SNES/Satellaview split?

[Powerpuff]

The Famicom Disk System argument is the best and sums it all. According to the No-Intro logic, the SNES and the Satellaview DATs should be split.

[xuom2]

I wait a decision by NGE, then i can split them in db.

[kiddocabbusses]

While we're on the subject with the possibility of the Satellaview split, mind if I give some suggestions on how to handle the streamlining of the theoretical new DAT? Of course, one is free to debate these if desired.

1) Last I recalled nearly every Satellaview ROM on the DAT currently has (BS) and (Japan) tags. Under a split into it's own DAT, these would basically be redundant - (BS) because in the SFC DAT that's used to distinct standard SFC from Satellaview releases, and (Japan) because there's little to no evidence that the Satellaview had any international penetration and thus all games are Japanese region only with the exception of Hacks/Public Domain material.

2) Redumps should be organized by the download date in the ROM. Since it's been brought up before - do NOT alter the dates in the ROMs for the time being, as there's potential for two perfectly verified good dumps to have different build data.

3) Have marks that differentiate games, magazines, and data pack/expansion ROM data. If further differentiation is required, then between Soundlink games and standard downloads.

[Powerpuff]

I wait a decision by NGE, then i can split them in db.

No "Nintendo - Satellaview" dat has surfaced yet at DAT-o-MATIC. No decision was taken or is NGE opposed to the split?

[xuom2]

NGE's tradition is to be very slow with reply.. no worry! Don't are we all very patient?

[kiddocabbusses]

Bumping this thread because I'm noticing a startling issue with Satellaview ROMs on some no-intro DATs around the internet and I figured this thread would be the best place to express my thoughts on it.

I'm not sure if someone at no-intro thought this was a good idea or they have been mistakenly been collecting these ROMs in this manner, but I've checked multiple (BS) ROMs in the no-intro DAT and found that they had their header values modified so that 7FD6-7FD7 are FF FF, effectively blanking the download date values. Furthermore, whoever's been doing this let a hacked Dr. Mario (BS) ROM in the no-intro DAT when I was pretty certain GoodSNES was originally distributing this pure.

I've noticed this in the BS Zelda, Dr. Mario, and Arkanoid: Doh It Again! ROM dumps and I figured after three of them that this is a disturbing current trend in the DAT.

I will express my reasons for stating that blanking the date values in the no-intro sets are a bad idea here;
1) With header dates blanked, it's more difficult to sort through the ROM dumps. In particular, if this earlier version of Excitebike: Bunbun Mario Battle ever got a good redump, it would be a lot more difficult to tell at first glance that it was any different from the common ROM if you couldn't look at the header and find the date values.
2) Hacking the date values may cause unintended side affects in emulation. For example, SNES9XPP XE refuses to load Arkanoid: Doh It Again (BS) No-Intro as a Satellaview ROM through the BIOs, but the version that was on GoodSNES, which has the date values, detects fine.
3) Removing the date values from the ROM is, at it's core, a base hack, and is against No-Intro's policy of maintaining pure dumps.

I'm certain in the case of most of these dumps that you can either go back to GoodSNES and find them, or find them on Satellablog.
My ideal for No-Intro would be that all the dumps on there be pure 8M Pack dumps with few modifications.

[C. V. Reynolds]

I didn't know about this before (my knowledge of Satellaview isn't good), but I think this is unfortunate. Is there any way to restore the header to be as original as possible? That is, to make the game as it was when it was first downloaded.

[kiddocabbusses]

I didn't know about this before (my knowledge of Satellaview isn't good), but I think this is unfortunate. Is there any way to restore the header to be as original as possible? That is, to make the game as it was when it was first downloaded.

Modifying the header is pretty easy (and in the cases of ROMs which appear to be fully functional but were hard-deleted and thus missing xFDA and xFDC-xFDF values, would be considered by some a necessary evil), but knowing the exact date values that were originally there would require knowledge of the specific ROM or the 8M Pack of origin. Especially true of things which were rebroadcast often. Header names, as well, are something that should be carefully maintained, especially since there are many ROMs that will boot right into gameplay without even a title screen. One thing to note is that regardless of whatever else the BS-X ROM may mess with on the header, both the name and date (along with ROM type) are kept until the file is overwritten outright. This may be important to note if someone goes on the prowl, filtering for ROMs with headers that have been tampered with by hackers vs. ROMs which have had the header modified by BS-X itself due to being DRM-locked/Hard-deleted.

Since as far as I can tell this started somewhere within no-intro, what I'd probably do is go back and compare every (BS) ROM to the GoodSNES counterpart and/or the ones from Satellablog. From my experience, GoodSNES, even with hacked ROMs, rarely tampered with the date values in the ROMs (Although names tended to be a prime candidate for header tinkering, usually changing a Kana-name to English for heck-knows-what-reason. I'll want to go back and check on this later but I'm now pretty certain one of the Excitebike (BS) ROMs going around has a hacked header due to think, and it'd be fairly apparent by comparing it with it's otherwise-duplicate counterparts.).

[kiddocabbusses]

Just a bit of an update on the headers issue. I've gotten some folks to help me examine more of the no-intro ROMs. I also found someone who had a GoodSNES ROM set. I'm not entirely done this yet, but I hope to have a full list of every problematic ROM shortly.

From what I can tell, the blanked-date ROMs are all ROMs that were previously on GoodSNES and were modified at some point either before on No-Intro, or some time between them being on No-Intro and the point where Callis started dumping ROMs. I've attempted digging up multiple GoodSNES sets, and from what I can find there is no GoodSNES set which has ROMs that have the date header hacks the no-intro ones got. It unfortunately seems like it could very well be every single ROM that was released before "Konae-chan no DokiDoki Pengin Kazoku" that got tinkered with in this way.

I think it would be best to find the "previous" GoodSNES ROMs and swap out the hacked header ROMs for those, to start. In nearly all the cases it's obvious the hacks were done -from- the GoodSNES ROMs anyway. (I noted in the current no-intro DAT that Dr. Mario (BS) was replaced with one of the Satellablog redumps. I'd suggest having the GoodSNES one over that as well, if this is all the case.)

It may also be a good idea that, while this is being done, every ROM in the DAT has the date on the ROM listed in the filename.

The ROMs that were dumped from the moment I started Satellablog onwards do not appear to have been altered in this manner.

I'll probably post later going into deeper detail if necessary. For now, I'm trying to figure out which ROM sites still have GoodSNES ROMs so I can do a comparison on my own.

[C. V. Reynolds]

Hey. Just appearing to mention that after a little Satellaview education (much thanks to Kiddo for the examples), I went through every Satellaview ROM image (may have missed one, tell me if so) and:

A: Added the date back into the header. I believe that when the No-Intro dat was first built, the source images used were the Goodsnes files that were marked as "[h]" (IE: hacked). These hacked images were indeed a part of the Goodsnes sets, and though the modification was not done by No-Intro, the adoption of these hacked images indeed was. These were the files that had their header dates blanked (rewritten as "FF FF"). While I understand why this was originally done (for purposes of unifying the same game, different broadcast ROMs), the fact is that these games never had a blank date header in the wild, and it is thus inaccurate to present them that way, no matter how many possible combinations of dates are possible for those headers. I assume we can agree that accuracy wins over convenience here.

Additionally:

B: On any game that had "80" (locked) as its boot byte, I changed the byte to read "84" (one boot remaining) instead. As emulation grows more accurate, it will likely become natural to emulate the Satellaview lockout. To ensure that these games remain bootable while maintaining source accuracy, I employed an idea (not my own) to "unlock" these games. It is impossible to know with certainty how many boots these games had at their times of download, so restoring the game to its original number of boots is not possible. However, at some point in their lives, these games did in fact have at least one boot remaining, and so it is safe to change this byte to "one boot remaining" without damaging the integrity of the dump. I'm aware that some might think this is controversial, but for the reasons I stated, I don't think it's a problem. There is literally no difference between a ROM image dumped (properly) when it has one boot remaining and one altered to have one boot remaining.

The points from A and B should be the new standard at No-Intro.

Unfortunately, there are certain Satellaview images that I could not find header dates for, and I could not repair those games as a result. Here is the list of those games:

BS Marvelous - Camp Arnold Course - Dai-1-shuu
BS Marvelous - Camp Arnold Course - Dai-2-shuu
BS Marvelous - Camp Arnold Course - Dai-3-shuu
BS Marvelous - Camp Arnold Course - Dai-4-shuu
BS Marvelous - Time Athletic Course - Dai-1-shuu
BS Marvelous - Time Athletic Course - Dai-2-shuu
BS Marvelous - Time Athletic Course - Dai-3-shuu
BS Marvelous - Time Athletic Course - Dai-4-shuu
BS Zelda no Densetsu - Inishie no Sekiban - Dai-1-wa
BS Zelda no Densetsu - Inishie no Sekiban - Dai-2-wa
BS Zelda no Densetsu - Inishie no Sekiban - Dai-3-wa
BS Zelda no Densetsu - Inishie no Sekiban - Dai-4-wa
Kodomo Chousadan Mighty Pockets - Chousa 1 - Junk-ya Black no Ie
Kodomo Chousadan Mighty Pockets - Chousa 2 - Hanzai Toshi Big Apple

If anyone can find what dates were in these headers, please let me know. Other than those, the issue with the missing dates should be a closed case.

Should anyone find any inaccuracies in the Satellaview section at No-Intro, please let me know. I know I still have work to do in this section.

PS: I haven't gotten around to documenting all of the header dates for the file-names, but I shall get to it in the future, unless someone else beats me to it.



EDIT: It has been brought to my attention that editing the boot byte on the Soundlink games may be an error. Upon download, the games immediately booted, and upon a timed expiration, they immediately became unbootable (and with an "80" for the boot byte); no Soundlink game has ever been found without 80 as its boot byte. My lack of Satellaview knowledge really bit here, and my method wasn't as foolproof as I thought it was (and I am dumb). Anyway... It may be that these games always have 80 as their boot byte (the "unbootable" byte) and have a unique method to boot regardless of that byte. We can wait for the experts such as Kiddo to collect more info on this, but it may be wise to revert the Soundlink games' (the ones I changed anyway) boot byte to 80 where applicable.

[kiddocabbusses]

BS Zelda no Densetsu - Inishie no Sekiban - Dai-1-wa
BS Zelda no Densetsu - Inishie no Sekiban - Dai-2-wa
BS Zelda no Densetsu - Inishie no Sekiban - Dai-3-wa
BS Zelda no Densetsu - Inishie no Sekiban - Dai-4-wa

Check to see if the non-hacked ROMs at the BS Zelda Homepage will do here. I think they also have Marvelous ROMs, you may want to check those as well.

I'll dig around for Mighty Pockets 1 and 2. These were all previously on GoodSNES at some point so...

[C. V. Reynolds]

Check to see if the non-hacked ROMs at the BS Zelda Homepage will do here. I think they also have Marvelous ROMs, you may want to check those as well.

I'll dig around for Mighty Pockets 1 and 2. These were all previously on GoodSNES at some point so...

I checked those before, and just checked again. Unfortunately, they're missing the dates as well. I did not know about the Satellaview Marvelous ROM images on the site before. Unfortunately, they too are missing the dates. Thus, the search continues.

[kiddocabbusses]

Guh, I guess this means the missing dates thing pre-dates No-Intro and came from something else, and may have "normalized" in some other area of the ROM keeping scene while GoodSNES tried to keep dumps with dates where applicable.

I went on UG, got the "SNES Scene Collection" and fetched out the very original "Scene releases" of Time Athletic 2-4, and the dates are STILL missing from them. It appears these and Sekiban may be lost causes. Did they somehow get DUMPED like that? Why these, in particular, when dates are the norm in most the other ROMs? The unfortunate question is, how do I know if the scener released his ROM as dumped, or tinkered with these values before release?

The BS Marvelous games seem pretty normal otherwise. Huh.. As did the Mighty Pockets ones.

Unfortunately if these ended up scene release'd like this, the only solution I can think of is placing in dates that match one of the known broadcast dates (easiest ones would be the premiere dates, unless we got confirmed rebroadcast builds).

[Connie]

I would be tempted to use the dates from the following site >>> http://en.wikipedia.org/wiki/List_of_Sa ... broadcasts <<< like this:

BS Marvelous - Camp Arnold Course - Dai-1-shuu >>> November 3, 1996
BS Marvelous - Camp Arnold Course - Dai-2-shuu >>> November 10, 1996
BS Marvelous - Camp Arnold Course - Dai-3-shuu >>> November 17, 1996
BS Marvelous - Camp Arnold Course - Dai-4-shuu >>> November 24, 1996
BS Marvelous - Time Athletic Course - Dai-1-shuu >>> January 7, 1996
BS Marvelous - Time Athletic Course - Dai-2-shuu >>> January 14, 1996
BS Marvelous - Time Athletic Course - Dai-3-shuu >>> January 21, 1996
BS Marvelous - Time Athletic Course - Dai-4-shuu >>> January 28, 1996
BS Zelda no Densetsu - Inishie no Sekiban - Dai-1-wa >>> March 30, 1997
BS Zelda no Densetsu - Inishie no Sekiban - Dai-2-wa >>> April 6, 1997
BS Zelda no Densetsu - Inishie no Sekiban - Dai-3-wa >>> April 13, 1997
BS Zelda no Densetsu - Inishie no Sekiban - Dai-4-wa >>> April 20, 1997
Kodomo Chousadan Mighty Pockets - Chousa 1 - Junk-ya Black no Ie >>> September 7, 1997
Kodomo Chousadan Mighty Pockets - Chousa 2 - Hanzai Toshi Big Apple >>> September 14, 1997

Kodomo Chousadan Mighty Pockets - Chousa 3 - Kyakusen Queen Patra no Nazo (Japan) (BS) >>> March 22, 1998 <<< Please check this.

It seems to tally up considering 3/4 dates for 3/4 parts.

It's the best I've found so far.

[kiddocabbusses]

For reference, here's a list of every Soundlink game and their "premiere" date.

While this is decent enough for a general ballpark, one major issue with simply putting the premiere as a placeholder date is that it's very possible the data from the premiere date may not match a "Saihousou" rerun, or that for other reasons there may be two different builds/revisions of a broadcast. This Excitebike dump, though technically a bad dump with some attempted restoration, shows a very good example of what can be expected.