HCS Forum - What's going on in other xsf formats

by Nisto at 4:00 PM EDT on September 4, 2017: Hi mogikihei. I uploaded the PSF set this morning.

Metal Gear Solid (1998-09-03)(KCE Japan)(Konami).7z

Please let me know if there is a problem with any of my code. I have only changed the code of the main function (800148B4). I have an IDA database with lots of notes and named functions also, if he wants it.

Thank you for helping.
by Kirishima at 4:08 PM EDT on September 4, 2017: This likely won't be helpful, but could the issues be related to this one: http://mametesters.org/view.php?id=6388
by bxaimc at 6:50 PM EDT on September 4, 2017: At least MGS is getting some love this late in the game. Other games like FFXII's psf2s are still broken with missing instruments and side-hacks like Xenosaga II's Suzuki sound driver hack are still nowhere near fully playable except for like the first couple seconds (if even that). And then Xenosaga I is completely out of the question with that Procyon studio SMD/SED/SWD combination of sequences.
by zubren at 6:13 AM EDT on September 13, 2017: Nisto, this is probably a stretch, but you wouldn't happen to have the early MGS pilot demo discs would you? There is a bunch of early / unreleased tracks on those, likely in the same format as the final MGS release.

There are three MGS Pilot discs:
PAPX-90044
SLPM-80254
SLPM-86098

I'm very curious to see what audio is on those, I've seen some videos of unused content being played back via memory editing an emulator.
by Nisto at 6:38 PM EDT on September 13, 2017: The only thing I found on SLPM-86098 was an early version of REX'S LAIR. It has less percussion mostly, that's all.

s16a_01.psf

I use a script to automate creating PSFs for MGS -- the PSFs generated from PAPX-90044 are identical to the ones generated from SLPM-86098, and there are no additional tracks. So nothing new there. SLPM-80254 have less tracks, but that's it, nothing is different as far as the music goes.

edited 8:13 PM EDT September 13, 2017
by zubren at 12:49 AM EDT on September 14, 2017: Huh, I wonder if the videos I saw with the memory editing were just offsetting instruments or something, interesting nonetheless.

Thank you for looking into it for me! This is really cool!
by zubren at 6:01 AM EDT on September 14, 2017: Any chance this also works on MGS: VR Missions?
by Nisto at 11:31 PM EDT on September 14, 2017: Wow, I did actually manage to patch the original MGS executable to work with MGS Integral as well (there was an issue with the instruments). I had a theory to fix it, and it was quite a long shot, but it was quick and easy enough to do, and seems to have fixed it for all tracks.

I haven't yet checked through these PSFs, so I'll let you guys do it. Let me know if you find anything that sounds strange.

https://www.sendspace.com/file/2iymfo

Note that the unpacked size is rather large (>500MB) ... I'm too lazy to do minipsfs for this game, I guess.
by zubren at 5:17 AM EDT on September 15, 2017: First off, I want to thank you so much. I have been waiting for YEARS for movement on the MGS1 audio format. This is like a gold mine for me. Forgive the wall of text! :)

I did not find The Best Is Yet To Come (Unused) in the MGS Integral PSF archive. I used an application to sort through dupes and went through every unique file. Maybe the different discs have different file contents, despite the names being the same? (It looks like it just appends numbers to the end of the stage name, so perhaps the number of files are the same?)

Would you be willing to post full PSF archives of MGS1 Standalone, VR Missions Standalone and the three pilot discs just for giggles? (Or, are you comfortable sharing your script in it's current state?)

MGS Integral is a different build than the standalone versions, alternate skins for Snake, Meryl, Cyborg Ninga, First Person Mode, Very Easy Mode + MP5 Silenced SMG, PocketStation support, etc. Likely a lot of changes and they probably slimmed down their build in the process.

Anyway, here are my finds from the MGS Integral archive you just uploaded.

***Neat Finds***
Integral:
s05a_03.psf - 140.66 Integral Hidden Frequency - Theme of Solid Snake (from Metal Gear 2) in MGS1 style. Very cool.

s12a_03.psf / s12c_03 - 140.66 Integral Hidden Frequency - Zanzibar Breeze (from Metal Gear 2) in MGS1 style.

s10ar_03.psf / s10a_03 - 140.66 Integral Hidden Frequency - Full Discovery track, the compiled Metal Gear Solid (1998-09-03)(KCE Japan)(Konami).7z file uses brf_01.psf which is the Briefing menu track, which is a short into on the OST track. I always wondered where the full Discovery track came from, it never played in the original game. I never knew Integral had a hidden frequency with hidden tracks.

s09ar_01.psf - Wolf Cavern track with some light ambience.

VR Missions:
vrtitle_03.psf / vrsave_03.psf - Unused Theme of Solid Snake variant meant for title screen. Different from Integral's hidden frequency version.

***Broken Stuff***
I am playing these back in mednafen, so the results below are from mednafen-0.9.47-win64.

Integral:
I suspect these are Rex's Lair, given that s16a_01 from SLPM-86098 was a variation of Rex's Lair. I did not find a working version of Rex's Lair in any of the other files in the archive.
s16a_01.psf - No Audio
s16b_01.psf - No Audio
s16c_01.psf - No Audio
s16d_01.psf - No Audio
d16e_01.psf - No Audio
s16a_02.psf - No Audio
s16b_02.psf - No Audio
s16c_02.psf - No Audio
s16d_02.psf - No Audio
d16e_02.psf - No Audio
s16a_03.psf - No Audio
s16b_03.psf - No Audio
s16c_03.psf - No Audio
s16d_03.psf - No Audio
d16e_03.psf - No Audio

VR Missions:
vrtitle_01.psf - Strange sound in right channel in loop after intro. Perhaps this is just how the file is? This is clearly a variation of the Introduction track, but I don't believe Introduction was ever used on the title screen for VR Missions as it had it's own: vrtitle_02.psf

I suspect these are the Genola (Giant MGS Guards) tracks. I didn't find a working version of the Genola tracks in any other files in the archive.
vcd_n05_01.psf - No Audio
vcd_n05_02.psf - No Audio
vcd_n05_03.psf - No Audio
vcd_n05_04.psf - No Audio
vcd_n05_05.psf - No Audio
vcd_n05_06.psf - No Audio
vcd_n05_07.psf - No Audio

***Final Note***
I can't seem to find the Evasion track that occurs when you escape line of sight of the guards. Example: https://www.youtube.com/watch?v=20UtG003AeQ

It seems to fade out from the Encounter track. Maybe hidden behind an optional loop point somewhere?

edited 5:18 AM EDT September 15, 2017

edited 5:19 AM EDT September 15, 2017
by Nisto at 12:53 AM EDT on September 16, 2017: Well, I'm glad there's some interest in this. Makes me feel like a month of reverse engineering wasn't a complete waste. Thanks for sharing your finds.

The track numbers for each stage between the various releases doesn't necessarily match the same song. The song numbers are added automatically by my script; it just loops through the amount of tracks specified in the MDX header. The unused version of The Best Is Yet To Come is also not present in any release prior to the final release (SLPM-86111) as far as I remember. So they probably removed it by the release of MGS Integral, making it exclusive to SLPM-86111 (and maybe the US equivalent).

If I had to guess, the problem with the PSFs with no audio / bad instruments is probably due to one of the following reasons:

A: The driver requires more patching to be compatible with MGS Integral's data. I'm out of creative ideas here, though. Maybe at this point it would be easier to just rip the original MGS Integral driver... provided that most of the code looks the same, otherwise I'm honestly not up for ripping it.

B: The tracks are muted, so the track volumes must be manually adjusted (see Silent Hill). Although, I did convert them to MIDI using the tools by SECA Project, and at least it doesn't seem like they're muted. But then, maybe the converter isn't processing all volume-related commands? I don't know...

C: An emulation bug in Mednafen?

As for ripping "VR Missions Standalone" -- do you mean the US release? Isn't disc 3 of MGS Integral supposed to be the same? I don't actually have any of the US releases of MGS1, so I can't help in that case. I'm okay with sharing my script (more of a kit, really), but I'll assume you have some basic knowledge with programming as you'll need to configure the script for your own needs. And you'll need Python 3.

https://www.dropbox.com/s/4joqco4kkih8zxx/mgs_psfkit.7z?dl=0

About the Encounter track: there's a special music mode called Kaihi Mode (which of course translates to "evasion mode") which is triggered specifically for that track. I think they just adjust the volume of some of the sub-tracks to get that effect. It's not a separate song. That's not to say it's impossible to rip it, if you really wanted it. However, does it have to have the normal variation playing at the beginning? Because frankly, I think I'd have a hard time figuring out how to time it.
by zubren at 5:18 AM EDT on September 16, 2017: MGS1 is pretty much my favorite game of all time, and my favorite soundtrack of all time.

As for the files that didn't play, I tried them with foo_psf and they did play, albeit broken due to the way foo_psf reads these files. So, looks like it's C. They were the Rex's Lair / Genola tracks + dupes of other tracks, so nothing new there. Seeing as Rex's Lair extracts fine from SLPM-86111, it's no issue.

In terms of VR Missions, there are technically 3 builds. NTSC-J: Integral, NTSC: VR Missions, PAL: Special Missions.

Integral has PocketStation support, plus a requirement to check saves from MGS1 to unlock the Ninja missions.

VR Missions dropped PocketStation support, and is standalone so Ninja missions have an alternate method for unlocking.

Special Missions requires the MGS1 discs, it's really strange. It loads voices from the MGS disc into memory, and then plays the rest on the Special Missions disc. So you need to disc swap in order to play.

Anyway, given the way software builds work, I'm always curious to dig into different releases in case someone forgot to remove some unused tracks.

I also do know some Python, so this is pretty convenient for me. Thank you for sharing your kit! I am pretty busy this weekend, but when I get to playing around with it, I will post my results here.

As for the Encounter track, if you were up to fiddling with it and seeing if it is just fading sub-tracks in or out, that would be awesome. I'm not super picky, so I don't need the transition.
by Nisto at 3:28 PM EDT on September 16, 2017: Oh, hah, it seems I don't even have to do the timing myself for the evasion mode. There's already some delay after triggering it.

https://www.sendspace.com/file/g5kmwv
by Kirishima at 2:18 PM EDT on September 17, 2017: Viopsf has been updated and now supports mgs: https://ux.getuploader.com/foobar2000/download/195
by Nisto at 2:41 PM EDT on September 17, 2017: That's terrific. Even the Integral tracks that wouldn't play in Mednafen works! Thanks for sharing the news, Kirishima.

edited 2:41 PM EDT September 17, 2017
by mogikihei at 5:13 PM EDT on September 17, 2017: @Nisto
According to viopsf author (anonymous), it seems to be a problem of HLE BIOS.
HLE BIOS code has changed significantly, so PSF of other games may be affected.
There seems to be no problem with MGS PSFs you made.
by Kirishima at 6:57 PM EDT on September 17, 2017: @mogikihei

All the psf's that I have and tested so far haven't had anything broken so far, and that includes some hacked up Final Fantasy 6 psf's I made a long time ago as a test.

I will post here if I come across something in the future.
by Nisto at 8:35 PM EDT on September 17, 2017: @mogikihei good to know that there's nothing broken in my code. Thank you once again for relaying the information. I'll check some other PSFs for errors as well.

EDIT: While it's not viopsf(1), I noticed viopsf2 doesn't play back the PSF2s from Silent Hill 2 properly. The note delays/releases are a bit too short. This is especially noticeable in Ordinary Vanity (BGM50000) and Laura Plays The Piano (BGM50016). It would be great if the developer could have a look at it.

I was contemplating running foo_psf alongside viopsf1, but unfortunately foo_psf doesn't allow disabling PSF1 support, so there would probably be a conflict.

edited 3:29 AM EDT September 18, 2017
by Kirishima at 2:47 PM EDT on September 24, 2017: @Nisto

A new version of the viopsf2 plugin was released. It should fix that Silent Hill 2 bug, but you probably should check for yourself: https://ux.getuploader.com/foobar2000/download/196

edited 2:48 PM EDT September 24, 2017
by Nisto at 4:32 PM EDT on September 24, 2017: @Kirishima: For the most part, it seems fixed. I've been comparing it to foo_psf. It's hard to tell, but it feels like there's still a very subtle difference in release times or maybe the reverb. It could just be a placebo effect though.

I'm also just noticing that there's a constant crackling/clipping noise in BGM50011 (Pianissimo Epilogue), especially noticeable in the right channel (unless I just have a bad left ear?), between 0:02~0:10 and other parts. It's not audible in foo_psf -- or at least not as noticeable.

Perhaps the issues are related?
by Kirishima at 8:47 PM EDT on September 24, 2017: I think it's noted in the readme that there's some cracking, if that's what "プチノイス" is. If I read correctly, the author doesn't know what's causing the problem, but says disabling adpcm caching fixes it. However, I don't see an option to test that.

EDIT: I finally tested the track across HE for winamp, foo_psf, and viopsf. Highly Experimental and Viopsf have the same amount of popping, foo_psf still has it, but much less than the other 2. It's probably like a couple of Fear Effect 2 and The Crow tracks I have that no matter what is used, they will have popping.

edited 9:05 PM EDT September 24, 2017
by Nisto at 9:21 PM EDT on September 24, 2017: I think he just removed the ADPCM caching in the code. And yeah, I also think that's what プチノイス refers to. I honestly don't think it's all that noticeable in Ordinary Vanity though. More so in BGM50011. Oh well. Thanks for the heads up either way.
by mogikihei at 2:13 AM EDT on September 25, 2017: @Nisto
According to viopsf author (anonymous),
BGM50016 has 18 seconds of silence, that is, music is played back after 18 seconds.

Please set "Detect Silence" to 20 seconds.
by Nisto at 2:32 AM EDT on September 25, 2017: @mogikihei: I am aware of that. How come you are mentioning it? That shouldn't be related to the noise issues.
by mogikihei at 2:43 AM EDT on September 25, 2017: I translated what the author mentioned, Ignore if you know it.

(I misunderstood from the lack of understanding.)

edited 2:51 AM EDT September 25, 2017
by Nisto at 3:19 AM EDT on September 25, 2017: Ah, okay. I guess he thought that I thought that it was silent? Anyway, I didn't mean to sound disrespectful.
by Kirishima at 12:57 AM EDT on September 28, 2017: I just remembered something. Isn't Kode54's aopsf based on mame's emulation? Because I was testing out the foo_psf plugin included with viopsf, and I'm sorta reminded of this bug in mame when playing the mgs psf's with it: http://mametesters.org/file_download.php?file_id=4567&type=bug

edited 12:59 AM EDT September 28, 2017
by Mouser X at 5:55 PM EDT on October 11, 2017: I was trying to listen to Mega Man Battle and Chase, but the PSFs wouldn't play on Winamp, and crashed Foobar2000. I quickly realized that Winamp wouldn't play because it couldn't find the "DRIVER" file (that's the full file name), due the combination of in_zip.dll (archive reading plugin for Winamp), and the lack of a file extension (in_zip will extract PSF, miniPSF, and PSFLIB files. But without an extension, it doesn't know that "DRIVER" needs to be extracted). As for FB2k, it's a separate component that's crashing (foo_wave_seekbar), and there's no update for it.

I extracted the PSFs (with the "DRIVER" of course), and they play in Winamp (it can find "DRIVER" now), but only if I disable "suppress opening silence." Doing that, though, results in "PRE_SND0_1.PSF" only playing for 8 seconds, even though it's timed for 1:21 (including fade time). By forcing the files to play indefinitely, it plays much past the 8 seconds mark (about 2:30 for a full two loops). Apparently, the silence at the beginning of the file is counted against the overall playtime (aka, the playtime as defined by the tags).

As for FB2k, even when extracted, they still don't play. It throws the error "Decoding failure at 0:00.000 (Unsupported format or corrupted file):
"N:\VGM PP rips\PS1\M\Mega Man Battle & Chase [RockMan Battle & Chase] (1997-03-20)(Capcom)\PRE_SND0_1.PSF"" This implies that the PSF might be built incorrectly. I *did* rename "DRIVER" to "DRIVER.psflib" and I updated the tags accordingly (using Winamp, because I can't for the life of me figure out how to alter tags in FB2k. Really? Why is it so hard to do something that should be really easy?). The altered files still play in Winamp, and still fail in FB2k.

If someone wouldn't mind looking into this, and figuring out if it's a plugin problem, or if its a PSF problem, I'd really appreciate it. For Winamp, I *could* just alter the time tags, adding on the appropriate amount of time to get past the opening silence (it'd add about 1:20 to the current time, all of which is silence), but considering that they don't work in FB2k either, I think it'd be ideal to figure out where the problem lies, so that they're playable on more than one program.

Thanks in advance for anyone who's willing and able to look into this. Mouser X over and out.

edited 9:53 PM EDT October 11, 2017
by Kirishima at 8:52 PM EDT on October 11, 2017: @Mouser X
I just tested those in foobar with viopsf. They work but have an absurdly long starting silence, like 1 minute or so. I didn't rename the driver or anything. Because of this, songs take forever to startup.

edited 8:54 PM EDT October 11, 2017

edited 8:56 PM EDT October 11, 2017
by Mouser X at 11:54 PM EDT on October 11, 2017: I'm aware of the excessive silence at the beginning of the songs. It's why (as I already mentioned), in Winamp, I disabled the "suppress opening silence" in the Highly Experimental preferences. That doesn't explain why they don't play in FB2k. I'm using Highly Experimental in FB2k, not vio_psf. That implies that there's something wrong with HE in FB2k. It also doesn't explain why the files have the silence at the beginning, especially when the silence is counted against the overall play time. It'd be one thing if the original tagger took into account the excessive silence, and planned for it. They obviously didn't (since most of the songs play for less than ten seconds as a result, when they should play for about a minute).

Outputting the PSFs as WAVs shows that there's an average of 1:19.5 of silence at the beginning of each song (as I guessed when I said "(it'd add about 1:20 to the current time, all of which is silence)"). So, if they had accounted for the silence each file should be about 1:20 longer. Specifically, (to accommodate the extra silence, which is counted against the overall play time) "PRE_SND0_1.PSF" should be 2:24, "PRE_SND0_2.PSF" should be 1:55, "PRE_SND0_3.PSF" should be 2:27, "PRE_SND1_1.PSF" should be 2:38, "PRE_SND1_2.PSF" should be 1:55, "PRE_SND2_1.PSF" should be 2:03, and "PRE_SND2_2.PSF" should be 2:22. But they're not. That implies that either the player they used to time the files doesn't count the leading silence against the overall play time, or the tagger (incorrectly?) assumed that it wouldn't.

Regardless, Highly Experimental is failing to play the PSFs in FB2K (but plays them mostly fine in Winamp. I noticed a few pops/clicks, and one oddly high-pitched beep, which I assume might be emulation errors). And, while re-timing the PSFs (to make up for the leading silence) is a work-around, it doesn't seem to me to be the best method. But since I've never actually ripped a PSF, I don't know if there's a way to fix it either. Though, it looks like this set was created using Mark Grass's generic driver, so theoretically, perhaps, if his tools (and driver) have been updated since this set was ripped, maybe it'd fix the silence problem? Anyone know how to use Mark Grass's generic driver to rip PSFs? I have the physical disc for this game (somewhere.... I know bought it on eBay years ago and I played it on an emulator), so I could probably do it. If I had any idea how to, of course. Mouser X over and out.
by Kirishima at 12:18 AM EDT on October 12, 2017: You could use vgmtoolbox. It uses the latest MarkGrass driver as well as Davironica's driver. Markgrass also share a tool for making psf's with the latest driver's here.

Now that I think of it, or you using kode54's aopsf plugin by any chance? There are issues with it that make MarkGrass's older generic drivers not work on that, and he released another tool to automatically update the driver. I think it's still here somewhere.

EDIT: Forget about the update tool for this one, it didn't work. I ended up updating one of the psf's manually, and it did fix the issue. I didn't have get a strange noise while testing it with winamp, but that maybe because I had to manually rip out the vb file from the psf(probably that high pitched beep you were hearing). Didn't hear it in foobar, though.

edited 12:41 AM EDT October 12, 2017

edited 12:43 AM EDT October 12, 2017
by Kirishima at 1:30 AM EDT on October 12, 2017: Try these:

https://www.dropbox.com/s/4zulcqakvjlczna/MMB%26C.7z?dl=1

These are from the EU rom. They seem a bit off to me, but that's the best I can do. Let me know if you received them so I can remove them.
by Mouser X at 11:09 PM EDT on October 12, 2017: Thanks for the re-rip! They sounded fine to me (in both Winamp, and FB2k, which are both using Highly Experimental for PSF playback), but I'm also not sure what you meant by "They seem a bit off to me". As far as I remember, Mega Man Battle and Chase never released in the US (I bought the PAL disc, which is why I said I played it on an emulator), so the EU version (since it's in English) would have been the one I'd* be familiar with anyway. I suppose the JAP version is NTSC, so I can see it making sense to rip from that version, if it's available.

I guess now the PSFs should be "re-timed" (aka, I plan on copying the tags from the old set. I'll give them a listen, of course, to make sure the loops are good), the set updated (if they're broken, I don't see why they should be hosted in the archive, when a fixed version exists), and re-uploaded. Thanks again! Mouser X over and out.
by Kirishima at 3:42 PM EDT on October 13, 2017: It's probably just me then. Anyways the songs sounded like they had reverb from the video I watched, but I didn't bother to set any when ripping these. You can use PSF-O-Cycle to set one if you want to.
by icantthinkofagoodusername at 1:35 PM EST on November 20, 2017: I made a tread about asking how to rip sf3 3rd strike samples(and sequences, possibly) with the best audio quality possible, should i ask here instead?

edited 1:45 PM EST November 20, 2017
by Kirishima at 4:51 AM EST on November 21, 2017: @icantthinkofagoodusername

You should probably go ask at vgmrips.net where their specialty is ripping music from various arcade boards in vgm format. CPS3 ripping may not be supported in the vgm format, but you probably might get more help there.

If you're wondering why cps3 ripping isn't supported in the vgm format, the answer I got a long time ago was because there we're only a handful of games on that board, thus supporting it wouldn't be worth it...
by sunk at 7:13 AM EST on November 21, 2017: Not sure if this is the right thread, but it would be nice if someone made a gsf rip from "Asterix & Obelix: Bash Them All!". I tried using Saptapper, but it doesn't use Sappy engine. I've contacted CaitSith2 a while back, asking if he could do it, but didn't get a reply. If any gsf rippers are still here, I'd really appreciate a rip from that game. It's got some nice tunes by good ol' Alberto Jose Gonzalez.

edited 7:26 AM EST November 21, 2017
by icantthinkofagoodusername at 9:31 AM EST on November 21, 2017: @Kirishima

Thanks for the help before, and i might check it out, if i have any luck.
by Kirishima at 8:53 PM EST on December 22, 2017: This is feedback for the foo_viopsf author if he even comes here. I just found this issue with a track from Brave Fencer Musashi:

https://www.dropbox.com/s/164mix1zmiwr523/Sword%20Fight.wav?dl=1

Strange notes.
by Kirishima at 4:09 PM EST on December 24, 2017: Update to last post - It seems to be a plugin issue. The notes do not happen with the mame plugin, just mednafen and pete's.
by Nisto at 10:21 AM EST on December 26, 2017: As I was timing the Silent Hill 2 set last week, something struck me: shouldn't PSF plugins actually factor in time skipped for opening silence when processing length/fade metadata? It seems like foo_psf doesn't do that now, and that may create a discrepancy depending on whether the user has the feature enabled or not, especially when there's a lot of opening silence (like Laura Plays the Piano in Silent Hill 2).

It may be better to skip silence using a 3rd party plugins/components anyway, if you ask me. I use foo_skip for foobar2000, and not just for PSF files.

Anyway, just some food for thought.

@Kirishima: I may be wrong, but that track actually doesn't seem to play correctly in any of the plugins. What you describe seems to be because of an incorrect/corrupt instrument playing in mednafen, whereas the notes are completely inaudible in mame. That could be caused due to improper ripping though (too little space allocated for the VAB data maybe), and not necessarily due to plugins.

I think it's supposed to be a stringed instrument as you can hear on the OST CD: https://www.youtube.com/watch?v=BnaHIt9wYe8&t=6. It could just be different (i.e. badly transferred) on the console version though of course. Is there an inline rip somewhere?

edited 11:03 AM EST December 26, 2017
by Kirishima at 5:10 PM EST on December 26, 2017: @Nisto

Looks like you're right. This is from a ps2 play-through iirk, if that's close enough to what you're looking for:

https://www.youtube.com/watch?v=gHJ3gOv9Uwo&list=PL91CC02D36809A3C3&index=13
by Nisto at 5:42 AM EST on January 4, 2018: I'm also leaving a note for the viopsf developer in hopes of a bugfix (these things may not necessarily be a bug in viopsf, but I didn't have these issues with foo_psf or any other decoders I have):

foo_input_viopsf seems to have an incompatibility issue with the foo_wave_seekbar (Waveform Seekbar) component.

The waveform isn't being updated asynchronously as with all other decoders (including foo_psf). The option "incremental update of waveforms being scanned" in particular isn't working like it should. Well, sometimes it does, but only by luck it seems.

Also, sometimes the player hangs for nearly a minute, and I don't know the exact cause, but I do believe Waveform Seekbar is also the culprit there as I have now temporarily disabled scanning/analysing of files and there hasn't been any issues since.

A report of my components can be found here:
https://pastebin.com/raw/DTvtZYtJ
English Tagged HGSS 2sf Set? by Bonboon228 at 6:03 PM EST on January 13, 2018: Anyone have a English Name retagged version of the Heart Gold and Soul Silver 2sf Set? Just Curious...
by Kirishima at 6:15 PM EST on February 14, 2018: I think I found a couple sets that don't work with viopsf. Both Crash Bandicoot 2 and 3 do not work currently. They both use the same psf driver (modified Crash Bandicoot 2 driver) and probably may need to be reripped using the generic psf driver.

Quick links to both sets if someone else wants to verify:

http://psf.joshw.info/c/Crash%20Bandicoot%202%20-%20Cortex%20Strikes%20Back%20%5bCrash%20Bandicoot%202%20-%20Cortex%20no%20Gyakushuu!%5d%20(1997-10-31)(Naughty%20Dog)(SCE%20America).7z

http://psf.joshw.info/c/Crash%20Bandicoot%203%20-%20Warped%20%5bCrash%20Bandicoot%203%20-%20Buttobi!%20Sekai%20Isshuu%5d%20(1998-10-31)(Naughty%20Dog)(SCE%20America).7z
by kode54 at 8:23 PM EST on February 14, 2018: Last I checked, both of those sets require PAL timing, and include a _refresh=50 tag to indicate such.
by vajuvaju at 6:46 AM EST on February 15, 2018: This is probably why VGMtrans fails to properly convert the Crash Bandicoot 3 theme to DLS+MIDI.

edited 6:46 AM EST February 15, 2018
by Kirishima at 9:25 PM EST on February 26, 2018: I'm a bit late in relaying(for lack of a better word coming to mind), but a new revision of viopsf has been released. It fixes the Crash Bandicoot issues and possibly the others except the Brave Fencer Musashi one (that's a problem with the psf itself).

https://ux.getuploader.com/foobar2000/download/199

edited 9:28 PM EST February 26, 2018
by Nisto at 1:06 PM EST on February 27, 2018: Thanks for the heads-up once again Kirishima. I always forget to check back on that repository.

Sounds like the issue with foo_waveform_seekbar isn't/can't be fixed then. Blargh!

It sucks, really, because going back to foo_psf isn't much of an alternative either, as it cannot play the Metal Gear Solid set for example. And kode54 doesn't want to fix support for SPU interrupts either. I'm not at all sure I can patch the driver any other way either. So I'm forced to go between foo_psf and foo_viopsf for specific PSF sets. "/

@kode54 (if you happen to read this): do you know if there is perhaps any alternative way I can patch the MGS driver or something, to make it work with foo_psf? I figure you can probably easier debug the real issue with it (and where exactly foo_psf is really choking) since you actually know/maintain the foo_psf code. I'm also horrendous with C++, so I don't really want to fiddle with the foobar2000 SDK and the whole development environment it requires, to be honest.

edited 1:11 PM EST February 27, 2018
by kode54 at 3:02 AM EST on February 28, 2018: I gave away the source code. I do not want to fix it myself, because I don't want to expend that much effort. I barely worked on the core libraries to begin with. I'm not even using Windows as my primary OS most of the time.
by Kirishima at 8:22 AM EST on February 28, 2018: Isn't it based on Mame's spu core anyway?
by kode54 at 7:09 PM EST on February 28, 2018: No. It's Neill's own SPU core. SPU interrupts were even written into the spec as something unsupported, and to be avoided.
by Kirishima at 7:36 PM EST on February 28, 2018: Ok, thanks for the clarification. I thought it used mame's/audiooverload's core for some reason, and the modified foopsf that was included with viopsf had a very similar issue with the mgs psf's that mame's core had with the game itself.
by kode54 at 10:15 PM EST on February 28, 2018: Maybe I was wrong, maybe it was only SPU interrupts for PSF2 that were meant to be avoided. I don't recall. I know his original design called for having SPU interrupts be so predictable as to be able to calculate a maximum cycle count until next interrupt, then run up to that count before triggering an interrupt, if not sooner if applicable registers are changed.
by Nisto at 5:46 AM EST on March 1, 2018: I've always followed the emulation notes when ripping original drivers. As far as I can see, SPU interrupts at least isn't one of the things that explicitly should be avoided.

I do seem to recall seeing some non-DMA4 (DMA1?) appearances in the PSFLab event log, however. But I don't think I could actually trace it back to a specific instruction/address.

edited 6:16 AM EST March 1, 2018
by bxaimc at 11:00 AM EST on March 4, 2018: @Nisto: I tried experimenting with your modules from Graffiti Kingdom and Shadow Hearts but neither of them worked with Xenosaga Ep 1 =/
by Nisto at 3:54 PM EST on March 4, 2018: @bxaimc: It does seem like that's an older version of the SMD/SWD formats. I think you should be able to get it working with the Shadow Hearts driver though, since that was released around the same time. I believe whoever ripped Shadow Hearts (not me, by the way) added a small custom header (containing the file size, and a fixed (?) RPC command number if I recall it correctly; it's blurry now) to every SMD/SWD, so you may need to add that to the data from Xenosaga as well. And make sure everything is properly done of course.
by Missingno_force at 9:11 AM EDT on May 4, 2018: @Nisto: i just got linked your post about the MGS1 psf set (i dont know how i missed that, but w/e) and you mentioned having an idb. I got a couple people (myself included) who would be very interested in getting our hands on it. Do you still have it and can share it?
by Nisto at 8:56 AM EDT on May 5, 2018: @Missingno_force: Yeah, I still have it. I'm not at home at the moment though, I'll dig it up when I get home tomorrow night. May I ask what/who it's for? Always fun to see your work come to use.
by Missingno_force at 1:49 PM EDT on May 5, 2018: @Nisto: Im on a discord which does research and REing of all things MGS and kojima in general, and while we have someone who works on rewriting the PC version of MGS1 based on a disasm, we havent really gotten off the ground with the playstation version, aside from memory hacks, so your work so far would be more then welcome :)
by Nisto at 11:59 AM EDT on May 6, 2018: Okay, well, here goes.. I decided to create a repository on Github. Never know if I'll upload others in the future, or if maybe someone can contribute updates.

https://github.com/Nisto/idb

You will need a 32-bit version of IDA v6.8 or later to open this database.

Disclaimer: Please note that this is not a comprehensive database by any means, and there may be mistakes. Read the notes under the database notepad tab. Some parts of it are only theories/guesswork (but hey, I guess that's exactly what reverse engineering usually is?), and some of the terminology might not even really be a thing, or simply used improperly. Oh, and forgive any of my confusing or cringeworthy comments.
by Missingno_force at 12:45 PM EDT on May 6, 2018: Awesome, thanks a lot!
by Kirishima at 12:47 PM EDT on May 13, 2018: More viopsf feedback. Crash Team Racing appears to not be playable probably for similar reasons to Crash 2 and 3 (PAL rip). It sorta is playable if skip silence on start is disabled/set to 0, but apparently there is alot of silence in these files.

I was contemplating re-ripping the game with markgrass's generic psf driver, but the readme stated the game is using its original driver for features a generic driver can't provide.

Anybody know of other pal rips that could be problematic?
Metal Gear Solid PSF's by Enclave123 at 7:49 AM EDT on June 22, 2018: I wonder if there any recommended ways to play the PSF of MGS? it not really playing well in Winamp for some reason. It have no sound when playing them.

I am really happy to see a PSF of MGS since i really enjoy the soundtrack.

Is there going to be a MGS VR missions PSF in there future?
by Nisto at 9:04 AM EDT on June 22, 2018: @Enclave123: I recommend viopsf with the Mednafen SPU DLL. Most PSF players don't have (proper) support for SPU interrupts, so that's why they won't play it correctly (if at all).

I uploaded VR Missions (as part of MGS Integral) a few pages ago, but it has since disappeared from Sendspace (guess no one uploaded it to Josh W or elsewhere). It's a rather large set since it's not a minipsf set. But I have the "PSF kit" still up on my Dropbox if you want to build the PSFs yourself. You need to know what you are doing though; there are some pretty crucial values in the Python script you might want to configure before you run it:

https://www.dropbox.com/s/4joqco4kkih8zxx/mgs_psfkit.7z?dl=1

edited 9:17 AM EDT June 22, 2018
by Kirishima at 1:29 PM EDT on June 22, 2018: If only it could have the alternate reverb that's in the "Peopsound" plugin. That's one of the few that sounds similar to a real spu's reverb rather than most of the others which, to me, sound like an enhanced snes reverb.

Also since I'm here, viopsf was updated to fix Crash Team Racing some time ago: https://ux.getuploader.com/foobar2000/download/201
by Nisto at 12:55 PM EDT on July 25, 2018: @ viopsf developer

Can you please update to SDK v1.4? It's still listed as a legacy decoder in foobar2000 v1.4, which finally has a stable release.

edited 1:12 PM EDT July 25, 2018
by Nisto at 12:06 PM EDT on July 30, 2018: I caved in and decided to install VC++2017 in an attempt to build viopsf for foobar2000 v1.4 myself. After spending the whole evening with this yesterday, I think I finally managed to do it. But as I have zero past experience with the foobar2000 SDK, and not much experience with the extended features of C++ (templates, classes, and all that), I can't guarantee a proper build. I hope someone who perhaps does have better experience can point out any mistakes I may have made!

Here's what I did to build it for v1.4

1. Install VC++2017 (I got the free Community edition of Visual Studio 2017 personally)

2. Create a folder named viopsf

3. Download the foobar2000 SDK and unpack the contents directly into the "viopsf" folder

4. Download viopsf and unpack the contents of the src folder in src.7z directly into the "viopsf" folder. You should now have three folders in the "viopsf" folder created in step 2 -- "viopsf", "foobar2000" and "pfc"

5. In the viopsf\foobar2000 folder, create the directory tree test\components. Without this, Visual Studio fails to execute a copy command at the end of the build process, even when run as administrator.

6. Open viopsf\foobar2000\foo_input_viopsf\foo_input_xsf.cpp

7. Change line 368 to:

class input_xsf : public input_stubs

8. Add this code between line 997 and 998:

static const char *g_get_name() { return "PSF Decoder"; }
static const GUID g_get_guid() { return XSFDRIVER_GUID2; }
static const GUID g_get_preferences_guid() { return XSFDRIVER_GUID1; }

9. Open the viopsf\foobar2000\foo_input_viopsf\foo_input_viopsf.sln file (solution) in Visual Studio

10. Build foo_input_viopsf by right-clicking the item in the left-hand pane and clicking "build"

edited 1:57 PM EDT July 30, 2018
by Kirishima at 3:50 PM EDT on August 11, 2018: @Nisto

Viopsf has been updated to SDK 1.4.
Soma Bringer by RetroFanatic at 9:38 AM EST on December 22, 2018: No success in ripping this one yet?
by SmartOne at 11:38 PM EST on January 5, 2019: in_vgm 0.40.9 was released!

That's right; for the first time since October 19, 2017, we're able to used the "nuked," cycle-accurate YM2612 core to listen to VGMs!

https://vgmrips.net/forum/viewtopic.php?f=3&t=112
https://github.com/vgmrips/vgmplay

But if I were you, I'd skip that official Winamp plugin (for use in XMPlay, of course; you're not crazy, are you?) and get the unofficial foobar2000 port plugin, creatively named "VGM input" and maintained by mysterious Japanese:

https://ux.getuploader.com/foobar2000/download/213

"VGM input" is functionally superior because you can configure the core type in the GUI, unlike "in_vgm," which requires a secret INI setting change. Here are the completely intuitive GUI help tooltips:

https://github.com/vgmrips/vgmplay/commit/f712603913050feca455dcd547f981dc1f2ebc9a

Also, the foobar2000 plugin sounds less bad on manual track change. I believe the correct technical term is "less poopy."

It's a great time to be alive, folks. Accurate YM2612 (which has been available in Genesis Plus GX for over a year)! Much hype.
by marcusss at 7:01 AM EST on January 6, 2019: Im sure it is a false positive but my antivirus..and foobar also reports it a virus..

"Failed to load DLL: foo_input_vgm.dll
Reason: Operation did not complete successfully because the file contains a virus. "

Anyway disabled AV it installed it and works fine ;-)

Thx for the update

edited 7:03 AM EST January 6, 2019

edited 7:04 AM EST January 6, 2019
by Nisto at 7:46 AM EDT on August 24, 2019: Question for kode54:

With foo_psf, what happens when the TEXT size specified in an EXE header is LESS than the actual decompressed size?

The reason I ask, is because Mega Man X6 (both the old rip and re-rip) has this anomaly, but both rips still play back fine.

Just trying to get an idea of how best to handle this, as it is a problem in psfisolate as well (it only loads the TEXT size specified in the header, so it won't find any samples in the case of the old Mega Man X6 rip).

I tried looking at the foo_psf source code, but can't quite make it out. I see there's a condition for when the decompressed size is LESS than the specified size in the header in the "psf_load_section" function, but nothing for the opposite case.

edited 7:48 AM EDT August 24, 2019
by kode54 at 7:37 PM EDT on September 2, 2019: See:

https://bitbucket.org/losnoco/foo_qpsf/src/master/psf.cpp

Notice I completely ignore t_size and only go by the decompressed size of the EXE block minus the size of the header, with some checks to verify we won't totally overflow the emulated RAM size.
by Nisto at 12:53 PM EDT on September 3, 2019: Oh, guess I was looking in the wrong project to begin with. Thanks for confirming!
by kode54 at 8:53 PM EDT on March 18, 2020: Speaking of looking in the wrong project, I've renamed the projects to make more sense, so more people like you don't run into the same issue again.

foo_psf -> foo_input_aopsf

foo_qpsf -> foo_psf

And the official source currently lives on my personal Gitea server, which I will occasionally remember to mirror on other places if I update the source.

Note that the Gitea server will be down for a brief while today, as it is migrated to a new server.
by Lunar at 11:32 AM EST on November 1, 2020: I converted the Famicompo mini contests from NSF to NSFE with timers. 1,300+ files.
by Lunar at 6:01 AM EST on November 4, 2020: Double Moon Densetsu / Legend of Doublemoon (NSFE) - tags, timers and internal playlist.

I'm surprised that this Dragon Warrior clone's great soundtrack by Motoaki Takenouchi went neglected for so long. There are three unknown songs; let me know if you recognise any of them.
by hcs at 11:00 AM EST on November 4, 2020: Thanks for the new NSFE sets Lunar!
by Kirishima at 5:51 PM EDT on April 27, 2021: I'm not sure if this is the best thread to ask, but apparently vgmtrans recently updated their ps1 akao converter to support more earlier games like FF7 and PE. I'm just wondering if anybody has their own up to date build since there isn't one available on the git repository.
Re-ripping Yu-Gi-Oh FM by ArtFromCode at 8:29 PM EDT on May 28, 2021: IS anybody going to re-rip Yu-Gi-Oh! Forbidden Memories? Most of the tracks have sounds missing.
EDIT: Seems to only affect the title screen.

edited 8:40 PM EDT May 28, 2021
by maruru at 7:06 AM EDT on June 26, 2021: Is it possible to rip music from Zook Hero ZX4?
by Lunar at 7:37 AM EDT on September 11, 2021: I have tagged and timed the USF set for International Superstar Soccer 64, aka RrRrRrRrRrRonaldinho meme football 64. And not before time!

One unknown "song" (a cool samba drum riff) and three unknown jingles, if you can identify where any of them are use in-game do let me know.

My previous tagged sets (Bomberman 64 Japan, Star Soldier, Deadly Arts, possibly others) haven't appeared on usf.joshw yet, I'm not entirely sure what the submission process is these days beyond plopping links on the forums here? I would like them uploaded. I'd like to take them off of my Dropbox storage soon. Here they are again, one last time:

Bomberman 64 Japan
Star Soldier: Vanishing Earth
Deadly Arts (forgot to mention this one before.)

edited 11:19 AM EDT September 11, 2021
by Knurek at 1:21 PM EDT on September 11, 2021: All 4 up on server @lunar
by Lunar at 8:13 AM EDT on September 16, 2021: Based, thanks Knurek. <3
If I can remember anything else I've tagged & timed I'll let you know. There probably are things I'm just crap about remembering to share them.
by SoundGLSI at 5:36 PM EDT on September 16, 2021: Something I've found: In the Sonic Colors set for the Nintendo DS, there's a mini2SF that doesn't play with the current 2SF decoder (I use foobar2000), it is called "18 - Theme of Sonic Colors.mini2sf". This is the theme. I've tried to re-rip it to see if the problem is solved, but it doesn't. It should be included in NCSF format, then?
In the set are various mini2sfs that can't be played, such as the "WiFi Ver." files.

edited 6:06 PM EDT September 16, 2021
by SoundGLSI at 3:26 PM EDT on October 24, 2021: I've noticed a few SPC rerips, and I've encountered a song that doesn't play very well in players. It is "08 Wild Life.spc" from Pop'n Twinbee. The previous version of the rip played fine.

Edit: Oooh, "14 Mechanical World.spc" also plays wrong at the beginning like the other track I mentioned.

edited 3:36 PM EDT October 24, 2021
by Lunar at 7:07 AM EDT on November 6, 2021: It's unfortunate about those new Pop'n Twinbee SPCs. I appreciate having the sfx removed the beginning of each stage song but the residual issues with playback aren't an acceptable compromise. Hopefully it gets resolved.

Meanwhile I'm back with another tagged & timed set: Wario's Woods.
I didn't download every SNSF set, but the sets I have downloaded are all untimed & untagged. I may attempt Sim City 2000 later.

EDIT: Here is Sim City 2000. This one isn't 100% tagged, there are several assets I was unable to identify, and basically didn't bother with the SFX. Also not 100% certain about the composer. But hey, the not-being-looked-at-for-a-decade is the enemy of the good, right?

edited 4:41 AM EST November 7, 2021
SimCity 2000 Credits by Yntec at 10:10 AM EST on November 7, 2021: The composers of the Sim City 2000 Soundtrack were Sue Kasper, Brian Conrad, and Justin McCormick, because it's the same music of the Macintosh II version (though it features exclusive tracks).

Also, some songs have their own names (such as "Dancing Megalopolis", "Happy Land" or "Pale City"), if you want to cross-compare names here's a youtube video:

https://www.youtube.com/watch?v=ubEP88U4mhA

And here's the credit page of the music that was used for the port:

http://patcoston.com/co/CREDMAC.aspx
by Lunar at 11:59 AM EST on November 8, 2021: Thanks for the info, I've updated the pack.

Download

As you mentioned, it doesn't perfectly correlate to the Mac version's soundtrack, but there is some crossover.

edited 11:59 AM EST November 8, 2021
Pokémon Black 2 & White 2 by SoundGLSI at 12:48 AM EDT on March 16, 2022: I had this tagged B2W2 2SF set since 2019, and I recently added the 5 downloadable musicals to it. I added custom versions of Tubeline Bridge and Skyarrow Bridge, so that means they sound like the official CD. I Also added custom versions of two musicals: "Charming Munna" (without vocals) and "Pokémon Party!" (without vocals).
I included the musicals after reading this thread, but it seems that didn't go too far so I tried to do it myself.
The tagging style I used is the same used in other Pokémon rips from here, i.e. the Pokémon HGSS 2SF set.

Download link. And for the NCSF fans out there, here are the extra musicals in NCSF format.

edited 12:36 PM EDT March 16, 2022
Tempo proper file doesn't exist anymore by Nowf at 10:29 AM EDT on April 8, 2022: The sendspace link is deleted
Misadventures of tron bonne to Mp3 by Nowf at 10:52 AM EDT on April 8, 2022: Okay I downloaded the psf, how to convert it to Mp3?
Misadventures of tron bonne sendspace link is deleted by Nowf at 11:51 AM EDT on April 8, 2022: Can you please re-upload it again?
by hcs at 2:17 PM EDT on April 8, 2022: PSF_Misadventures of Tron Bonne.7z from MarkGrass in 2015, though note the concerns in this thread.

NoWool's pack (from here in 2019) might be more complete.

You can convert to wav using foobar2000; install the foo_psf component, add the files and you can right click and select Convert.

---

Please don't post several times about the same thing, this is a low-traffic forum, and you can edit posts for a day after posting them. I deleted your posts in other threads.

edited 2:38 PM EDT April 8, 2022

edited 2:39 PM EDT April 8, 2022

edited 2:41 PM EDT April 8, 2022
by SoundGLSI at 4:48 AM EDT on June 13, 2022: Recently, a new Japanese foobar2000 component (Mednafen PSF Decoder) was uploaded to the upload.jp foobar2000 components page.
It requires some steps to get it running, detailed in the readme.txt file.
by Squaresoft74 at 9:33 PM EDT on July 2, 2022: Any idea who's the author of this component ?
by SoundGLSI at 1:07 AM EDT on July 3, 2022: The developer of the component is anonymous.
by Squaresoft74 at 2:48 AM EDT on July 3, 2022: Ah that's too bad, wished i could get in touch.
I noticed in the release notes that they were planning to investigate about the current Xenogears rip "issue" where some tracks have some unwanted noise at start.
Turns out it's just a broken driver issue, nothing wrong with the component itself.
The current set also doesn't work on real hardware.
I've been working on a re-rip last year which work just fine where the current one fails.
by SoundGLSI at 11:19 PM EDT on July 3, 2022: Mednafen PSF Decoder got an update.
by vajuvaju at 11:36 PM EDT on July 9, 2022: Thanks for the heads up, @SoundGLSI!

edited 11:37 PM EDT July 9, 2022

edited 11:38 PM EDT July 9, 2022

Previous Page | Next Page
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

Search this thread

Show all threads

Reply to this thread:

HCS Forum Index
Halley's Comet Software
forum source

User Name		Tags: bold: [b]bold[/b] italics: [i]italics[/i] emphasis: [em]emphasis[/em] underline: [u]underline[/u] small: [small]small[/small] Link: [url=http://www.google.com]Link[/url] [img=https://www.hcs64.com/images/mm1.png]
Password
Subject
Message