Shockwave .DCR file format: Difference between revisions

From Computers Wiki
Jump to navigationJump to search
(Add link to ABMP info)
(Reorganize article)
Line 1: Line 1:
These are some WIP notes on the '''Shockwave .DCR file format'''.
These are some WIP notes on the '''Shockwave .DCR file format'''.


The file is composed of a series of chunks. Each chunk has a header consisting of a four-byte identifier and either a four-byte integer or variable-length integer for the length of the data, followed by the data itself. The identifier may be either forwards or backwards depending on the endianness of the system that created the file. The format of the chunk size (four-byte or variable) depends on the chunk type. Headers in the following document will be identified primarily by their little-endian names and secondarily by their big-endian names.
== Format ==


Variable-length integers are multi-byte sequences. Each byte encodes seven bits of the integer, with the high bit being 0 if this is the end of the integer, and 1 if not. The lower seven bits of the first byte correspond to the highest seven bits of the decoded integer.<ref>https://github.com/Earthquake-Project/Shockky/blob/1308dc7ff0ac2347dcd5f3879bce1eac54e1992e/Shockky/IO/ShockwaveReader.cs#L173</ref> For example, the variable-length integer <code>81 3a</code> (binary <code>10000001 00111010</code>) represents the integer <code>ba</code> (decimal <code>186</code>, binary <code>00000000 10111010</code>).
* 0000-0003: <code>RIFX</code> (little) or <code>XFIR</code> (big)
* 0004-0007: Size of file minus 8, always little-endian (? was LE on a BE file (Hannah and the Ice Caves))
* 0008-000b: <code>MDGF</code> (? for BE -- assuming this is <code>FGDM</code> for LE?)
* 000c-000f: <code>revF</code> (<code>Fver</code> for LE?)
⚫
* 0010-001f: Binary data, followed by a human-readable version string (for HatIC: <code>0f8a 0101 8e3a 0938 2e35 2e31 2331 3034 .....:.8.5.1#104</code>)
* 0020-0023: <code>rdcF</code> (<code>Fcdr</code> for LE?)


== RIFX / XFIR ==
<code>Fcdr</code> is then followed by a variable-length integer. This appears to mark the compressed size of the following data. On HatIC, this is 0x813a, which becomes 0xba (186 decimal) according to the following algorithm.<ref>https://github.com/Earthquake-Project/Shockky/blob/1308dc7ff0ac2347dcd5f3879bce1eac54e1992e/Shockky/IO/ShockwaveReader.cs#L173</ref>


This is the first chunk in the file.
<syntaxhighlight lang="csharp">

public int ReadVarInt()
* 0000-0003: Four bytes (RIFX / XFIR)
{
* 0004-0007: Four bytes, little-endian, representing the size of the file in bytes minus eight
int value = 0;
* 0008-000b: Four bytes (FGDM / MDGF)
byte b;

do
{
== Fver / revF ==

b = ReadByte();
This chunk follows RIFX / XFIR and appears to encode version data.
value = (value << 7) + (b & 0x7F);

}
* 0000-0003: Four bytes (Fver / revF)
while (b >> 7 != 0);
* 0004-????: Variable-length integer, representing the size of the chunk data
return value;
* ????-end: Binary data, followed by a human-readable version string
}

</syntaxhighlight>
=== Data sample from HatIC ===

⚫
<code> 8a 0101 8e3a 0938 2e35 2e31 2331 3034 ....:.8.5.1#104</code>

* 0000-0005: Unknown
* 0006-000d: Human-readable version string (8.5.1#104)

The author of Shockky believes the unknown portion may be composed of multiple variable-length integers.<ref>https://github.com/Earthquake-Project/Shockky/blob/master/Shockky/Chunks/FileVersionChunk.cs#L18</ref>

== Fcdr / rdcF ==

This chunk follows Fver / revF and appears to encode help text about the compression used, and something else?

* 0000-0003: Four bytes (Fver or revF)
* 0004-????: Variable-length integer, representing the size of the compressed chunk data
* ????-end: Zlib-compressed data

=== Data sample from HatIC ===


The tool offzip decompresses this to 257 bytes of data.<ref>https://aluigi.altervista.org/mytoolz.htm#offzip</ref>
The tool offzip decompresses this to 257 bytes of data.<ref>https://aluigi.altervista.org/mytoolz.htm#offzip</ref>
Line 47: Line 60:
00000100: 00
00000100: 00


== ABMP / PMBA ==
* 00e0-00e3: <code>PMBA</code> (<code>ABMP</code> for LE?)

* 00e4-00e8: Unsure, is <code>a842 00de 13</code>
This appears to stand for AfterBurner Map.<ref>https://docs.google.com/document/d/1jDBXE4Wv1AEga-o1Wi8xtlNZY4K2fHxW2Xs8RgARrqk/edit#heading=h.x638haejy3bd</ref>


* 0000-0003: Four bytes (ABMP or PMBA)
offzip then detects 5183 bytes of compressed data (12051 uncompressed) starting at 00e9. Info.<ref>https://github.com/Earthquake-Project/Format-Documentation/blob/0a362560596f34295d1fd4f24f3e2d889938d829/structure/core/formatNotes_DCR.txt</ref>
* 0004-????: Variable-length integer, representing the size of the chunk data (including the few bytes after this and before the compressed map)
* ????-????: Variable-length integer, representing some sort of identifier, appears to always be <code>00</code>
* ????-????: Variable-length integer, representing the size of the decompressed map
* ????-end: Zlib-compressed map


== References ==
== References ==

Revision as of 22:42, 25 July 2022

These are some WIP notes on the Shockwave .DCR file format.

The file is composed of a series of chunks. Each chunk has a header consisting of a four-byte identifier and either a four-byte integer or variable-length integer for the length of the data, followed by the data itself. The identifier may be either forwards or backwards depending on the endianness of the system that created the file. The format of the chunk size (four-byte or variable) depends on the chunk type. Headers in the following document will be identified primarily by their little-endian names and secondarily by their big-endian names.

Variable-length integers are multi-byte sequences. Each byte encodes seven bits of the integer, with the high bit being 0 if this is the end of the integer, and 1 if not. The lower seven bits of the first byte correspond to the highest seven bits of the decoded integer.[1] For example, the variable-length integer 81 3a (binary 10000001 00111010) represents the integer ba (decimal 186, binary 00000000 10111010).

RIFX / XFIR

This is the first chunk in the file.

  • 0000-0003: Four bytes (RIFX / XFIR)
  • 0004-0007: Four bytes, little-endian, representing the size of the file in bytes minus eight
  • 0008-000b: Four bytes (FGDM / MDGF)

Fver / revF

This chunk follows RIFX / XFIR and appears to encode version data.

  • 0000-0003: Four bytes (Fver / revF)
  • 0004-????: Variable-length integer, representing the size of the chunk data
  • ????-end: Binary data, followed by a human-readable version string

Data sample from HatIC

8a 0101 8e3a 0938 2e35 2e31 2331 3034 ....:.8.5.1#104

  • 0000-0005: Unknown
  • 0006-000d: Human-readable version string (8.5.1#104)

The author of Shockky believes the unknown portion may be composed of multiple variable-length integers.[2]

Fcdr / rdcF

This chunk follows Fver / revF and appears to encode help text about the compression used, and something else?

  • 0000-0003: Four bytes (Fver or revF)
  • 0004-????: Variable-length integer, representing the size of the compressed chunk data
  • ????-end: Zlib-compressed data

Data sample from HatIC

The tool offzip decompresses this to 257 bytes of data.[3] 

00000000: 0300 04e9 99ac 7000 360b 0000 0800 0737  ......p.6......7
00000010: 7a34 2e98 99ac 5d00 500d 0000 0800 0737  z4....].P......7
00000020: 7a34 89a8 0472 d0af cf11 a222 00a0 2453  z4...r....."..$S
00000030: 444c 4d61 6372 6f6d 6564 6961 207a 6970  DLMacromedia zip
00000040: 6c69 6220 636f 6d70 7265 7373 696f 6e2c  lib compression,
00000050: 2063 6f6e 7461 6374 204d 6163 726f 6d65   contact Macrome
00000060: 6469 612c 2049 6e63 2e2c 2053 616e 2046  dia, Inc., San F
00000070: 7261 6e63 6973 636f 2c20 4341 004d 6163  rancisco, CA.Mac
00000080: 726f 6d65 6469 6120 6e75 6c6c 2043 6f6d  romedia null Com
00000090: 7072 6573 736f 7200 5357 4120 4465 636f  pressor.SWA Deco
000000a0: 6d70 7265 7373 6f72 2058 7472 6120 6672  mpressor Xtra fr
000000b0: 6f6d 204d 6163 726f 6d65 6469 612c 2049  om Macromedia, I
000000c0: 6e63 2e20 2028 6874 7470 3a2f 2f77 7777  nc.  (http://www
000000d0: 2e6d 6163 726f 6d65 6469 612e 636f 6d2c  .macromedia.com,
000000e0: 2063 6f6e 7461 6374 2073 7570 706f 7274   contact support
000000f0: 406d 6163 726f 6d65 6469 612e 636f 6d29  @macromedia.com)
00000100: 00

ABMP / PMBA

This appears to stand for AfterBurner Map.[4]

  • 0000-0003: Four bytes (ABMP or PMBA)
  • 0004-????: Variable-length integer, representing the size of the chunk data (including the few bytes after this and before the compressed map)
  • ????-????: Variable-length integer, representing some sort of identifier, appears to always be 00
  • ????-????: Variable-length integer, representing the size of the decompressed map
  • ????-end: Zlib-compressed map

References

  1. ↑ https://github.com/Earthquake-Project/Shockky/blob/1308dc7ff0ac2347dcd5f3879bce1eac54e1992e/Shockky/IO/ShockwaveReader.cs#L173
  2. ↑ https://github.com/Earthquake-Project/Shockky/blob/master/Shockky/Chunks/FileVersionChunk.cs#L18
  3. ↑ https://aluigi.altervista.org/mytoolz.htm#offzip
  4. ↑ https://docs.google.com/document/d/1jDBXE4Wv1AEga-o1Wi8xtlNZY4K2fHxW2Xs8RgARrqk/edit#heading=h.x638haejy3bd
Retrieved from "https://computers.huntertur.net/index.php?title=Shockwave_.DCR_file_format&oldid=499"