Here's a quickbms script to extract files based on the .toc:
http://hcs64.com/files/ebtoc00.bms
Run the script with the .res or .sng file, and it will locate the .toc file if it is in the same directory.
A lot of these files have no header and are just 32khz big endian 16-bit signed PCM. The rest seem to be DSP, except the .sn0 which I have no interpretation for.
Thanks! I returned later to find that importing as 16bit Motorola @16KHz, although still beneath what I knew it should have been, did correctly interpreted the stereo image. I guess I only chose Intel before and discarded the results because of the noise level. Starting to remember based on past experience that these corporate brandnames are another representation of endianness? Anyway, it was still interleaved.
Also, what looks like severe clipping in the second half is simply a bit heavy use of signal compression by its author.
Funny, I have just been using another tool by the maker of QuickBMS... No, not funny at all after looking at his webpage. Jeez.
I've added support for .sf0 and .sfx in vgmstream r930. I don't know if there's any point in dealing with .sng and .sn0. You can get stereo out of the .sfx files by renaming mus_credits.sfx and mus_credits.2.sfx to mus_credits_L.sfx and mus_credits_R.sfx (and same with mus_ui), and vgmstream will pick up both files if you try to play either.
If anyone wants these files, I've uploaded them. It's a short tracklist, so I threw in the game sound effects to balance it out. http://www.multiupload.com/WGP7I1YF82
Anyone, feel free to post a 'proper' version on the tracker, I don't care.
First of kudos to hcs for the script its awesome getting a proper rip of this game.
I looked at the music.res file and found no songs related to the tracks (which is what I need). So I look at the track names such as Nebulatest.toc + Nebulatest.trk. [These can be opened if you rename them to .res] and found the streams in them.
Couple of issues I found:
1- Most songs with [the exception of China, tutorial and a a few others] have a right channel distortion. In fact it sounds original if you listen to it through the left channel only by muting the right channel [used audacity for this].
2- Does not loop in vgmstream r931 although I can do it manually in audacity
3- has a terrible fade at the end of the songs
Is there a way I can make this into full quality stereo? Btw looping and fade are just bonus if you'd like :)
I don't think I actually got those files, could you upload them again? Do you remember what I was supposed to do with these? I'm not too clear from the thread.
Well you updated a bms script to make it compatible with Excitebots. The music files are in a .res format (a compressed folder which requires the .toc file in order to extract it). You accomplished just that. However the music.res file only contains the menu music, intro and credits all which work perfectly through vgmstream in terms of looping, complete song and fading out.
However the level [China, Kilimanjaro, etc.] music (which I'm gunning for) are not in the music.res file which I thought they'd be located.
So I looked carefully through the iso and found the level music to be in the trk files (which also require its acompanied .toc file as well in order to extract the audio file which is in a vgmstream playable format, I forgot what extension specifically). You've updated to make the script work excellent on .res + .toc extraction. On a later post I then told you that the rest of the music was in the .trk files. Of course I'd like to add that if you run the current bms script using .trk and its accompanied .toc file it will extract the level music but messy'.
'Messy' being: starting point of the songs are halfway through the original music, terrible random static noise during the songs, and only one channel is played properly while the other is not syncronized at all and 5 different files of .sf0 (Now I remember, this is the format the music is in).
Therefore I'm asking if you can update the current bms script to make it compatible with .trk + .toc to hopefully fix the song issues. The mediafire link up a few posts should contain the trk sample. Your response made my day thank you for helping me with this.
Well, good news and slightly unpleasant news. I figured out what was up, and here's an updated extractor which should get the music out of the .trk properly: ebtoc01.bms
The trick is that the first portion of these files are compressed, I assume that is what is loaded directly into RAM. The rest, then, is streamed as needed (and thus must not be compressed in the same way).
The bad news is I'm not going to make any efforts towards decompressing that just now, though it looks like it might just be LZ77. It shouldn't be necessary for music, but I'm not sure.
edited 7:59 AM EDT May 21, 2011
That other stuff would probably be useful for setting loop points for these tracks, I suspect that the tiny .sng and .sfx files have loop info in them for the .sf0. Or maybe the track songs don't loop?
Used your updated bms and here's the follow up of what I got:
The ending portions of the songs have all been fixed Woot! :)
You mentioned that the beginning of songs seem to be compressed and I believe this results in a problem I just found out. The first couple of seconds of the songs are not in the .sf0 when converted using vgmstream. Thus making it loop very inaccurate as you'll hear a hiccup in the transition of the loop. The first few seconds of the song are really vital otherwise a decent amount of .wav editing is needed.
The sfx could be loop points I dunno, but if you manage to get it to loop that'll be the final icing on the cake.
I'd like to give thanks for co-operating with me in this and I know a lot of people would be happy when it lands its spot on tracker :)
What I mean is that the songs are missing the first 3-5 seconds so you can't even loop it properly manually using audacity. Unless of course you can fix the script/sf0 to loop.
Its been a while since any mention of this, but this just needs a 'minor' fix to complete, can anyone try to decompress the first portion of the songs in the bms script to make the songs play correctly, and possibly looping it?
I know hcs said that he'd look into decompressing the files when he had 'time', is 2 months good enough?
I've spent many hours staring at the hex without any luck, I haven't tried digging into the binary yet. If you could send it I might take a stab but it's hard to guarantee as I haven't had much luck with projects lately.
I'm still not sure what your asking. I'm a bit of a noob at this, what is the 'main executable' your talking about? Are you referring to the streams? Please be a bit specific on what I'm suppose to give.
It's not in the filesystem, you should be able to find it in the menus somewhere or by right clicking on the main partition. I don't know what software you're using.
[edit] Never mind, arbin got it for me.
edited 1:29 AM EDT October 9, 2011
[edit] Located decompression at 801666F0, should just be a matter of time now, we'll see how much time I get to study it.
The compression algorithm used for the .res files in this game is QuickLZ 1.2.0 developed by Lasse Mikkel Reinhold.
https://github.com/TheGameCommunity/EQuicklz
I have more detailed file specifications for .res and .toc files if you guys want them.
I've successfully been able to extract most files from the archives, but I don't know how to reverse engineer the sound effect audio files (specifically the files in racesnd.res). It appears you guys have had some success in the past?
Some help on this would be appreciated. The best way to get ahold of me is on the Excitebots Discord server, but I will also check this thread periodically.