The current state of sequenced game music ripping scene, in my honest opinion by Katsur at 1:35 AM EST on January 14, 2024
It's only now the best I can understand the depth of complexity and scale of this... crisis, this mess that we have unwillingly helped to create...
Once there was Kode54, who, in my opinion, understood sequenced video game music from consoles, like Nintendo DS (.2sf/mini2sf) and Game Boy Advance (, Sega Saturn and Dreamcast, Sony PlayStation 1 and 2, very well. He created a 32-bit-only Foobar2000 components. Just like the Neill Corlett documented the PSF audio format and made a plugin for now-dead WinAmp that plays this format back in 2003. Now both Kode54 and Mr. Corlett left the game music ripping scene for good, and the future of sequenced gamerip music looks increasingly uncertain, with some hardships in ripping sequenced game music, such as PSF/PSF2 ripping process, still persist. If this wasn't bad enough, Kode54's game music decoders for Foobar2000 are still 32-bit-only, but unfortunately nobody updated his, for instance, PSF decoder to support 64-bit. I have 64-bit Foobar2000 and Windows 10 operating system. I still hope that someone will reverse-engineer such components and update them to modern norms and specifications. Kode54 is currently still working on various random software, which is not as popular as his past jobs. I'm afraid that he has lost momentum, much to my sadness :(
I personally tried to rip sequenced music from NTSC-U Gran Turismo (1998, PlayStation), but I was deeply overwhelmed by difficulties. Software used: VGMToolbox. I tried to find a comprehensible instruction manual to assist me in doing so, but many instructions felt outdated and... vague. I gave up out of bad luck and frustration. I'm not even a programmer, mind you. The PSF making process involved ripping SEQ and VAB files and turning them into playable PSF/MINIPSF/PSFLIB files. Now I wonder if SEQ+VAB music player would exist, optionally with OpenMPT-style pattern display.
They said that one of the PS1 emulators had debug functions to show sequenced music's inner workings, and also disable one of its sound channels. That's the best I can say.
Back then, there were so many successful sequenced gamerip music, from Sega Genesis (.vgm), Nintendo Entertainment System (.nsf), Arcade (.vgm), to Nintendo 64 (.miniusf) and Sony PlayStation 2 (.psf2/.minipsf2). However, the extensions I was talking about were emulated; the music sequence files couldn't work without discrete specific audio chips depending on each console. Metroid (1986, Nintendo Famicom Disk System), Ridge Racer (1993, Namco System 22 arcade), Sonic the Hedgehog (1991, Sega Genesis), Super Mario World (1990, Super Nintendo Entertainment System), just to name a few. Reading Copetti's analysis on SNES architecture and its Audio chapter, I wondered how did the writer managed to get the multi-audio channel display work. Maybe I can enjoy seeing how sequenced DSP channels behave for a specific game OST in the future... Seriously, look at it: https://www.copetti.org/writings/consoles/super-nintendo/#audio.
As a coping mechanism towards the hardships, I fantasize about the highly advanced sequenced game music player with OpenMPT-style sequence pattern display and lots of formats to boot, such as PlayStation SEQ, Nintendo DS SSEQ, and also PSF, PSF2, USF, GSF, 2SF, DSF, SSF, NSF, VGM, SPC, etc.
In conclusion, the sequenced game music ripping scene is really wonderful, interesting, beautiful, yet convolouted. The current state is, in my honest opinion, a mess. Ripping process can be painful in any video game files. Things with streamed game music ripping scene are doing well, but it's a different story. Maybe I was wrong or downright pessimistic, you can clarify. I wanted to express my actual feelings towards the overwhelming hardships of this interesting scene. I'm sorry if I was ignorant or upset towards hardships.
What do you think about the current state? Share your thoughts and feelings.
It's a very complex topic indeed. The slow progress does not surprise me at all. I don't know where to start to attempt to shed some light on the matter. Not only are there many annoying technical reasons while some stuff takes ages to rip but there's also the question of motivation and interest. What if someone is no longer interested in the topic? What if the time it takes to do all of this is not worth it in the end? I can only talk about N64 (USF) and to some minor extent Playstation music but it shows how hard it can be to rip/play the actual sequences properly:
First of all (and it was explained numerous times over the past 20+ years) is the difference between an actual sequence (MIDI, AKAO, SEQ, MOD…) and an emulated file like USF, PSF and the like. The emulated ones are NOT sequence files! This comes up time and time again. People ask if certain "channels" of a USF can be muted or if the "samples" can be extracted or if there's a "sequence" specification for USF and so on. Nothing of this makes sense because USF is simply not an audio file in the technical sense. It may seem like it to the ordinary user who "plays" them in Foobar or Winamp but they're fundamentally different. The USF player does not process a sequence, does not load samples, does not load sound banks. Instead it runs a stripped-down emulator which runs game-specific code which does all of the above and even more. USF/PSF is a black box where all of the stuff that happens inside is "out of reach" to the player so to speak. You run a portion of game-code and route whatever it produces to the output of the player. That's all.
There are 2 big main problems with those formats:
1) the sequence and sample format is NOT the same across all games 2) games can modify sequences and other aspects while the game is running
Number 1 is by far the biggest problem because it means a real sequence player has to implement an entire playback engine for EVERY possible format that a USF can contain. This is a massive undertaking that almost no one is going to do. And even if you somehow theoretically manage to implement all of this, it still leaves us to problem 2: games can modify the music and even the underlying formats in any way they see fit. The official N64 SDK even encourages the developer to modify formats to their needs as long as certain basic criteria are met. This basically makes a 100% accurate playback engine impossible to implement!
Extracting the sequence is one thing. But it does not account for changes that are made during gameplay. This is the reason why such a thing as USF was invented. Because USF contains everything needed to "play" the music correctly, including all modifications and stuff that happens while the sequence is playing. This makes USF way more accurate than any theoretical sequence player can be. But this also means USF is way harder to create. Extracting samples and sequences from N64 ROMs is a matter of 2 minutes, but creating a USF takes days/weeks/months. So in short:
raw formats: > fast to extract > smaller then USF > full control over all aspects > lack of features, game-specific modifications not taken into account
USF: > (very) hard to create > no control over anything > bigger then raw formats > 100% accurate but ONLY if ripped correctly
Yes, I've been around for an extremely long time myself. I'm still semi-active in the SPC ripping scene, including making SPCs of normally undumpable games. I've got five sound drivers' worth done, and I'm working on a sixth (that being Advanced Realtime Dynamic Interplay).
However, I do not consider Halley's Comet Software as my main place. It's a nice collection, but I find myself missing SNESMusic.org, mostly because of the searchability aspect (I was able to search under more categories with SNESMusic.org). And I'm still interested in perhaps attempting to give that site some life even after all of these years (and even after the forums went down in mid-August 2019, just as I had made a few submissions on the normally undumpable SPC sets), since I have quite the update collection.
But I've also been having my own fun by creating my own sound driver for the SPC700, so I've also been doing way more than just ripping: I've been making my own music for it, too!
Someone has already managed to rip 1997 Gran Turismo sequenced music very well... It is divided in two regions: Europe-USA and Japan. Link: https://hcs64.com/mboard/forum.php?showthread=62736&showpage=0#post_62748
VGMTrans has been update to V1.2 in August 2024, and that's good. But...
VGMToolBox is no longer updated as of April 2024, according to this GitHub page: https://github.com/Manicsteiner/VGMToolbox
PSX Multiconverter isn't updated since 2014 and remained trialware...
PSound is abandoned too!
These end-of-life cases raise even more questions about the future of sequenced game music ripping... the state with USF is still a mess nowadecade (10+-year version of "nowadays").
And, heard of libvgm (https://github.com/ValleyBell/libvgm)? This rewrite sounds promising imo