Programming the C64

Introduction
VIC-II Colour codes
Image-Application formats
Hardware Keyboard Matrix and Digital Joystick Scanning
GCR - Group Coded Recording
Comments to nonC64 assembler programmers

Introduction

This section is intended for programmers of the C64 (Mainly Assembler coders), or other programmers who need information about the programming of the C64. It contains some practical information and it is not the intention to give you an course in assembler programming here.

VIC-II Colour codes and Luminance intensity
	Dec Hex Colour Brightness Vice Palette R,G,B CCSC64 palette R,G,B 0 0 Black 0/0 00, 00, 00 10, 10, 10 1 1 White 4/8 FF, FF, FF FF, FF, FF 2 2 Red 1/2 68, 37, 2B E0, 40, 40 3 3 Cyan 3/6 70, A4, B2 60, FF, FF 4 4 Purple 2/3 6F, 3D, 86 E0, 60, E0 5 5 Green 2/5 58, 8D, 43 40, E0, 40 6 6 Blue 1/1 35, 28, 79 40, 40, E0 7 7 Yellow 3/7 B8, C7, 6F FF, FF, 40 8 8 Orange 2/3 6F, 4F, 25 E0, A0, 40 9 9 Brown 1/1 43, 39, 00 9C, 74, 48 10 A Light Red 2/5 9A, 67, 59 FF, A0, A0 11 B Dark Gray (1) 1/2 44, 44, 44 54, 54, 54 12 C Medium Gray (2) 2/4 6C, 6C, 6C 88, 88, 88 13 D Light Green 3/7 9A, D2, 84 A0, FF, A0 14 E Light Blue 2/4 6C, 5E, B5 A0, A0, FF 15 F Light Gray (3) 3/6 95, 95, 95 C0, C0, C0
	Note: Brightness is on a 0-4 scale on the original C64, and a 0-8 scale on C64C & C128(D)

Image-Application formats

Formats for application image types:
Start addr.	Start address/Load address (first two bytes of the file)
Length	File length in bytes (including load address bytes)
Bitmap	Bitmap offset (8000 bytes)
Colour	Colour data offset (1000 bytes). In Multicolour mode,  Copy (25*40) 1000 bytes to the area $D800 and upward.
Background	Background colour (1 byte), stored in lower nybble of $D021

Common file formats for Hires raw-images on the C64
Name (Hires)	Start Addr.	Length (Bytes)	Bitmap Offset Addr.	Screen Offset Addr.	Colour Nybbles Offset Addr.	Background Offset Addr.
Art Studio	$2000	9009	0	8000	Not used
Doodle	$5C00	9218	1024	0
Image System (Hi)	$4000	9194	0	8192

The upper 4 bits of screen memory becomes the colour of any bits that is set (to 1) in the 8*8 pixels attribute controlled by the screen memory. The lower 4 bits is the colour of any pixels cleared (0).

Common file formats for multicolour bitmap raw-images on the C64
Name (Multicolour)	Start Addr.	Length (Bytes)	Bitmap Offset Addr.	Screen Offset Addr.	Colour Nybbles Offset Addr.	Background Offset Addr.
Artist 64	$4000	10242	0	8192	9216	10239
Adv. Art Studio	$2000	10018	0	8000	9016	9001
Amica Paint (*)	$4000	10257	0	8000	9000	10000
Blazing Paddles	$A000	10242	0	8192	9216	8064
Koala Painter	$6000	10003	0	8000	9000	10000
Image System (MC)	$3C00	10218	1024	9216	0	9215
Vidcom 64	$5800	10050	2048	1024	0	2024
CDU Paint	$7EEF	10277	273	8273	9273	10273

(*) Amica Paint can save in packed format, using control-codes ($C2) for loops etc. For depacking : $C2 00 marks end of picture, $C2 xx yy means copy byte yy xx-times. Other bytes combinations must just be stored as they are. Furthermore, Amica Paint has an extra feature included in the file; The colour-Rotation-Table at offset Addr. 10000 with a total of 256 bytes.

Common formats for FLI raw-images on the C64
Application (Format)	Background *=Rasterbars	Colour Nybbles	Colour Screens (steps of $0400)	Bitmap	Notes
Saga FLI *	$3B01-$3BC8	$3C00-$3FFF	$4000-$5FFF	$6000-$7F3F	$3B00?
Interpaint (*)	$3B00-$3BFF	$3C00-$3FE7	$4000-$5FFF	$6000-$7FFF	*Options

* Rasterbars is an EXTENSION to the FLI concept.

Hardware Keyboard Matrix and Digital Joystick Scanning

The keyboard is read as a matix of 8 rows by 8 columns. The row is selected by clearing (set to 0) the corresponding bit in the CIA1 port A ($DC00). The column data bits are then read from the CIA1 port B ($DC01), and a cleared bit indicates that the key is pressed at the time of scanning.

Notice: A keypress is of course not buffered in hardware. If a key is pressed, the bits are cleared at that moment, not before or after. You can scan for more than one key at a time if you want, withing the limits of course. For a buffering input stream, you should write your own routines or use kernal calls.

For scanning Joystisk 0 and 1 (Port A and B) you use the same registers, only this time you read from Port A also if you want to scan Joystick 0 in Port A.

X = Not important in connection with Joystisk

To give a simple example how keyboard scanning works, here is an example of how to wait for a press on the SPACE key.

You can make versions which clean up the registers before and after use, or you can make quick and dirty versions for shorter code. It depends on what you need in your final code. This is just to give you an example.

GCR - Group Coded Recording

This is the disk encoding scheme of the Commodore 1541 (and similar) diskdrives , and does not take advantage of the index hole present on the disks (Some systems are Index-hole oriented) but is a SYNC-oriented format. It does however take advantage of the possibility of larger tracks towards the edge of the disk (Some systems don´t).

Each 1 bit on the magnetic media (disk) is represented by a fluxation in the magnetic field, either from North to South, or the other way. Each 0-bit is represented by a -nonchanged- magnetic field. A flux in the field generates a small current in the wires of the coil, this current is amplified and, in turn, interpreted as a High level (1) by the digital logic.

The disk is divided into tracks. Each track is divided into a number sectors. Each sector consists of a block-header and a datablock. In order to mark the beginning of a blockheader or datablock, a SYNC of ATLEAST 9 magnetic fluxations, e.g. (N)SNSNSNSNS (= 111111111) are written to disk. All this is in order to keep knowledge of where a sectors starts (without it, you would not be able to tell (in a simple way) which bits are what).

Loop data ($55 = %01010101) are part of Block-headers (10 * $55) and Data-blocks (X * $55, if needed).

The basics rules for recordings which are not Sync signals:

1. Not more eight 1-bits in sequence, recorded on the magnetic media.
2. Not more than two 0-bits in sequence, recorded on the magnetic media.

In order to achieve this, the each nybble is recorded as 5 bits-GCR code.

GCR conversion table
Dec	Hex	Binary	GCR value		Dec	Hex	Binary	GCR value
0	0	0000	01010		8	8	1000	01001
1	1	0001	01011		9	9	1001	11001
2	2	0010	10010		10	A	1010	11010
3	3	0011	10011		11	B	1011	11011
4	4	0100	01110		12	C	1100	01101
5	5	0101	01111		13	D	1101	11101
6	6	0110	10110		14	E	1110	11110
7	7	0111	10111		15	F	1111	10101

Example of GCR encoding

Four bytes plain data (in this case $12, $34, $AB and $CD) are converted into 5 bytes GCR format:

Bytes:              $12           $34           $AB           $CD
Binary:           00010010      00110100      10101011      11001101
Nibbles:         0001  0010    0011  0100   1010    1011   1100   1101
GCR codes:      01010  01011  10011  01110  11010  11011  01101  11101
Nibbles:        0101  0010  1110  0110 1110  1101 0110  1101 1011 1101
GCR bytes:       01010010    11100110    11101101  01101101   10111101
Converted:         $52         $E6         $ED       $6D        $BD

As already mentioned, the scheme takes advantage of different track length. It simply adds more sectors to each track.

Sectors/Track on 1541 drives
	Track	Sectors
	1-17	21
	18-24	19
	25-30	18
	31-35	17
	Tracks 35-40 must be supported by an extra system (e.g trackloader or systems like Dolphin Dos.)

Examples of alternative disk encoding schemes, used on other systems:

MFM - Scheme used on many diskdrives for PC etc, also used on some drives of the C64/C128.

Comments to nonC64 assembler programmers

As I have coded in assembler on a wide range of microprocessors; the 6502, 8051, Motorola 68K, 56K (DSP), Intel 80X86, Pentium and ARM (Advanced Risc Machines) based processors, I know there is variation in the syntax of how to represent something in the assembler code, I feel the need for having a small section explaining the special signs and symbols used when it comes to coding the C64 in assembler.

A guide for common signs, symbols and terms used by C64 coders in the assembler syntax.

$	Hexadecimal number; Eg. $C000 (49152 decimal) same way as 0xC000 in C and some other systems. Used alone in an assembler instruction, it normally is a memory address (e.g. STA $D020 - Store the value from the accumulator to the memory location $D020 (In this case the VIC-II chip - Border colour).
%	Binary number. Is used similar to $ (Hexadecimal)
#	Numerical value. E.g. LDA #0 - Load accumulator with the value 0 (Not the value stored at address 0). Can be combined with the $ sign to represent a hexadecimal value, e.g. LDA #$0F will Load the accumulator with the numerical value 15 (Decimal). Similar syntax is common on many other platforms.
<	LSB, Least Significant Byte. E.g. LDA #<$ABCD will load #$CD into the accumulator. The final instruction will be LDA #$CD. Usefull with constants or to make the source easy to read.
>	MSB, Most Significant Byte. E.g. LDA #>$ABCD will load #$AB into the accumulator. See <
Nybble	4 bits (1 bytes = 2 nybbles. A nybble is either the uppermost or lower 4 bits)

Titles for reading:

General readings

Commodore 64 Programmers Reference Guide
- Published by Commodore Business Machines, Inc. and Howard W. Sams & Co., Inc.

Programming the 6502
- Rodnay Zaks, Sybex. (ISBN : 0-89588-046-6)

Advanced readings

The MOS 6567/6569 video controller (VIC-II) and its application in the Commodore 64
- Christian Baur