Oh, silly me. I never posted my updated research done in JANUARY. Most of the experiments were the same, but I *did* test some more CDs.
So, for your reading pleasure, here is the latest (and, mind you, I'm nearly LAUGHING at myself when I say that--7 months really isn't "new").
This experiment is being run by Grant Guenther (grantguenther@mac.com). Please do not redistribute this document unless it is in full and also has the current date.
Version History:
1.0 (Grant Guenther, Jan 7, 2001): Original typeup. 1.0.1 (Grant Guenther, Jan 8, 2001): Noted possiblity of a secondary lookup table. Added Correlations and Unanswered Questions sections.
This document is best view with a fixed font (such as Monaco).
0123456789 ..........
Test Results:
MR2-Control CD Burn Time: 24:02:37 CD Type: Audio Tracks: 6
Track Start and Total Times: 00:02:00 04:00:07 04:02:07 04:00:07 08:02:14 04:00:07 12:02:21 04:00:07 16:02:28 04:00:07 20:02:35 04:00:00
Note: Unlabeled CDs were CDs that were burnt incorrectly. However, they may still provide information that will later be useful.
Coorelations & Summery:
CD Tests A modified the number of tracks, but kept the same burn time. Result: No change. However, there are very few test samples. The difference may have to be greater than 3 to affect the creature type.
CD Tests B modified the number of frames, but kept the same number of tracks. Result: No change. However, Unlabeled 2 also differed in the number of frames (as well as the number of tracks) but its type was quite different. This still lends the possibility that the number of frames does, indeed, affect the type of creature produced. This is an area that should be investigated further.
CD Tests C modified the number of seconds, but kept the same number of tracks. Result: Slightly different creatures were produced. The total seconds do, indeed, affect the type of creature generated. That, or each track time determines the type of creature.
CD Tests 2-A are determining the span of frames until a new creature is generated. This research has not been concluded yet.
Unanswered Questions:
1) Does the number of tracks actually affect creature type? 2) Are each track time used to calculate what creature is produced, or is it the total time? One way to test this is to create a CD in which all tracks are 4 seconds, and one is 5. On the second CD, swap one track for the next. Do this again. If the same monster is produced each time, odds are that individual track time is not considered, but the total burnt time. 3) By how much does the number of frames have to be in order to change the creature type? 4) Does MR2 also actually READ data off the drive? Does it use block size and block data to determine the type of creature and its stats? All tracks, thus far, have been of silence.
Of Particular Note:
1) There are 392 creature cards. This may be used in calculating which creature is generated. (Note: this may even be incorrect. It may not be possible to pull some creatures from CDs. The value may be 255.)
2) All tracks burnt were of silence. Another test will determine if the actual data affect the creature type or stats.
3) All CDs were burnt using the same program with similar parameters. They were all burnt using the same brand of CDs. Later I try out similar burns with CD-RWs (in order to save the expense of burning CD after CD).
Personal Conclusions & Theories:
1) Although the first few CDs (1B, 2B) demonstrated little evidence that the frame-time did not affect the type of creature was produced, CD (Unlabeled 2) countered this evidence. It is quite possible that the difference of one, two, or three frames do not affect the type of creature, but the difference of 5 or more may. This should be investigated more thoroughly.
2) Consider this: If the main and subtypes were calculated strictly by times and track numbers, that means any breed--ANY--are possible. However, the game clearly does not allow certain combinations (such as Undine/Beaclon). I feel that there is a table used to calculate the type of monster (and, along with it, a template of abilities). Furthermore, I believe that this table is the same used for the card numbers. However, without proof, this is merely a theory.
Further, the lookup table may not be precisely the card numbers; there may be a secondary lookup table that calcuates which card matches such creature. This would explain the erratic patterns of the C-test CDs.
3) The number of tracks did not affect the type of creatures (1A, 3A), or at least in these tests. It is still my personal opinion, however, that the number of tracks do affect the type of creature, but the difference must be greater than 3.
Why do I believe this? CD databases use a CD's total time (actual burn time) and number of tracks to uniquely identify CDs. I don't believe that it's coincidence that Mariah Carrey's Christmas album just HAPPENS to produce a Satan (Santa) Claus, or that Tecmo's CDs just so happen to produce the special creatures they do. I suspect that there is a preset "exception" case statements just prior to the equation to calculate the table entry for each creature.
I don't believe that Tecmo would overlook this simple check and throw out this piece of input.
4) But why would different pressings of the same CD, same track times, etc, produce different monsters? The answer would be in the frames.
Let's say you own BMG music. You have a deal with Griffen to press you own copies of My •••• Christmas by the Foo Bars. However, there are plenty of pirates out there who are more than willing to forge your company's name, logo, etc etc, to appear to be the real deal.
But how can you PROVE that these pirates are stealing your CDs? One method is to read the internal information of the CD. But this can be easily copied and forged. The method to deter and reveal piracy is by minor changes. Dollar bills have watermarks; CDs have frames.
There is usually a silence between tracks. (There is *always* a lead-in time for audio CDs). Let's say that Griffen uses 2.1 seconds as a track-buffer time. If you, owning BMG, want to differentiate between your CDs, Griffen's, and a pirate's, what should you do?
CHANGE THE BUFFER TIME. Instead of 2.1 seconds, use 1.9. A precise CD reading program will tell you the *exact* burning times. And, if you were a pirate, you may overlook this little detail and burn your pirated copies at 2.0 seconds.
Of course, I may be making the simple very complex (or violating Occam's Razor). Human beings are at the heart of burning, and still are prone to mistakes. What if the buffer time, at the first pressing, was 2.5 seconds, and the next was 2.25? A very minor change, but would greatly affect the total burnt time. Or, even simplier, perhaps a given track was 5:20:03 and, remastered or redubbed, it is now 5:20:44 (0.547 seconds added). The time printed on the CD cover would not be updated (for a half-a-second change??), CD players would not show the exact change, but any program that pulled the exact burnt time would notice (such as CDDB or (gasp!) Monster Rancher 2).
5) Seconds definately change the type of creature produced. In these examples, the creatures were all the same main type and differred in sub-type. Their cards were also very close to each other (however, the card values themselves differed as the seconds increased; in other words, an increment of 1 second did not necessarily increment the card count by one).
6) Offhand, I do not see any "magic number" in 392, the total possible creatures that could be generated from a CD.
7) Possibility of a secondary lookup table: Tecmo *did* change the type of creatures generated between their lines. For example, MR1 will generate one type of creature with a given CD, but MR2 will generate a different one (even if the creature from MR1 exists in MR2). This makes sense--what fun is it to "find" new creatures when each title creates the same creature, time and time again? Solution: A secondary lookup table.
--Cats
By Quincunx on Tuesday, August 7, 2001 - 08:24 am:
Um... isn't this the same research posted as What Do The Games Read From CDs? An Experiment Into The Depths! under Ongoing Research?
Oh, wait... you've added more to it. But I just knew I'd heard of Griffen and the Foo Bars before. «grin»
By CatsGodot on Tuesday, August 7, 2001 - 08:32 am:
I was thinking that too--my gut was telling me, "Hey, you never posted this 'cause you never really FINISHED THIS...and since it's SO FREAKIN' LATE, you MIGHT AS WELL." So I compared between the two, and almost didn't post it, 'cause I thought they were the same.
However, the one that's posted is version 1.0, and this is 1.0.1. The major difference I noted was the Gaboos.
And, funny, I completely forgot about the Foo Bars until I noticed the posting here with the A ???? Christmas...
--Cats
By Quincunx on Tuesday, August 7, 2001 - 08:40 am:
Ah... what a difference a day makes. Actually, I kind of expected those two CDs would make identical creatures before I looked at the results, but that's only because of all the other research I've perused here.
I'm still firmly in the "frames are irrelevant except for rares" camp... with the caveat that if you add up enough frames over several tracks, you're talking about real seconds. «smile»
Home
LC