[Note: This is technical information for people who are interested in writing their own software to modify samples in the LSDj ROM images]
How The Sample Kits Are Laid Out
Each sample kit is $4000 bytes long. The kits start at offset $20000 in the ROM image. All in all, there is space for 21 kits, of which the last 4 kits are reserved for the speech synthesizer.
Sample Kit Contents
Each sample kit can contain up to 15 samples.
When a kit is used, it is mapped to the memory area $4000-$7FFF...
ADDRESS CONTENT
$4000 $60
$4001 $40
$4002-$401F Sample end addresses *)
$4020 0
$4021 0
$4022-$404E Sample names **)
$404F 0
$4050-$4051
N / A
$4052-$4057 Kit name (6 chars)
$4058-$405B
N / A
$405C Force Loop (1 bit/sample; samples
8–1) ***)
$405D Force Loop (samples
15–9 (MSB not used))
$405E-$405F
N / A
$4060-$7FFF Sample data (4 bits/sample point @ 11468 Hz, in chunks of 16 bytes)
*)
The end addresses specify the address of the first byte after the sample. So if the first sample is laid over the area $4060-$473F, the end address is $4740. If the sample is not used, set the end address to 0.
The addresses are 16 bits long, stored in little-endian order (meaning: address $4250 should be stored as $5042 in memory).
**)
Each sample has a name of 3 chars. When entering a sample name with less than 3 chars, pad with '-'. For unused samples, fill in the sample name with null chars.
***)
The force loop bit is used for sample kits, so that some vocals can be looped automatically: for example aaaaaaa
contains some more information concerning Rom structure
http://wiki.howardforums.com/index.php/Understanding_ROM_Structure