan made HD music pack eventually uploaded, which I already volunteer to start making once everything's figured out.
- by shockdude at 11:13 PM EST on December 2, 2018
- I tried to learn how vgmstream processes Switch OPUS by reading/editing the source code and by comparing the original W06 lopus with the converted W06 opus.
I can't say I fully understand how it all works, but I have convinced myself that vgmstream is processing the files ok. It's just copying the original raw OPUS data and building standard OggS packets & headers around that data.
Now I really want to hear a line-out of W06 (or N07/N12). Even if the output is filtered, I still think the audio artifacts would be audible (or inaudible!)
edited 11:15 PM EST December 2, 2018
- by Ultrafighter at 1:41 AM EST on December 5, 2018
- Hello Bnnm, thanks for Aerial assault TXTH!
Also may I kindly ask you to check out a bunch of samples? This time it's 2 thematic groups & 4 VG titles in total.
Category #1 is old UbiSoft banks, the games I'd like to see supported by plugin are original Splinter Cell + its followup Pandora Tomorrow (both for PC).
Regarding SC1 there're a few big Music_*.SS0 files & the same amount of smaller Music_*.LS0 files, should this be enough to add support for these to vgmstream? BTW there're also 2 more SS0 bigfiles which have multiple companion LS0 files, the latter seem to be language-specific.
As for SC2 its audio archives can be much more difficult to implement as file structure is more similar to later Ubi VGs - there're numerous *.SS0 files but no tiny accompanying index / list files, there's just 1 small *.SP0 file plus enormous MAPS.SM0 bigfile (I'm not even sure it has anything to do with SS0 banks).
So is it possible to somehow play those without help of decubisnd? It'd be great! And the reason why I haven't upped samples yet is because I have to hear if you're going to need any additional files to look into or not.
The 2nd group are the latest Command & Conquer installments (C&C4 Tiberian twilight and C&C3 Tiberium wars, both for Windows OS) and they prove to be just as bothersome as Red alert 3 (which I didn't rip in the end although I tried 2 platforms).
Most (all?) sounds seem to be headerless or even compressed (their "cdata" extension might suggest precisely that) and very few files can be converted by ealayer3. I tried expecting some convertible streams via hex editor but I can seem to discover any special traits that make them different from inconvertible tracks, I'm not an expert though...
BTW sometimes output MP3s are a few times smaller than their respective CDATA files, those should be 6ch streams I presume.
I hope you'll be able to help me out with this, I believe in you man! And 1 more thing to keep in mind: I added (conv) prefix to the tracks which can be converted ATM, this might be useful I suppose.
edit: Do you think it's possible to make !tags.m3u for bigfiles with multiple tracks in them, for example that MUSIC.PKG/txth from Tribes - Aerial assault? I experimented a bit and this is what I got: A or B. It can be me doing something wrong but neither FB2K nor xmplay doesn't want to add those M3Us to playlists.
edited 3:06 AM EST December 5, 2018
- by bnnm at 3:02 PM EST on December 5, 2018
- @Ultrafighter - SM0 see here.
C&C are normal EA stuff. Rename to .snr if file is small-ish (with a hex editor value at 0x02 is sample rate) otherwise rename to .sns and use snr-maker bms (those 'conv' files), may use EA-XAS or EALayer3. A few files don't look extracted correctly or may not be audio (ex. GlobalStream~166b084d.3107d99b.1a91be1f.296f1810)
!tags subsongs probably in the future, for now use TXTP to select subsongs (remember you can create an empty 0-byte file named: MUSIC.txth#2.txtp instead of a full TXTP).
edited 6:39 AM EST December 6, 2018
- by bnnm at 12:38 PM EST on December 7, 2018
- KBPS PSA
People seem confused by those mysterious bitrate numbers so let me explain. When possible (barring bugs) vgmstream calcs the average bitrate of the *stream*. This number is correct, but can be unexpectedly different compared to the bitrate of the *codec*.
By "codec bitrate" I mean raw data stripped from any of the container parts. Say you have a Nintendo Opus file. The Opus data may be 64kpbs (ex. size 0x32000), but Nintendo adds lots of extra data, thus the stream becomes ~5% bigger 67kbps (size 0x34800, but same quality).
Again, this isn't wrong per se, as what is part of the codec and part of the stream is debatable. For example, Wwise Vorbis removes many unnecessary codec parts, so a Ogg Vorbis of 100kbps (codec) may become a Wwise Vorbis of 95kbps (codec? stream?), with the same sound quality. Or, FSB MP3 have padded frames so 128kbps could become 130kbps.
Technically vgmstream could be modified to show the base codec bitrate but I think time is better used elsewhere.
ALSO: note that numbers going up or down don't necessarily imply upper or lower quality. Wasteful codecs could have higher bitrate, but actually worse quality than better codecs at the same or lower bitrate.
Adventurous users may notice SSBU's files have some unused garbage data at the end. This garbage seems to be part of the song at 96kbps, and the file plus garbage is ~96kbps.
A probable explanation is that all songs were originally 96kbps, but were changed to 64kbps later (for performance reasons?), and they didn't bother to redo the whole thing and just overwrote part of the file (now you can get depressed).
edited 1:39 PM EST December 7, 2018
- by Ultrafighter at 4:44 AM EST on December 8, 2018
- Hello Bnnm, thanks for advice as I've just begun to see the light at the end of the tunnel in regards to those C&C VGs! They all seem to be even more pesky & troublesome to rip than UE3 games though! Not only all the files are virtually nameless there're lots of cases when renaming to SNR doesn't work & I have to make SNRs by method of trial & error...
As for unpacking initial BIG4 archives - I hope it was done correctly, I tried 2 tools I know of and they output the same files so no malfunction occurred there I guess. Maybe BGM isn't just in "separate tracks" style among those *.cdata assets...
And getting back to SC1 & Pandora Tomorrow: let's say I've got only LS0 & SS0 files from the former and just SS0s from the latter (SP0 seems unimportant, there're no SB0s & SM0 is an entirely different story). Is it impossible to add them to plugin or you'll have to look into them to say for sure?
edited 4:47 AM EST December 8, 2018
- by bnnm at 5:25 AM EST on December 9, 2018
- latest vgmstream changes (autobuilds (MSVC) / foobar / test+winamp+xmplay (GCC) / bug list)
- Add Vicarious Visions .zss/zsm/ens/enm [X-Men Legends II (multi)]
- Fix .fag num_samples/subsongs [Jackie Chan Stuntmaster (PS1)]
- Add .opusx/adpcmx [Touhou Genso Wanderer -Reloaded- (Switch)]
- Add looping Ogg Opus [Monster Boy and the Cursed Kingdom (Switch)]
- Fix audacious builds
@Ultrafighter - It is impossible to add SS0 without SM0, which has thousands of subsongs. I won't add it.
- by Ultrafighter at 5:50 AM EST on December 10, 2018
- I'm sorry then, I suppose I didn't get it right from here - I took it as info about specs of audio streams contained in SS0 or SM0 bigfiles can be found in (usually) tiny LS0 or SB0 files. I presumed that SM0 giants can/may store every single piece of audio used by a game while SS0s contain some big streams (that can probably be found in respective SM0s too).
When it comes to Splinter Cell 1 though it doesn't even look like this anymore, for example there's Sounds\ENGLISH\ folder and it holds a single big MAPS.LM0 + lots of rather big *.LS0, it looks like the latter contain actual sounds too. I guess I misunderstood the general structure when I was inspecting select assets with "Music" in their names.
So if it is possible to add SS0(s) with help of SM0(s), does it mean that if I post entire audio directory (not random files which might store BGM but everything) you might implement most/all SS0 files? Or is it always a game specific question and you can't tell before inspecting those files?
All in all it has to be a lot of work, I think I'll postpone ripping SC1 and hope that something can be done about SC2 archives. So may I ask you to look into Pandora tomorrow example files if I upload those? I can make a list of folder contents with file sizes if it helps (that way I won't have to up the whole directory).
edited 5:52 AM EST December 10, 2018
Previous Page | Next Page
HCS Forum Index
Go to Page 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222
Search this thread
Show all threads
Reply to this thread:
Halley's Comet Software
Generated in 0.0408s;