===== page 1 =====
“Nicole Eggert”
Vice snapshot with Vice palette
Made with the GIMP from a NE photo
and converted to C64 320x200
Hires Mode Bitmap
by Stefano Tognon
in 2006
“EVS memory”
...
Free Software Group
1
===== page 2 =====
SIDin 10
version 1.00
16 July 2006
2
===== page 3 =====
SIDin Contents
General Index
Editorials...............................................................................................................................4
News.....................................................................................................................................5
Goattracker 2.23/2.35.......................................................................................................5
JCH New Player 21.g4 Final............................................................................................6
Fap 0.2.3..........................................................................................................................7
SID DUZZ' IT v2.0 beta....................................................................................................9
SIDPLAY for MAC OS X 3.4.............................................................................................9
HVSC Update #45..........................................................................................................10
JCH New Player 21.g5 Final..........................................................................................10
Tiny Sid 2 Compo result.................................................................................................11
Goattracker 2.4/2.5.........................................................................................................11
Big Sid Compo................................................................................................................12
Richard Bayliss Interview!...................................................................................................13
Tiny Sid 2 (part 1)...............................................................................................................16
XL5 (256b)......................................................................................................................16
Mini Digi Drum (512b).....................................................................................................21
RM (512b).......................................................................................................................25
Miniature Rhapsody (256b)............................................................................................32
Conclusion......................................................................................................................35
Goattracker 2......................................................................................................................36
Start................................................................................................................................36
Improvements.................................................................................................................40
Conclusion......................................................................................................................41
3
===== page 4 =====
Editorials
Stefano Tognon <ice00@libero.it>
Hi, again.
This is the 10° issue of the magazine.
I did not imagine that when I started to wrote SIDin, I will reach a so high number of issues.
I very hope to continue with this trend.
I report here a comment by Frantic about the previous number that corrects some my mistakes:
Some comments on my player was clearly wrong (the tune is long not because of filtersweeps but because
of different combinations of the songsequences, and also, not mentioned explicitly but still, my player wouldn't
benefit from being placed in ZP since the code to copy it would outnumber the bytes earned), but it's a nice ef-
fort and a nice issue in general. I like it! Player coding nerds unite! Don't think I have time to participate in Tiny
Sid compo #2 though, which is too bad since that would give me the opportunity to demonstrate that almost
the same player structure would indeed be capable of complex sounds too (although I didn't focus on that in
the former compo entry tune).. ;)
Thanks for this.
In this number we see some entries from the inside of the Tiny Sid Compo 2, the others will be in
the next number. I think that we need more musician/coder that are able to produce tiny music, as
this little music are essential for having a full features minigame, so don't blame if in all numbers I
present such mini players :)
In the end article, I show how I use the new Goattraker 2 for creating a cover of a tune. This was
for me a way to learn the new features of the program. I so don't go to describe all the features of
GT2, as the Readme is full completed of all you need for use the program, but all the process in-
volved for produce the tune.
Bye
S.T.
4
===== page 5 =====
News
Some various news of players, programs, and competitions:
• Goattracker 2.23/2.35
• JCH New Player 21.g4 Final
• Fap 0.2.3
• SID DUZZ' IT v2.0 beta
• SIDPLAY for Mac OS X 3.4
• HVSC Update #45
• JCH New Player 21.g5 Final
• Tiny Sid 2 Compo result
• Goattracker 2.4/2.5
• Big Sid Compo
Goattracker 2.23/2.35
The Cadaver PC music editor available at http://covertbitops.c64.org is available:
v2.23
• Fixed SHIFT+INS in wavetable (inserts $E9 $00 now as it should).
• Added SHIFT+U to unlock table scrolling.
• Added SHIFT+F5/F6 to change speed multiplier.
v2.24
• Added SHIFT+F7 to edit the hardrestart ADSR parameter.
• Added SHIFT+F8 to switch between 6581 and 8580 SID models.
v2.25
• Added detail to table overflow error message in relocator.
• Added stop of playback if a tablepointer gets pointed directly to a jump (would bug in
packed/relocated tune).
• Added relocator error message for the above case.
• Added relocator error message for illegal song restart position.
• Not possible to save erroneous packed/relocated songs anymore (now that there is ade-
quate error location).
• Added SHIFT+
,. to move song startposition on all channels and restart last playmode
v2.26
• Fixed relocator complaining of FF parameter in speedtable.
• To increase stability, in multispeed mode relocator doesn't use the "same gatetimer/1st
wave" optimization.
• Changed Shift+O to optimize the currently edited table, not just speedtable.
• Added an option to optimize everything with the SHIFT+ESC clear function. This clears un-
used patterns, instruments and table-entries.
v2.27
• Fixed optimize function with different pattern lengths.
• Added marking & copy/pasting orderlist data like patterns & tables
v2.28
• Added sort of mouse control.
• Sound reinitialization when changing multiplier or other parameters is now faster.
v2.29Beta
• Internal reorganization.
• New assembler-based relocator that leaves out unused portions of player code.
• Instead of choosing a player version, one can choose combinations of player features.
• Relocator does not save selfcontained duplicate parts of tables.
• Added keys 0-9 & A-Z in the fileselector to move to the corresponding part of the filelist (if
5
===== page 6 =====
found).
• Added SHIFT+H to calculate left/right shifted "hifi" speedtable values, according to the fre-
quency of a note.
v2.3
• Fixed songdata optimize function.
• Fixed relocator with no-hardrestart instrument in slot 1.
• Fixed relocator handling of effect C (Set cutoff).
v2.31
• Fixed effect copy/paste.
• Added possibility to disable pulseskipping with /O command line option. Warning: causes
huge rastertime increase.
• Tempo 2 is allowed. See the tips section (3.6) for details.
v2.32
• Maximum pattern length increased to 128. Do not complain of "pattern too complex" errors :)
• Optimized playroutine init in case of only 1 subtune.
v2.33
• Fixed a bug where relocator was checking pattern 208 (illegal).
• Optimized playroutine in case there is no keyoff/keyon used.
• Optimized playroutine in case there are no pattern effects used.
• Optimized playroutine channel variables in case there is no transpose, repeat, wavedelay
and pattern effects.
• If not using the last instrument (63), you can control song startup default tempo by using its
Attack/Decay parameter.
v2.34
• Added relocator optimization for using 1 or 2 channels only. To enable, you must use only
channels 1 or 1,2 and not be using sound effect support or zeropage ghostregs.
• Turning off buffered writes in relocator options works now even if sound effects or zeropage
ghostregs are selected (they will be turned off too).
v2.35
• Auto-creation of speedtable entries will always generate a new entry (except for song load-
ing), duplicate values will be optimized by the relocator.
• Pressing SHIFT+ENTER over a speedtable-using pattern command while the parameter is
zero, goes to the next empty speedtable entry and sets the command parameter to point to
that entry.
For version 2.25 it is available even for Mac OS X.
JCH New Player 21.g4 Final
Released in January 2006 the player coded by Laxity to use with JCH
Updates in Final:
● Fixed a bug in the vibrato feel, that caused an overflow and made the note go out of tune..
● Fixed the synchronization (initialization) bug.
Download from:
http://noname.c64.org/csdb/getinternalfile.php/14838/NewPlayer v21.G4 Final.zip
6
===== page 7 =====
Fap 0.2.3
Fap is a new simple and lightweight audio player by "Dr. Fiemost"
.
It features a dirlist view, where you can browse through your disk and bookmark directories in
which you've stored your music, and a playlist view which is automatically filled with all the support-
ed files contained in the selected directory
Formats currently supported:
• modules: mod, s3m, xm, it
• compressed modules: xmz, itz, mo3
• sidplay: sid (includes songlength db
support)
• ogg-vorbis:ogg
• others: wav, voc, au
Changelog:
0.1.0 2006-01-12
• first public release
0.1.1 2006-01-16
• fixed a couple of bugs
0.1.2 2006-01-23
• Some improvements with songlength
db usage
• Use memory mapped files for OGGs
7
===== page 8 =====
reading
• Modules can now play at 48KHz
• 8 bit output is disabled because it's broken
• Support for WAV, VOC and AU files
0.1.3 2006-01-28
• Some minor fixes and changes
0.1.4 2006-02-04
• Fixed some regressions introduced in last release
• Always start with ascending sorting
• Added a dialog for displaying song info
0.1.5 2006-02-22
• Experimental support for hardsid
• Now it's possible to disable SID filter
• Sort type is now saved again
• Various bugfixes
0.1.6 2006-03-01
• Doubleclick on a song start playing
• Display current and total subtunes
• Select between playlist and song mode: the first play all the songs in the playlist, the latter
play only selected song
• Various bugfixes
0.1.7 2006-03-06
• Display OGG comments with correct codec (UTF8)
• Experimental support for mp3 using libmad
• Various bugfixes
0.1.8 2006-03-12
• STIL support
• Now it's possible to disable unwanted plugins at compile time
• Various bugfixes
0.1.9 2006-03-27
• If both "filename
_
PSID.sid" and "filename.sid" exists display only first
• Various bugfixes
0.1.10 2006-04-09
• Fixed a possible error with samplerate conversion
• Some minor fixes and changes
0.1.11 2006-04-23
• When a song ends and directory has been changed start playing new playlist
•
* Added ffmpeg plugin with support for wma, ra, aac and other formats (need recent ffmpeg
form cvs, enable with --enable-ffmpeg-plugin)
• Bookmarks are associated with plugins
• Added playlist support (pls, m3u)
• Sndfile plugin now detect song duration
• Formats supported by the sndfile plugin are automatically detected at runtime
• Fixed a memory leak
• Some minor fixes and changes
0.1.12 2006-05-10
• Fixed an error when configuring with --disable-src
• Some minor changes
0.2.0 2006-05-27
• Now it's possible to create and save custom playlist
0.2.1 2006-06-04
• Fixed opening of compressed modules when extension is not lower case
• Fixed comments for platform other than Unix/Linux
• Ogg and Mad backends can play 8 bit
• Some minor changes
8
===== page 9 =====
0.2.2 2006-06-12
• Fixed config dialog: cancel didn't work as expected
• Fixed a bug when saving settings
• Added Game Music Emulator backend
• Fixed a bug with non-PSID file filtering
0.2.3 2006-07-15
• Added random sorting option to playlist
• Do not reopen STIL and songlength DB if HVSC path is not changed
• Added two options to the configure script:
• --disable-psid-filter - disables non PSID filtering in Sidplay backend
• --disable-noise-shaping - disables noise shaping in Mad backend
• Fixed opening files with spaces in filename in Modplug backend
Download at http://www.geocities.com/drfiemost/fap.html
SID DUZZ' IT v2.0 beta
SID DUZZ' IT is the C64 music editor by GRG/Shape. A public beta release was released into the begin of
this year:
• Player sources have been rewritten and optimized.
• New filter effects and detuning in the player.
• Some editor routines have been fixed.
• Fixed two bugs in the dumper. (Tracks that was off was not dumped properly.)
• Load/Save/Dump and dos commands are new.
• Works nicely on CBM, CMD and IDE64 drives.
Download at http://home.eunet.no/~ggallefo/sdi/
SIDPLAY for MAC OS X 3.4
SIDPLAY for Mac OS X is a complete rewrite of SIDPLAY for Mac OS, which was originally released in June
of 1996. Version 3.4 was released in March 2006
News in 3.4:
• SIDPLAY is a universal binary now and runs great
on the new Intel-based Macs
• Added Spotlight plug-in and changed search feature
to use Spotlight
• Added scrollbar for info window
• New anti-aliased oscilloscope rendering using
Quartz 2D
• Update to latest libsidplay2 source
• Updated to PSID64 0.7
• Fixed oversampling problem
• Fixed bug in PRG export if path to SIDPLAY appli-
cation contains a space character
• Added songlength display for RSID tunes
• Added display of SID chip model settingFixed vari-
ous major and minor bugs
Download from: http://www.sidmusic.org/sidplay/mac/
9
===== page 10 =====
HVSC Update #45
HVSC Update #45 was released on April at www.hvsc.c64.org
After this update, the collection should contain 32,000 SID files!
This update features (all approximates):
663 new SIDs
1004 fixed/better rips
11 fixes of PlaySID/Sidplay1 specific SIDs
8 repeats/bad rips eliminated
466 SID credit fixes
135 SID model/clock infos
17 tunes from /DEMOS/UNKNOWN/ identified :-)
10 tunes moved out of /DEMOS/ to their composers' directories
7 tunes moved out of /GAMES/ to their composers' directories
Main Composers featured in this update:
(Artists marked with NEW are either completely new to the HVSC or they get their own directory
in this update)
DRAX Laxity Barry Leitch
Jeroen Tel Arthur Keleti (Arthur) Dwayne Bakewell
Richard Bayliss Bizet Butka
Dalezy Dr. Piotr Hein Holt
Josstintimberlake Harold Klink (NEW) Klax
Monk (NEW) Nata Psycho
Randy Reason Rygar
Sidder Skysurfer Square (NEW)
TC Van Styer Watchman (NEW)
Wuiti Xonic the Fox (NEW) Yaemon
Yuro Zabutom (NEW)
JCH New Player 21.g5 Final
Released in May 2006 the player coded by Laxity to use with JCH
Updates in Final:
● Fixed the periodical retrig bug, in liveplay mode.
● Implemented stop ($fe in pattern list). Forgot that completely in pre .g5. DOH!
Download from:
http://noname.c64.org/csdb/getinternalfile.php/22927/NewPlayer v21.G5 Final.zip
10
===== page 11 =====
Tiny Sid 2 Compo result
The second Tiny Sid Compo was concluded on 21 April 2006 with this result:
256B RESULT:
Pos. Points Title Author
512B RESULT:
-------------------------------------------------------------------
1 96 BLOCK ACID DUB Frantic
2 94 XL5 Stefano Tognon
3 86 Miniature Rhapsody Eric Odland
4 55 Back To Basics Jaymz Julian
Pos. Points Title Author
-------------------------------------------------------------------
1 103 Plaster GRG
2 92 RM Stefano Tognon
3 64 Resolution Eric Odland
4 57 Mini Digi Drum Leif Bloomquist
Download the result pack with comments from: http://digilander.iol.it/ice00/tsid/tinysid2
Goattracker 2.4/2.5
The editor has reach a major update as the internal file format is changed as new features was
addes:
v2.4:
● Added one bit of accuracy to pulse modulation speed. A new song- format (identifier GTS4)
reflects this, old songs will be converted upon loading.
● Added possibility to execute pattern commands from the wavetable.
v2.41
● Fixed player code assembly errors when using some pattern commands only from the
wavetable.
● Fixed bug with command BXY when command AXY is not used.
● Fixed SHIFT+O removing table entries pointed to from wavetable pattern commands.
v2.42
● Fixed frequency table size determination.
● Added SHIFT+ENTER when parameter is zero to get you into an empty table entry for 8XY,
9XY and AXY pattern commands as well.
● Added adjustment of table views when changing instrument number in unlocked table
scrolling mode.
● Added ENTER over nonzero instrument number (in patterns) to go editing that instrument.
v2.43
● Added duplicate pattern detection/removal to song optimize.
● Added pattern shrink/expand (SHIFT+O,P).
● Added pattern join/split (SHIFT+J,K).
v2.44
● Fixed editor playroutine to not execute tick n commands, when a command has already
been executed through the wavetable.
v2.45
● Added possibility to disable realtime pattern command optimization at tick 0 (/Rxx command
line option or click the "RO" text on the statusbar) for tempo-independent portamento & vi-
brato.
● To make room on the status bar for the realtime optimization status indicator, finevibrato &
pulse optimization indicators have been shortened to "FV" and "PO"
.
11
===== page 12 =====
v2.46
● Added transpose of speedtable portamento speeds (SHIFT+Q,A,W,S) SHIFT+W,S (octave
transpose, multiply/divide by 2) also work in pulse & filtertable.
● When editing the wavetable, RETURN & SHIFT+RETURN work for wave-table executed
commands as in patterns (go to table position pointed by command, create new speedtable
entries)
v2.47
● Added experimental random delay to write of reSID register, in order to make potential
ADSR-bugging on C64 audible. The amount of delay can be adjusted with /Z command line
option.
● SHIFT+L in pulsetable creates repeated "set tempo" commands if speed is greater than 127.
● Unbuffered playroutine timing slightly changed.
v2.48
● Fixed playroutine to use configurable baseaddress again.
v2.49
● Added /G command line option, guard 1stwave/gatetimer parameters. When this is on (de-
fault), only silent waveforms can be entered as firstwave (only thing that makes sense) and
editing gatetimer of one instrument affects all of them, as otherwise bugs in playback result
easily. Use /G0 to turn off.
v2.5
● Fixed "optimize everything" removing the tempo setting in instrument 63.
● Fixed default tempo of multispeed tunes in packer/relocator.
Big Sid Compo
This is a tentative to produce some big sid tune where more composers will work onto the same
peaces using the same editor. I try to make this in a compo like style, so maybe this could stimulate
the composers.
Check at: http://digilander.iol.it/ice00/tsid/bigsid and look at the forum
Submission is open until 2 December 2006
12
===== page 13 =====
Richard Bayliss Interview!
by Stefano Tognon
This time I go to interview Richard Bayliss that maybe you know for having played one of his innu-
merable number of games. Richard is very active and I think that in game music he is very talented.
The interview was made in this month.
Hello, Richard,
could you give some words about you and your real life?
I am 27 years old, will be 28 on Christmas Day. I work in a pharmaceutical distribution warehouse,
picking and packing orders as well as maintain the warehouse, keeping it the way it should be. I
don't mind the work that I do. I am a keen C64 programmer & musician (You might have already
known before). I originally wrote games since 1991 using the Shoot 'Em Up Construction kit. It was-
n't until 1994 I learned BASIC programming games and composing music, after purchasing a utilities
tape from a Public Domain library and played around with the editors. It wasn't until 1999 I learned
to code games in Turbo Assembler, then 2003 (or was it 2004) was where I started Cross-Assembly
:o)
Maybe you are the unique scener around today that can release a game a week? (even if
some people blame for the not optimal gameplay of many of your past games, but music is
always good). Have you some games now in develop?
Yes, one of which (I'm hoping to get finished before Christmas 2006, although there is a dealy possi-
ble) is called Sub Hunter. This game is based on the Vic 20 version of Sub Hunt by Mastertronic
software, but is more up to date with parallax scrolling, a few other sub games. Secondly is a co-op
game production with The Art Ravers, called "Crash Course"
, a kind of maze game, where you are
playing against an opponent and have to drive around the maze, collecting items to boost your
score on over 20 different courses. This sounds interesting.
From time to time you release a Music collection onto disk. Have you already planned a new
one in the future?
Ah, you're on about my DMC albums, etc. Yes, I am hoping to do a few more music collections in
the future, but I have not confirmed when.
What are the music editors that you use to composer sid tunes and why have you choose
them?
At first it used to be the Music Assembler and Music Mixer Module 5 (Formly known as Voicetrack-
er), because it was small in size and fun to compose. I tried Future Composer V4, it wasn't that easy
to use. I then moved on to DMC V2.0 which was a whole lot easier to use, as the utility disk had ex-
ample tunes. I needed to play around with the example tunes before I was able to create my own in-
struments. The composer which I used the most were between DMC V4.0 and DMC V5.0. At first I
originally had a version of DMC V4.0B by Graffity/Panic Design, found on the Magna Musica utility
disk. Sadly this version had a bug in the tune packer. So it wasnt until 2001 I had found a 100%
working version of this composer on the Internet (Between 1995-2000 I had no Internet access at
home at the time).
13
===== page 14 =====
I got in contact with a few people and one person in particular had introduced me to DMC V5.0. I
had found this to be one of the most complex version of DMC at the time, so I had to experiment a
lot with the features that were included in the DMC music composer. Occasionally, but not all the
time, I compose on Goat Tracker music editor, but don't feel comfortable with the sounds that are
used in this program. It does not beat the traditional real C64 old/new SID chip.
Are there some particular features that you want to find in an editor that are missing from to-
day ones?
Goat Tracker lacks the *real C64* SID environment, which I think would be great if it could emulate
exactly the same thing. Also with this composer, I miss the good old DMC V5 environment, for the
sector editor, where you could try and control filtered sounds to go up or down like in DMC V5 itself.
It is not a bad composer though.
Now some quick final (standard) questions:
Real machine vs emulator: what do you think of?
Real machine. Emulators cannot beat the real C64, because of its capabilities to sound and vision
handling. With an emulator, like VICE it does not sync sound correctly when playing C64 games.
For example a game written in SEUCK. Press fire to shoot something, your weapon appears but the
sound comes after the weapon move slightly.
6581 vs 8580 chip: any (musical) preference?
I like the new SID (8580), because of the features with the sound filters. Especially when it comes to
using cool composers like DMC V5.
What is the worst sid that you compose and the better one?
The music I wrote for Fake Game. It was that rubbish, but I only composed it as a little joke for the
title screen. It made me laugh after I composed it. I suppose I really should have preserved it for an
entry for the 2006 Crap Game compo Block Defenz, but instead I *programmed* the music with the
Turbo Assembler. :)
Who are your best sid authors?
The best SID authors? In the past, Martin Galway did amazing ambient tunes, like Insects in Space
and Parallax. I loved both of those tunes, not to mention the title music for Miami Vice. Secondly has
to be Rob Hubbard, one of the greatest legends. I loved his work on Mega Apocalypse. Awesome
sounds, but his best work had to be Lightforce. Thirdly Jeroen Tel of Maniacs of Noise. He done
some brilliant tunes for games between 1988 and 1990. Finally Edwin Van Santen (Who sadly died
earlier this year). He created excellent tunes for 20CC, I remember a disk which had the entire
20CC music collection between 1988-1991, featuring many of his great tracks. I'm also going to be
working on a tribute demo to EVS, which will feature some of his best tunes, and messages on
scroll text from various people of the scene today.
The best scene musicians of today has to be Drax/MON, Intensity/Cosine, Fanta, Laxity and a few
others.
14
===== page 15 =====
What are the best sids ever in your opinion?
It is hard to say as there are loads of great SIDs stored on the HVSC. Except that, I wish people
would cut down on the hardcore drum 'n bass style music, as to my opinion Drum 'n Bass is not mu-
sic at all - even the real drum 'n bass, but rubbish in general. SID has not been built for awful (what
people call) music like this. Of course, it is up to the SID musician themselves :)
Finally, many thanks for the time you give for this interview, and now you can say any things
you want that the people will read from you!
I'd like to say a big thank you to all of those who are reading this article right now. I am impressed
that although the C64 is over 22 years old, people are still supporting this great machine, like
Cronosoft, the CSDB and many other C64 enthusiasts. This impresses me a lot as I still like using
the C64 today, and still keep many of the old C64 classics.
Secondly I'd like to say how awful it has been to hear about the death of EVS/20CC, in May 2006
and I would like to pay tribute to the great scene C64 musician. To show respect, I'll be releasing a
demo by the end of July 2006 featuring a selection of his great tunes, and a few messages from oth-
er sceners, as well as myself (Possibly with a note file included). EVS's great tunes will be remem-
bered.
15
===== page 16 =====
Tiny Sid 2 (part 1)
by Stefano Tognon <ice00@libero.it>
In this first part I go to show 4 entries of the Tiny Sid 2 Compo, two 256 bytes entries and two 512
bytes entries. The other 4 will be in next number.
The tunes can be downloaded at: http://digilander.iol.it/ice00/tsid/tinysid2/tinysid2
_
res.zip
XL5 (256b)
As I like the Jeff “X-Large” series of tunes, I have choose the number five for trying to cover it in
256 bytes. The tune is about 0:36 seconds long and is quite simple, but with a good result that you
will listen for some minutes for sure.
A look to the sid2mid output, gives and idea of the tune:
Voice 1 Voice 2 Voice 3
Time Note Freq PW WF ADSR VL Note Freq PW WF ADSR VL Note Freq PW WF ADSR VL Filter
============================================================================================================
00:00.04 >G#1< 51 2048 40 00c9 -- >C-1< 32 0 10 0069 -- >C-1< 32 0 10 0029 -- L__ 1__ 380 f
00:00.04 +++ 51 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 2d0 f
00:00.05 +++ 51 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 260 f
00:00.06 +++ 51 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 220 f
00:00.06 +++ 51 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 1f0 f
00:00.07 +++ 51 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 1c0 f
00:00.07 +++ 51 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 190 f
00:00.08 +++ 51 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 160 f
00:00.09 >G#2< 103 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 440 f
00:00.09 +++ 103 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 390 f
00:00.10 +++ 103 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 320 f
00:00.10 +++ 103 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 2e0 f
00:00.11 +++ 103 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 2b0 f
00:00.12 +++ 103 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 280 f
00:00.12 +++ 103 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 250 f
00:00.13 +++ 103 2048 40 00c9 -- +++ 32 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 220 f
00:00.13 >G#3< 207 2048 40 00c9 -- >G#1< 51 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 500 0
00:00.14 +++ 207 2048 40 00c9 -- +++ 51 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 450 0
00:00.15 +++ 207 2048 40 00c9 -- +++ 51 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 3e0 0
00:00.15 +++ 207 2048 40 00c9 -- +++ 51 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 3a0 0
00:00.16 +++ 207 2048 40 00c9 -- +++ 51 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 370 0
00:00.16 +++ 207 2048 40 00c9 -- +++ 51 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 340 0
00:00.17 +++ 207 2048 40 00c9 -- +++ 51 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 310 0
00:00.18 +++ 207 2048 40 00c9 -- +++ 51 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 2e0 0
00:00.18 >G#1< 51 2048 40 00c9 -- >G#2< 103 0 10 0069 -- +++ 32 0 10 0029 -- L__ 1__ 380 f
The voice 2 and 3 plays the same pattern of voice 1, but with a delay of 2 and 4 notes. Else, it
uses a low sustain, so there is a sort of low echo. All notes are equal length and instrument are fixed
(with the exception of filter)
Those points are very important: we have only one track to play for each voice, without the prob-
lem of different note length and changing of instruments. However, the implementation of the filter
(that change each tick and use different resonance) is out of question for a 256 byte tunes with all
the note we have to play.
I agree that the filter is what that give to this tune his style, but the result without it is acceptable
having in mind the 256 bytes restriction.
When I take all the notes for imagine how the track/pattern should be written, I had the first delu-
sion: the used notes are 17 and so I cannot use nibble to store them. However, I find that the tune
could be represented by pattern and this means a little of compression that can be achieved.
Here you can see the final pattern disposition (the first I wrote where a little different but more nat-
ural to use):
16
===== page 17 =====
[1] G#1 G#2 G#3
[1] G#1 G#2 G#3
[1] G#1 G#2 G#3
[1] G#1 G#2 G#3
[1] G#1 G#2 G#3
[2] F#1
[1] G#1 G#2 G#3
[1] G#1 G#2 G#3
[1] G#1 G#2 G#3
[1] G#1 G#2 G#3
[1] G#1 G#2 G#3
[2] F#1
[3] G#1 G#2 D#4 G#1 G#2 D-4 G#1 G#2 C-4 G#1
[4] G#2 A#3 G#1 G#3 G#2 A#3
[3] G#1 G#2 D#4 G#1 G#2 D-4 G#1 G#2 C-4 G#1
[5] D-4 D#4 G#2 C-4 G#2 G#3
++
[3] G#1 G#2 D#4 G#1 G#2 D-4 G#1 G#2 C-4 G#1
[4] G#2 A#3 G#1 G#3 G#2 A#3
[3] G#1 G#2 D#4 G#1 G#2 D-4 G#1 G#2 C-4 G#1
[6] D-4 D#4 G#2 G-4 G#2 G#4
[3] G#1 G#2 D#4 G#1 G#2 D-4 G#1 G#2 C-4 G#1
[4] G#2 A#3 G#1 G#3 G#2 A#3
[3] G#1 G#2 D#4 G#1 G#2 D-4 G#1 G#2 C-4 G#1
[5] D-4 D#4 G#2 C-4 G#2 G#3
++
[3] G#1 G#2 D#4 G#1 G#2 D-4 G#1 G#2 C-4 G#1
[4] G#2 A#3 G#1 G#3 G#2 A#3
[3] G#1 G#2 D#4 G#1 G#2 D-4 G#1 G#2 C-4 G#1
[7] D-4 D#4 G#2 G#4 G#2 A#4
++
[8] F-1 F-2 G#4 F-1 F-2 G-4 F-1 D#4 F-2 C-4 F-1 D#4 F-4 F-1 C-4 D#4 F#1 F#2 F#4 F#1 F#2 F-4
[9] F-1
[10]C#4 F#2 G#3
[9] F-1
[11]A#3 F#2 C#4 F#1 G#3
[8] F-1 F-2 G#4 F-1 F-2 G-4 F-1 D#4 F-2 C-4 F-1 D#4 F-4 F-1 C-4 D#4 F#1 F#2 F#4 F#1 F#2 F-4
[2] F#1
[10]C#4 F#2 G#3
[2] F#1
[11]A#3 F#2 C#4 F#1 G#3
I so wrote the tune that use the above information and I had it reach 309 bytes. No, we are too
much out.
However, adding the filter and have a 512 bytes entry should be a good solution, but I want to op-
timize ever better that better. I so change little things at each step and save the source of modifica-
tion (that you find inside the released tune) and finally I could stay into 256 bytes.
Here the optimization used:
➢ I switch the basic startup SYS instruction by loading directly the tune at address $0326. This
seems to be the better way to made the tune running: you save 8 bytes.
➢ I used to synchronize the tune with raster position (interrupt is disable with SEI), then cycle
to the 3 voices (I start from the first and then go to the others). Using instead the Frantic
technique that use SBX undocumented instruction (and cycles from the 3° voices to the 1°)
we can save another 4 bytes
➢ The pattern 1 was long 15 notes, but I divide into only 3 notes. I obtain to save of 4 bytes.
One more byte I save using INX for passing from 0 to 1 in an initialization
➢ 2 more bytes were saved using the LDY absolute,X instruction in two points
➢ The best reduction is of 15 bytes and come out from having the pattern/track pointer already
set to 0. I have carefully look at address page 0 to find 6 location that are in sequence of
0,
+7,
+7 and are empty. I have compare them after loading a program and at startup of the
machine. Is not granted that if you made other things with the C64 before loading the tune
they are still 0.
➢ Other 3 bytes were saved porting the initialization of control register of the voices inside the
17
===== page 18 =====
main cycle (and remembering that $11 is for voice 2 and 3 and $41 for voice 1)
➢ Sustain/Release of voice are: $C9, $69 and $29. You should note that release is never acti-
vated into the tune, so we can look only to attack. The idea is to use $30, $60 and $90, as
$60=$30<<1 and $90=$60<<1. So 1 bytes are saved using the shifts.
➢ Other 4 bytes are saved by using a zero page location that is already set to 1 at startup. We
can use even other little above value, as this is only the delay for starting to play the first
note. Fortunately one location was found.
➢ 8 bytes were saved by this idea: we have that one pattern of note is terminated by a 0 value.
In some cases, one pattern is followed only by another one, like: 4 and 3. If we remove the
end of pattern 4, pattern 3 will be executed as it come sequentially in the source.
➢ Last bytes were saved by removing a redundant instruction and by moving the init of volume
in the end of cycle
Well I hope that this description is useful for learning some optimization techniques
; XL5
; 256b SID music
; by Stefano Tognon 2006
; This is a cover of Jeff "X-Large 5"
; step1: 309 bytes (normal code of the tune)
; step2: 301 bytes (use $0326 start up)
; step3: 297 bytes (use Frantic sbx/sync thecnique)
; step4: 292 bytes (rearrange pat1 and inx)
; step4: 290 bytes (use ldy directly)
; step5: 277 bytes (use 0 value already setted)
; step6: 274 bytes (use common set of control register+fix Frantic code) +1 must be added
; step7: 273 bytes (use different SR initialization)
; step8: 269 bytes (use location defined to 1 at sturtup)
; step9: 265 bytes (rearrange patterns removing END)
; stepA: 263 bytes (new rearrange patterns removing END)
; stepB: 261 bytes (removing other 0 bytes)
; stepC: 259 bytes (remove redondant instruction)
; stepD: 256 bytes (move volume init, lost first note, but nop not needed)
processor 6502
curTrack = $57 curPat = $58 duration = $2B DUR = 8
; this and +7 and +14 are 0 at sturtup of after loading
; this and +7 and +14 are 0 at sturtup of after loading
; init at 1
;org 2049
;.byte $0b,$08,$e8,$03,$9e,"2061",0,0,0
;.org 2061
.org $0326
.byte $2A, $03
.byte $ED, $F6
.org $032A
sei ; disable interrupt
;lda #$0F
;sta $D418 ; volume max
;lda #$11 ; triangle waveform
;sta $D40B ; voice 2
;sta $D412 ; voice 3
;lda #$41
;sta $D404 ; voice 1
;lda #$C9
;sta $D406 ; SR voice 1
;lda #$69
;sta $D40d ; SR voice 2
;lda #$29
;sta $d414 ; SR voice 3
lda #$30
sta $d414 asl
sta $D40d asl
sta $D406 ; SR voice 3
; SR voice 2
; SR voice 1
lda #$08
sta $D403 ; wave high
;ldx #$00
;stx curTrack
18
===== page 19 =====
;stx curPat
;stx curTrack+7
;stx curPat+7
;stx curTrack+14
;stx curPat+14
; inx
;ldx #1
;stx duration
lda #1+DUR+DUR
sta duration+7
lda #1+DUR+DUR+DUR+DUR
sta duration+14
sync:
;lda #$ff
loopSync:
cpx $D012
bne loopSync
;ldx #0 ldx #14 loopVoice:
dec duration,x
bne nextVoice
lda #DUR
sta duration,x
; start from first voice
; start from last voice
lda #$11
cpx #0
bne skip
lda #$41
skip:
sta $D404,x ; control register of voice
nextNote:
; read next note
ldy curPat,x
lda pat,y
beq nextTrack
inc curPat,x ; go to next note
; put note y to sid x
tay
lda freqLo-1,y
sta $D400,x
lda freqHi-1,y
sta $D401,x
; increment voice index and test for last voice
nextVoice:
;txa
;clc
;adc #7
;tax
;cmp #$11
;bcs sync
;bcc loopVoice
; nop ; without this it goes out of sync!
lda #$ff
sta $D418
.byte $cb,7 ; sbx #7
bpl loopVoice
bmi sync
nextTrack:
inc curTrack,x
ldy curTrack,x
lda track,y
bmi resetTrack
sta curPat,x
bpl nextNote
resetTrack:
lda #0
sta curTrack,x
;sta curPat,x
beq nextNote
; note declaration
END = 0
F1 = 1
Fd1 = 2
Gd1 = 3
F2 = 4
Fd2 = 5
Gd2 = 6
Gd3 = 7
Ad3 = 8
C4 = 9
19
===== page 20 =====
Cd4 = 10
D4 = 11
Dd4 = 12
F4 = 13
Fd4 = 14
G4 = 15
Gd4 = 16
Ad4 = 17
freqLo:
.byte 231 ;F1
.byte 20 ;F#1
.byte 116 ;G#1
.byte 207 ;F2
.byte 39 ;F#2
.byte 232 ;G#2
.byte 208 ;G#3
.byte 129 ;A#3
.byte 103 ;C4
.byte 112 ;C#4
.byte 137 ;D4
.byte 178 ;D#4
.byte 59 ;F4
.byte 157 ;F#4
.byte 20 ;G4
.byte 160 ;G#4
.byte 3 ;A#4
freqHi:
.byte 2 ;F1
.byte 3 ;F#1
.byte 3 ;G#1
.byte 5 ;F2
.byte 6 ;F#2
.byte 6 ;G#2
.byte 13 ;G#3
.byte 15 ;A#3
.byte 17 ;C4
.byte 18 ;C#4
.byte 19 ;D4
.byte 20 ;D#4
.byte 23 ;F4
.byte 24 ;F#4
.byte 26 ;G4
.byte 27 ;G#4
.byte 31 ;A#4
pat:
pat1:
pat2:
pat4:
pat3:
pat5:
pat6:
pat7:
pat8:
pat9:
pat10:
pat11:
.byte Gd1, Gd2, Gd3, END
.byte Fd1, END
.byte Gd2, Ad3, Gd1, Gd3, Gd2, Ad3;, END
.byte Gd1, Gd2, Dd4, Gd1, Gd2, D4, Gd1, Gd2, C4, Gd1, END
.byte D4, Dd4, Gd2, C4, Gd2, Gd3, END
.byte D4, Dd4, Gd2, G4, Gd2, Gd4, END
.byte D4, Dd4, Gd2, Gd4, Gd2, Ad4;, END
.byte F1, F2, Gd4, F1, F2, G4, F1, Dd4, F2, C4, F1, Dd4, F4, F1, C4, Dd4, Fd1, Fd2, Fd4, Fd1, Fd2, F4, END
.byte F1, END
.byte Cd4, Fd2, Gd3, END
.byte Ad3, Fd2, Cd4, Fd1, Gd3; , END next byte is already 0
track:
.byte (pat1-pat), (pat1-pat), (pat1-pat), (pat1-pat)
.byte (pat1-pat), (pat2-pat), (pat1-pat), (pat1-pat)
.byte (pat1-pat), (pat2-pat), (pat1-pat), (pat2-pat)
.byte (pat3-pat), (pat4-pat); (pat3-pat)
.byte (pat5-pat)
.byte (pat3-pat), (pat4-pat);, (pat3-pat)
.byte (pat6-pat)
.byte (pat3-pat), (pat4-pat);, (pat3-pat)
.byte (pat5-pat), (pat3-pat)
.byte (pat4-pat);, (pat3-pat)
.byte (pat7-pat)
;.byte (pat8-pat),
.byte (pat9-pat), (pat10-pat), (pat9-pat)
.byte (pat11-pat), (pat8-pat), (pat2-pat), (pat10-pat)
.byte (pat2-pat), (pat11-pat), $FF
20
===== page 21 =====
Mini Digi Drum (512b)
1/4 note 1/8 note 1/16 note
128 64 32 16
7 6 5 4
Speed Instrument Duration Bits Byte Sound Bar
12 0 1 12 0 1 15 1 1 12 0 1 12 0 1 dc.b 76 dc.b 76 dc.b 207 dc.b 44 dc.b 44 bump 1
bump
pshhh
da
da
12 0 1 12 0 1 15 1 1 12 0 1 dc.b 76 dc.b 76 dc.b 207 dc.b 76 bump 2
bump
pshhh
bump
1 0 1 1 0 1 1 0 1 1 0 1 5 0 1 5 0 1 5 0 1 5 0 1 dc.b 33 dc.b 33 dc.b 33 dc.b 33 dc.b 37 dc.b 37 dc.b 37 dc.b 37 bip 3
bip
bip
bip
bup
bup
bup
bup
12 0 1 12 0 1 15 1 1 12 0 1 12 0 1 dc.b 76 dc.b 76 dc.b 207 dc.b 44 dc.b 44 bump 4
bump
pshhh
da
da
12 0 1 12 0 1 15 1 1 12 0 1 dc.b 76 dc.b 76 dc.b 207 dc.b 76 bump 5
bump
pshhh
bump
15 1 1 12 0 1 15 1 1 12 0 1 dc.b 207 dc.b 76 dc.b 207 dc.b 76 pshhh 6
bump
pshhh
bump
15 1 1 4 1 1 12 0 1 4 1 1 15 1 1 4 1 1 12 0 1 4 1 1 dc.b 175 dc.b 164 dc.b 44 dc.b 164 dc.b 175 dc.b 164 dc.b 44 dc.b 164 pshhh 7
tsk
bump
tsk
pshhh
tsk
bump
tsk
15 1 1 12 0 1 15 1 1 12 0 1 dc.b 207 dc.b 76 dc.b 207 dc.b 76 pshhh 8
bump
pshhh
bump
15 1 1 4 1 1 12 0 1 4 1 1 15 1 1 4 1 1 12 0 1 4 1 1 dc.b 175 dc.b 164 dc.b 44 dc.b 164 dc.b 175 dc.b 164 dc.b 44 dc.b 164 pshhh 9
tsk
bump
tsk
pshhh
tsk
bump
tsk
15 1 1 12 0 1 15 1 1 12 0 1 dc.b 207 dc.b 76 dc.b 207 dc.b 76 pshhh 10
bump
pshhh
bump
15 1 1 4 1 1 12 0 1 4 1 1 15 1 1 4 1 1 12 0 1 4 1 1 dc.b 175 dc.b 164 dc.b 44 dc.b 164 dc.b 175 dc.b 164 dc.b 44 dc.b 164 pshhh 11
tsk
bump
tsk
pshhh
tsk
bump
tsk
15 1 1 12 0 1 15 1 1 12 0 1 dc.b 207 dc.b 76 dc.b 207 dc.b 76 pshhh 12
bump
pshhh
bump
15 1 1 4 1 1 12 0 1 4 1 1 15 1 1 4 1 1 12 0 1 4 1 1 dc.b 175 dc.b 164 dc.b 44 dc.b 164 dc.b 175 dc.b 164 dc.b 44 dc.b 164 pshhh 13
tsk
bump
tsk
pshhh
tsk
bump
tsk
12 0 1 12 0 1 15 1 1 12 0 1 12 0 1 dc.b 76 dc.b 76 dc.b 207 dc.b 44 dc.b 44 bump 14
bump
pshhh
da
da
12 0 1 12 0 1 15 1 1 12 0 1 dc.b 76 dc.b 76 dc.b 207 dc.b 76 bump 15
bump
pshhh
bump
15 1 1 12 0 1 15 1 1 12 0 1 dc.b 207 dc.b 76 dc.b 207 dc.b 76 pshhh 16
bump
pshhh
bump
15 1 1 4 1 1 12 0 1 4 1 1 15 1 1 4 1 1 12 0 1 4 1 1 dc.b 175 dc.b 164 dc.b 44 dc.b 164 dc.b 175 dc.b 164 dc.b 44 dc.b 164 pshhh 17
tsk
bump
tsk
pshhh
tsk
bump
tsk
4 1 1 4 1 1 4 1 4 1 4 1 4 1 12 0 1 12 0 1 dc.b 196 dc.b 196 1 dc.b 148 1 dc.b 148 1 dc.b 148 1 dc.b 148 dc.b 44 dc.b 44 tsk 18
tsk
tsk
tsk
tsk
tsk
bump
bump
15 1 1 15 1 1 15 1 1 15 1 1 15 1 1 15 1 1 dc.b 207 dc.b 207 dc.b 175 dc.b 175 dc.b 175 dc.b 175 pshhh 19
pshhh
pshhh
pshhh
pshhh
pshhh
21
This is the 512 bytes tune by Leif
Bloomquist, and it is quite unusual. In fact it is
a sampled based tune!
The tune use two instruments: kick and hi-
hat.
In this left table you will see the structure of
the tune: the instrument to play, the duration
of the note and the speed for each bars.
Changing the speed of the sample, will pro-
duce different sound that is wrote in the
Sound column.
All those information are packet into a byte:
● Bit 7: instrument to use
● Bits 6,5,4: duration
● Bits 3,2,1,0: speed
The samples for the two instruments are
packed together in one byte:
● high nibble: kick
● low nibble: hihat
and they eat half of the space available for
the tune.
The last words to say about this player is
that synchronization is made by counting the
current clock time in jiffies.
===== page 22 =====
; "Midi Digi Drums" By Leif Bloomquist Schema/AIC leif@schemafactor.com
;
; Entry into the TinySID2 Compo 2006 (512 byte category)
;
"Vars"
;$fb - placeholder for jiffy value when done current note
;$fc - holds location in track
;$fd - always 0
;$fe - points to current instrument - either $09 (kick) or $0a (hihat)
; ================ Setup ======================
processor 6502 ;C64
org $0801,00
basic ;
"SYS2061"
dc.b $0c,$08,$64,$0c,$9e,$32,$30
dc.b $36,$31,$00,$00,$00
;Digibooster
lda #$ff
sta $d406
sta $d406+7
sta $d406+14
lda #$49
sta $d404
sta $d404+7
sta $d404+14
;Fill in the "513th" byte - last byte of sample
lda #$14
sta $09FF
ldx #$00
stx $fd
stx $d011 ; Blank Screen
; Split up track
split
lda sampleraw,x lsr sampleraw,x lsr sampleraw,x lsr sampleraw,x lsr sampleraw,x and #$0f inx
bne split
; Grab the packed byte
; Rotate original right
; Rotate original right
; Rotate original right
; Rotate original right - high nybble should be low nybble now
; Strip high bits out of accumulator
sta sampleraw+$100,x ; Stash original low nybble in next page
; Reset to beginning of track
init
ldx #$00
stx $fc
; ================ Main Loop ======================
repeat
ldy $fc
lda track,y
; 1. Get a byte from the "track" and determine note duration.
beq init ; A zero byte means the end of the track. Start from the beginning again.
; Mask out the duration bits
and #$70
;Divide by 4 and store in the code ahead
lsr
lsr
sta duration+1
; Figure out the jiffy count at the end of this duration
lda $a2 ; Current clock time in jiffies
duration
sta $fb
adc #16 ; Ignore carry, we want wraparound (#16 gets overwritten)
; 2. Select instrument.
; Default to kick
lda #>sampleraw
sta $fe
; Get the original byte again
lda track,y
; Determine instrument through hi bit
22
===== page 23 =====
go
bmi go inc $fe ; Leave as kick
; Move to next page where hi-hat samples are
; 3. Determine the playback speed
and #$0f ; Mask the playback speed
sta samplespeed+1 ; Store in code ahead
; 4. Play the sample.
jsr play
; 5. Wait for timeout to expire
ldy $fb
rest
cpy $a2
bne rest
; Move to next entry in track
inc $fc
jmp repeat
; ================ Play Instrument ======================
play
ldy #$00
playloop
lda ($fd),y
sta $d418
sta $d020
samplespeed
ldx #12 playdelay
dex
bne playdelay
;some delay value - set from track (#12 gets overwritten)
iny
bne playloop kickdone
rts
; Samples are exactly 255 bytes long, so we're done when back to 0
; ================ Track ======================
track ;(Copy/paste from Excel)
dc.b 76
dc.b 76
dc.b 207
dc.b 44
dc.b 44
dc.b 76
dc.b 76
dc.b 207
dc.b 76
dc.b 33
dc.b 33
dc.b 33
dc.b 33
dc.b 37
dc.b 37
dc.b 37
dc.b 37
dc.b 76
dc.b 76
dc.b 207
dc.b 44
dc.b 44
dc.b 76
dc.b 76
dc.b 207
dc.b 76
dc.b 207
dc.b 76
dc.b 207
dc.b 76
dc.b 175
dc.b 164
dc.b 44
dc.b 164
dc.b 175
dc.b 164
dc.b 44
dc.b 164
dc.b 207
dc.b 76
dc.b 207
dc.b 76
dc.b 175
dc.b 164
dc.b 44
dc.b 164
23
===== page 24 =====
dc.b 175
dc.b 164
dc.b 44
dc.b 164
dc.b 207
dc.b 76
dc.b 207
dc.b 76
dc.b 175
dc.b 164
dc.b 44
dc.b 164
dc.b 175
dc.b 164
dc.b 44
dc.b 164
dc.b 207
dc.b 76
dc.b 207
dc.b 76
dc.b 175
dc.b 164
dc.b 44
dc.b 164
dc.b 175
dc.b 164
dc.b 44
dc.b 164
dc.b 76
dc.b 76
dc.b 207
dc.b 44
dc.b 44
dc.b 76
dc.b 76
dc.b 207
dc.b 76
dc.b 207
dc.b 76
dc.b 207
dc.b 76
dc.b 175
dc.b 164
dc.b 44
dc.b 164
dc.b 175
dc.b 164
dc.b 44
dc.b 164
dc.b 196
dc.b 196
dc.b 148
dc.b 148
dc.b 148
dc.b 148
dc.b 44
dc.b 44
dc.b 207
dc.b 207
dc.b 175
dc.b 175
dc.b 175
dc.b 175
dc.b 0 ; EOT
; ================ Packed Sample ======================
org $0900
sampleraw
incbin "samples-packed.bin"
24
===== page 25 =====
RM (512b)
RM is my cover of David Whittaker “Red Max”
. This is a repetitive tune with lot of interesting
sounds, so maybe it is a little difficult to fit in so little space: a good challenge for cover.
At the beginning, the voice 3 performs an arpeggio with 4 notes that it is combined with a long
pulse modulation that give the characteristic fabulous sound. Voice 1 uses 3 instruments for having
a drum sound using an octave arpeggio. Inside the tune there is long sound with decreasing porta-
mento.
In totally it is used 8 different instruments that are based onto ADSR and waveform, so we have
to manage for sure the AD, SR and control register for one instrument. We start using a fixed pulse
modulation for all the voice.
Looking at the notes, we see that there are lot of different notes used, so it is not convenient to
store only the high/low frequency of the used notes, as for sure generating them at runtime will take
less space.
Actually the best runtime generation that I see so far is with about 52 bytes, based onto a 11
bytes of frequency table. But when I reverse engineering Modulus tune, I remember that Ivan gener-
ates always his notes frequency based onto 22 values, and he not store the values into a table.
For madding this, he stored the note as: octave/base note
then he takes the frequency from the table (that are for the highest octave) and down sample until
the octave he has to produce. I so take the same idea, but using 11 bytes for the lower octave fre-
quency (that has the high byte always at 1) and over sampling until the octave to produce. The re-
sult is about a 41 bytes procedure (I use an ASR undocumented instruction for a AND/SHR instruc-
tions inside it).
Looking now at the structure of the song, we see that voice 1 plays a fixed pattern of 10 notes/du-
rations/instruments and voice 2 plays 16 notes, but each pattern has a transposition from the prece-
dent.
Voice 3 has different patterns, but with some exception of equal pattern and pattern transposition.
This make voice 3 a little more complicated.
Now the idea is simple: memorize the transposition that are to used for each track when the pat-
tern of note is over. For voice 3, we even memorize the transposition of the pattern, and even the
pattern to use for that transposition.
All is now good, but not the duration/base note used for memorize one note, as making a transpo-
sition is not like to add the transpose value to the actual note to play. This is a right reason for abort
the used note representation and to go to generate the frequency with the more expansive 52 bytes
routines, as for sure the transpose is just an ADC instead of a lot of base-12 addictions.
No. No so fast. Maybe the transpositions are apply to certain notes only and some accurate anal-
ysis could help us to use this system with the ADC for transposition at the same time.
Well, to make it short, it is possible and using only 2 additional bytes, and this is lot better than the
standard method. However I think it is a little complicated to show how this come from (I have made
lot of rearranging for reach this point), but I hope that you understand the idea below this by looking
at this final description.
25
===== page 26 =====
1. C C
2. C# D
3. D E
4. D# D# Voice 3: E F C A# G G# 0,
5. E F
6. F G Voice 2: A# 0,
7. F# A
8. G G# Voice 1: F C G G# 0,
9. A C#
10. A# F#
11. B D
12. G
13. A
14. .
15. .
+12
+2,
+4,
+7
+2,
+4,
+9
In the right part above we see all the notes (without octave) used by the 3 voices and the relative
arpeggio values to add to each note. In the left part we see the common notes sequence from 0 to
11. Notes from 12 are like the others, but 1 octave above. Here you see even the base notes stored
into the nibble: you see that the order is particular. Else notes at position 12 and 13 are the same of
position 6 and 7. Remember that note is stored as octave/base note.
Now, take those values for arpeggio:
Voice 1: 0,
Voice 2: 0,
Voice 3: 0,
+1,
+8,
+16
+12,
+9,
+13
+12
You will see that a simple addiction will perform the right note steps.
At this point we can describe how a pattern is made:
Byte Description
1000.xxxx xxxx= instrument to use
11nn.nnnn nnnnnn= note length
0ccc.nnnn ccc=octave nnnn=base note
Positive values are notes in the form octave/base note, while negative values are instruments or
note length according with bit 6 value.
Now that all seems perfectly, we have to make the arpeggio in voice 3 at the beginning (first 4
bars of song) and in the middle of song, as this is a sound that must be replicated. Well, this maybe
late me think to rewrite all the stuff not using octave/base note, as now it is not possible to make the
arpeggio by simple adding a value.
However, now we have save lot of bytes and maybe we could find a way to make the arpeggio in
another way. In fact, the arpeggio must be generated at each frame.
I so try to make arpeggio by adding a total amount of frequency to add at the last note played on
the voice 3 when it plays note of octave 6 (that fortunately is when arpeggio is used). Arpeggio
steps are taken from a table.
Unfortunately this code goes out of space, so I remove some value of arpeggio steps and even I
26
===== page 27 =====
manage only high frequency for reducing space. The final arpeggio (that it is different from the origi-
nal) come out after several try into changing the values until I find a sound that was acceptable.
; RM
; 512b SID music
; by Stefano Tognon 2006
; This is a cover of David W.
"Red Max"
; For having all the high/low frequency of notes, I use the inverse
; thecnique of Ivan Del Duca (he starts from octave 8 and go down,
; I start from octave 0 and goes up, saving 12 bytes for high values
; as they are to take all = 1)
; The base routine take about 41 bytes, some overhead is due to the use
; of transpositions of note
; note sequence are choosed for having a transpose simple by adding the value
; to octave/note. Fortunately only 2 bytes are needed for this
; optimization:
; -3 make pat4 followed by pat5
; -2 move trPos before trans1, removing two 0
; -1 use ASR for AND and LSR
; -6 use ADSR in same byte
; -1 skip a zero byte as next is 0
; -2 use bpl instead of jmp
; -5 do not init an already 0 value
; -2 put right value inside trPos and trPos+1
; all the other just removing arpeggio routine and modify until a good sound is reached
; devel note: AD SR could be combined as low/high nibble are in right position
; later fix: remove a typing error that I see while I was trying to create the psid version
.processor 6502
actFreqLow = $10
actFreqHigh = $11
curTrack = $57 curPat = $58 temp2 = $2A
delay = $2B ; init at 1
temp = $2C
duration = $2D ; actual duration +0 +7 +14
storeDur = $2E ; stored duration to use +0 +7 +14
instr = $2F ; instrument number to use +0 +7 +14
transp = $30 ; transposition for this pattern +0 +7 +14
note = $31 ; note +0 +7 +14
DUR = 12;48
.org $0326
.byte $2A, $03
.byte $ED, $F6
.org $032A
sei ; disable interrupt
lda #(trans1-trans) sta curTrack
; lda #$8
;sta $d402
sta $d403
;sta $d402+7
sta $d403+7
;sta $d402+14
sta $d403+14
; read initial track position
lda #(trans2-trans) sta curTrack+7
; read initial track position
; value to set is 0
; read initial track position
;lda trPos+3 ;sta curTrack+14
lda #1
sta duration
sta duration+7
sta duration+14
sync:
;lda #$ff
loopSync:
cpx $D012
bne loopSync
ldx #14
; this and +7 and +14 are 0 at sturtup of after loading current track/transposition position
; this and +7 and +14 are 0 at sturtup of after loading
27
===== page 28 =====
lda note,x pha
; read voice 3 note
jsr noteFreq
ldy delay
;lda delay
;lsr
;and #$03
;tay
lda arpeggio,y
tay
contArpeggio:
;cpy #0
beq skipArpeggio
lda actFreqLow
adc #$24
sta actFreqLow
lda actFreqHigh
adc #$06 ; 04
sta actFreqHigh
dey
bpl contArpeggio
skipArpeggio:
pla
cmp #$60
bmi ggg
jsr putFreq
ggg:
ldy #DUR
dec delay
bne sync
sty delay
;ldx #14 loopVoice:
dec duration,x
bne nextVoice
lda storeDur,x
sta duration,x
; start from last voice
nextNote:
lda #$40
sta $D404,x
; read next command
ldy curPat,x
lda pat,y
beq nextTrack
bpl putNote
cmp #$C0 ; len/instr?
bcc newInst
and #$3F ; store duration
sta storeDur,x
sta duration,x
inc curPat,x
bne nextNote
newInst:
and #$0F
sta instr,x inc curPat,x
bne nextNote
; store instrument to use
putNote:
sta note,x jsr noteFreq
; used for voice 3 arpeggio
jsr putFreq
ldy instr,x ; instrument number
lda control,y
sta $D404,x ; put control sid value
lda adsr,y
pha
and #$0F
sta $D405,x ; put attack/decay
pla
and #$f0
sta $D406,x ; put sustain/release
inc curPat,x
28
===== page 29 =====
; increment voice index and test for last voice
nextVoice:
lda #$7f
sta $D418
;jmp sync
.byte $cb,7 ; sbx #7
bpl loopVoice
;bmi sync
jmp sync
nextTrack:
ldy curTrack,x
lda trans,y
sta transp,x bmi resetTrack
cpx #7
beq tr2
bpl tr3
lda #(pat0-pat) setPat:
sta curPat,x inc curTrack,x
bpl nextNote
resetTrack:
txa
lsr
lsr
tay
lda trPos,y sta curTrack,x
;sta curPat,x
bpl nextNote
tr2: lda #(pat1-pat)
bpl setPat
tr3: lda track,y
bne setPat
noteFreq:
stx temp
; set frequency A=octave/note
clc
adc transp,x
pha
and #$0F
tax
lda freqSource,x
sta actFreqLow
lda #1
sta actFreqHigh
pla
;and #$F0 ;lsr
.byte $4b, $F0 lsr
lsr
lsr
tay
beq freqOk
calcFreq:
asl actFreqLow
rol actFreqHigh
dey
bne calcFreq
freqOk:
ldx temp
rts
putFreq:
lda actFreqLow
sta $D400,x
lda actFreqHigh
sta $D401,x
rts
fC0 = 12
fC_0 = 28
; store transposition for this pattern
; fix pattern for voice 1
; store current pattern
; read initial track position
; fix pattern for voice 2
; voice 3 use different patterns
; take the octave of note
; ASR #$0F
29
===== page 30 =====
fD0 = 45
fD_0 = 62
fE0 = 81
fF0 = 102
fF_0 = 123
fG0 = 145
fG_0 = 169
fA0 = 195
fA_0 = 221
fB0 = 250
; source values used to calculate frequency table. (taken from 4mat)
freqSource .byte fC0, fD0, fE0, fD_0, fF0, fG0, fA0, fG_0, fA_0, fC_0, fF_0, fB0, fG0, fA0
; instrument
control: ; $41
.byte $41, $11, $81, $41, $21, $41, $41, $41, $41
; SR/AD
adsr:
.byte $4A, $09, $09, $08, $30, $48, $49, $6a, $8C
; voice 1:
I1 = 0 ; 41 0a40
I2 = 1 ; 11 0900
I3 = 2 ; 81 0900
; voice 2
I4 = 3 ; 41 0800
; voice 3
I5 = 4 ; 41 0030
I6 = 5 ; 41 0840
I7 = 6 ; 41 0940
I8 = 7 ; 41 0a60
I9 = 8 ; 41 0c80
; length:
; 1 2 3 4 6 7 16 17
L1 = 1
L2 = 2
L3 = 3
L4 = 4
L6 = 6
L7 = 7
L16 = 16
L17 = 17
L32 = 32
INS = $80 ; instrument
LEN = $C0 ; length
C0 = 0
D0 = 1
E0 = 2
D
0 = 3
_
F0 = 4
G0 = 5
A0 = 6
G
0 = 7
_
A
0 = 8
_
C
0 = 9
_
F
0 = 10
_
B0 = 11
; notes:
C1 = $10 + C0 ; 12
F1 = $10 + F0 ; 17
A
_
1 = $10 + A_0 ; 22
C2 = $20 + C0 ; 24
A
_
2 = $20 + A_0 ; 34
C3 = $30 + C0 ; 36
D3 = $30 + D0 ; 38
E3 = $30 + E0 ; 40
F3 = $30 + F0 ; 41
G3 = $30 + G0 ; 43
G
_
3 = $30 + G_0 ; 44
A3 = $30 + A0 ; 45
A
_
3 = $30 + A_0 ; 46
C4 = $40 + C0 ; 48
D4 = $40 + D0 ; 50
E4 = $40 + E0 ; 52
G4 = $40 + G0 ; 55
C5 = $50 + C0 ; 60
G5 = $50 + G0 ; 67
G
_
5 = $50 + G_0 ; 68
C6 = $60 + C0 ; 72
D6 = $60 + D0 ; 74
30
===== page 31 =====
F6 = $60 + F0 ; 77
pat:
arpeggio:
.byte 3, 7;, 12
.byte $00 ; this must be zero as all pattern start from here (already init at 0)
pat0:
.byte INS+I1, LEN+L2, F1
.byte INS+I2, LEN+L1, G
5, C6
_
.byte INS+I3, LEN+L2, C6
.byte INS+I1, C2, F1
.byte INS+I2, LEN+L1, G
5, G5
_
.byte INS+I3, LEN+L2, C5
.byte INS+I1, C1, 0
pat1:
.byte INS+I4, LEN+L1
.byte A_1, A
_
1, A_1, A
_
2, A_1, A_1, A
_
.byte A_1, A
_
1, A_1, A
_
2, A_1, A_1, A
_
1, A_1
2, A_2, 0
pat2:
.byte INS+I5, LEN+L16, F6, F6, C6, C6, F6, F6, C6, C6, 0
pat3:
.byte INS+I6, LEN+L1, E3, F3
.byte INS+I7, LEN+L2, F3, F3, F3
.byte LEN+L6, F3
.byte LEN+L2, C3
.byte INS+I6, LEN+L1, F3, C4
.byte INS+I7, LEN+L2, C4
.byte LEN+L3, C4
.byte LEN+L1, C3
.byte INS+I6, C4, C4
.byte INS+I7, LEN+L2, C4, A
3
_
.byte LEN+L1, G
3
_
.byte INS+I8, LEN+L17,G3
.byte INS+I6, LEN+L1, G3, E3, C3, C4, G3, E3, E4, C4, G3, G4, E4, C4, C5, G4, E4, C4, 0
pat4:
.byte INS+I9, LEN+L16, D3, A_2;, 0
pat5:
.byte INS+I6, LEN+L1, A3, A3
.byte INS+I7, LEN+L2, G3, A3
.byte LEN+L1, G3
.byte LEN+L2, A3
.byte LEN+L3, D4
.byte LEN+L4, D6
.byte INS+I6, LEN+L1, F3, F3
.byte INS+I7, LEN+L2, E3, F3
.byte LEN+L1, E3
.byte LEN+L3, F3
.byte LEN+L2, A3, F3, E3, 0
pat6:
;, 0 .byte INS+I9, LEN+L32, E3
next byte is 0, so skip
; track1 transpose
trans:
; track3 transpose
trans3:
.byte 0, $10, 0, 0, 0
;0,
.byte 0, $FF
trPos:
.byte (trans1-trans), (trans2-trans)
;, 0, 0 taken below
; track1 transpose
trans1:
.byte 0, 0, 12, 12, 0, 0, 12, 12
.byte 12, 12, 1, 1, 12, 12, 1, 1
.byte 13, 13, 0, 0, 13, 13, 12, 12
.byte $FF
; track2 transpose
trans2:
.byte 12, 12, 8, 8, 12, 12, 8, 8
.byte 12, 12, 8, 8, 12, 12, 8, 8
.byte 9, 9, 0, 0, 9, 9, 8, 8
.byte $FF
track:
.byte (pat2-pat), (pat3-pat), (pat3-pat), (pat5-pat), (pat4-pat)
;, (pat5-pat),
.byte (pat6-pat)
31
===== page 32 =====
Miniature Rhapsody (256b)
This is the 256 bytes entries by Eric Odland, and how you can hear it is almost based onto arpeg-
gio.
Let we see some points about the player:
● It initializes IRQ by starting at $0326 (the short way around), disabling interrupt with SEI, and
then synchronized to raster position using 70 tests passages.
● Frequency table is built at startup using 12 values, then fixed values of Sustain/Release are
set for voice 1 and 2, and Attack/Decay for voice 3.
● A low pass filter is apply in all the 3 voices.
● Voice 2 will play the notes of voice 1 but with a frame of delay.
● Voice 1 plays the arpeggio, while voice 3 the bass instrument.
It is very interesting to see how data to play is represented into this player:
Value (xxyyzzzz) Description
xx Arpeggio (0-3) that are taken from a table:
1 4 3 - maj7 3rd inversion
2 3 2 - 4 root
3 4 3 - min7 root
min2 root
yy Bass (0-3) that are taken from a table that contains
the 3 values of note to play in each case
zzzz Root note (1-F) for the arpeggio
The tune is so composed only from a sequence of bytes in the above form (each notes have the
same duration), that are repeated as soon as the 00 value is reached.
Here the source:
; Miniature Rhapsody
; By Eric Odland
;
;
;
;
; Written for "Tiny SID 2" - Online SID Music-Programming Competition
; Platform:Commodore 64
; (c) 2006 Stone Monkey Software.
; File: mr.S
; Created on 3/22/06.
; NOTES:
; Idea to use undocumented opcode SAX borrowed from Stefano Tognon's xl5
processor 6502
; ************ VARIABLES ***********
SEG.U Data
; Zero Page variables.
org $E0
voice1Note dc.b 0
org $E7
voice2Note dc.b 0
org $EE
voice3Note dc.b 0
32
===== page 33 =====
org $FA
pattern1Position dc.b 0
arpeggioIndex dc.b 0
arpeggioVector dc.b 0
; arpeggio data
org $C000
arpeggioData ds.b 17,0
SEG header
org $0326
dc.b $2A, $03
dc.b $ED, $F6
; **************** CODE *****************
SEG code
org $032A
sei
SetupFrequencyTable:
; Setup frequency table
ldy #11
lda #1
.1 sta freq_hi,y
dey
bpl .1
ldy #0
.2 lda freq_lo,y
asl
sta freq_lo+12,y
lda freq_hi,y
rol
sta freq_hi+12,y
iny
bne .2
tya
Start:
sta pattern1Position
; Initialize SID
lda #$9D ; voice1 sustain/release
sta $D406
lsr ; voice2 sustain/release
sta $D40D
sta $D413 ; voice3 attack/decay
; Init player
bne GetNextArpeggio
UpdateFrame subroutine
; update voice 1
ldx arpeggioIndex
lda arpeggioData,x
sta voice1Note
ldx #14
.nextVoice:
lda voice1Note,x
tay
lda freq_lo,y
sta $D400,x
lda freq_hi,y
sta $D401,x
lda #$21
sta $D404,x
lda #$FF
dc.b $CB, 7 bpl .nextVoice
; SAX #7 : X = (A & X)-7
lda arpeggioVector ldx arpeggioIndex
cpx #16
bne .1
eor #%11111110
sta arpeggioVector
.1 clc
; arpeggioVector = -arpeggioVector
adc arpeggioIndex
sta arpeggioIndex
bne endGetNextArpeggio
inc pattern1Position
;--------- GETNEXTARPEGGIO ---------
GetNextArpeggio subroutine
ldy pattern1Position
33
===== page 34 =====
lda PatternTable,y
beq Start
; fill arpeggioData
pha
and #$0F
clc
adc #31
sta arpeggioData pla
; store first note in arpeggio
lsr
lsr
lsr
lsr
pha ; arpeggio/bass note values
and #%00000011 ;extract bass note
tay
lda BassNote,y
sta voice3Note
pla ; extract arpeggio value
lsr
lsr
tay
lda ArpeggioTable,y
ldy #0
ldx #3
.1 rol
rol
pha
rol
and #%00000011
sec
adc arpeggioData,y
iny
sta arpeggioData,y
; setup filter
; Put this here instead of at initizalization to save space
lda SIDsnapshot-1,x
sta $D416-1,x
pla
dex
bne .1
.2 iny
lda arpeggioData-4,y
clc
adc #12
sta arpeggioData,y
cpy #16
bne .2
stx arpeggioIndex
; arpeggioVector = 1
inx
stx arpeggioVector
lda #$20
sta $D412
;------- END GETNEXTARPEGGIO ----------
endGetNextArpeggio:
ldx #70
.wait lda $D012
bne .wait
dex
bne .wait
lda voice1Note
sta voice2Note
jmp UpdateFrame
;**************** DATA ****************
;
; Pattern data = xxyyzzzz
; xx = arpeggio 0-3
; yy = bass 0-3
; zzzz = rootNote 1-F
; $00 = end of pattern
PatternTable:
34
===== page 35 =====
;dc.b $75, $75, 0
dc.b $95, $86, $E7, $8B ; part 1
dc.b $D5, $22, $BC, $BB ; part 2
dc.b $29, $89, $29, $8E ; part 3
dc.b $1E, $9C, $73, $73 ; part 4
dc.b $00
ArpeggioTable:
dc.b %00111000 ; 1 4 3 - maj7 3rd inversion
dc.b %01100100 ; 2 3 2 - 4 root
dc.b %10111000 ; 3 4 3 - min7 root
dc.b %11010000 ; min2 root
BassNote:
dc.b 21, 24, 26, 0
SIDsnapshot:
freq_hi = . + $100
freq_lo:
freqMult dc.b $70, $8F, $1F ; filter cutoff_hi, res/filt, mode/vol
dc.b $0C, $1C, $2D, $3F, $52, $66, $7B, $92, $AA, $C3, $DE, $FA
Conclusion
For seeing the winners of the compo analyzed, you must attend the next number, but I think that
with these entries you have already lot of material available for starting to think how to build your
mini tunes, maybe for next compo or minigame.
35
===== page 36 =====
Goattracker 2
by Stefano Tognon <ice00@libero.it>
In this article I will go to look at the new version of Cadaver's Goattracker pc music editor. You al-
ready know that version 2 is now tables based and so it can compete with the best C64 music tools
available.
I think that the best mode to discuss this tool is by creating a tune with it. I so realize a cover of
“What is Love” of Haddaway. This is a tune that I like very much.
Start
The first step was to setting up the Goattracker in Linux (as no executable for it is released): it is
just to run make into the src directory, but after having already compiled the BCE library and put it in
the linux directory, otherwise the compilation did not finish.
Now I went to search some midi version of the tune to cover: I download many versions, that dif-
fers only for the kind of midi-mapped instrument used. But finally I simple choose the Karaoke one,
as I like to see the words of the song in it.
For extracting the notes from the midi, I had used the last version of Rosegarden. You find it at
www.sf.net and I have compiled it directly from the source using the last tarball available.
Rosegarden is a midi/notation sequencer that features:
● Matrix editor
● Matrix drum editor
● Notation editor
● Midi event editor
For my porpoise I had used the matrix/drum editors using a 1/16 rhythm scale.
If you want to test the midi instrument used into the midi tune, for then going to reproduce that
sound into the sid chip, the best way is to have timidity running as alsa server in Rosegarden:
timidity -Os -iA -B2,8 -s 16k
The -s option it to limit the cpu usage as in low system it can be too hard.
I also use to disable in runtime some voices of the midi track while running for better going into
choose the 3 tracks that I can fit into the sid.
There are 10 music tracks into the midi (as you can see from the main Rosegarden page), but es-
sentially I have choose to have:
● Drum
● Bass
● Melody / Lead/ Chords
for the 3 sid tracks.
36
===== page 37 =====
Rosegarden main page
Drum matrix editor
37
===== page 38 =====
Matrix editor
Notation editor
For my own use I make a table (build while creating the tune) in a spreadsheet that contains the
Goattracker patterns with the used instruments for each voice. I put even some colors that give bet-
ter looks to the global music (one color for one type of instrument).
This was for me the best way to remember each patterns into goattracker and even to control that
filter may stay always into voice 3
You will see that the first patter is very short: this is a good characteristic of GT that allow you to
set the length of each pattern independently.
In the tune, I use the filter in voice 3 for making the bass sound better. In some points I have to re-
produce the lead or synth voice instead of bass: in that case I have to make a special instrument
that stop filter execution as GT2 left the filter set up from the previous instrument (the bass).
I have also use a lot the pattern function Rxx that repeat xx time the pattern that follows. GT2 per-
mit this instruction and even to transpose a pattern.
38
===== page 39 =====
Bat.
V1 V2 V3 V1
V2
V3
V1
V2
V3
V1
V2
V3
V1
V2
V3
V1 V2 V3
V1
V2
V3
V1
V2
V3
V1
V2
V3
V1 V2 V3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
08 00 01 Cabana
02 Cabana+Tamburine+Hihat 03 C.+T.+H.+Snare+Drum
08
07 Lead 07 Lead 0D Chords 0D Chords
0A
0B Synth Voice 0C Synth Voice 0F Bass Synt Synt 0F Bass Synt
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
03 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum 04 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum
07 Lead 12 Synth Voice 13 Synth Voice 0B Synth Voice
Synt 10 Bass 11 Bass 11 Bass 11 Bass
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
03 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum 05 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum
0C Synth Voice 14 Guitar + vocals 14 Guitar + vocals 0D Chords
11 Bass 11 Bass 11 Bass 11 Bass
50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
05 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum 06 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum
0D Chords 12 Synth Voice 13 Synth Voice 0B Synth Voice
11 Bass 11 Bass 15 Bass 21 Lead
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81
03 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum 05 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum
0C Synth Voice 14 Guitar + vocals 14 Guitar + vocals 0D Chords
21 Lead 11 Bass 16 Bass Synt Synt 17 Bass Synt
82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
05 C.+T.+H.+Snare+Drum 00 1A Tamburine
1B C.+T.+H.+S +D +HC
0D Chords 00 00
Guitar 1D Synt voice
Synt 18 Bass Synt 0C Synth Voice 11 Bass 19 Bass
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
03 C.+T.+H.+Snare+Drum 20 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum
1F Synth Voice 1E Synth Voice 0B Synth Voice 0C Synth Voice
21 Lead 22 Lead Bass 21 Lead 21 Lead
114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129
03 C.+T.+H.+Snare+Drum 05 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum
14 Guitar + vocals 14 Guitar + vocals 07 Lead 07 Lead
11 Bass 16 Bass Synt 23 Synth Voice 0C Synth Voice
130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145
03 C.+T.+H.+Snare+Drum 05 C.+T.+H.+Snare+Drum 03 C.+T.+H.+Snare+Drum 28 C.+T.+H.+Snare+Drum
Guitar 25 Synth Voice Guitar 25 Synth Voice 07 Lead 07 Lead
11 Bass 11 Bass 24 Synth Voice 26 Synth Voice
146 147 148 149
00
00
27 Synth Voice
Patterns and instruments
One of the points that was not good in GT2 was the impossibility to change the ADSR of a instru-
ment into the wavetable. This is very useful for having two instruments in sequence to reproduce
(like a Drum + Bass) that may have different ADSR for better result.
But this was in version 2.35 (that I start to use): in next version the function was implemented let-
ting a pattern command to be executed from wavetable.
39
===== page 40 =====
Improvements
Before speaking about some improvements that I see for the editor, one point that now is perfect
is the delay between a note played and the time you listen it in the PC. With GT in Linux there was a
delay too high and you cannot follow a pattern being played. In GT2 all is in realtime.
I don't know if this depends by the editor itself or simple by improved sound libraries in linux, but
this is essential for a good use of the editor.
Now the points I see that could be better, but not essential, to have:
● You can start to play the song from the current start pattern position: with this function you
can, for example, listen carefully only the end part of a tune, without to start always from the
beginning. Using some keys you can move the starting pattern into the editor. The main
problem here is if you use the Rxx repeat command, as GT2 move command moves the po-
sition of the pattern of one voice for all the Rxx in one step: this means that after this you
have the relative position of the tree voices that are no more synchronized, and so you will
listen an out of tune song. This is not a big problem as you can simple don't use the Rxx
command or you may use it only when the tune is finished.
● When you edit a pattern where you use lot of different instruments at each ticks, it could be
very difficult to be sure to have insert the right instrument and in this case having different
colors for different instruments could give an eyes help. That means, for example that you
should have a color picker into instrument editor, and then that instrument will be showed
with that color into the pattern. Maybe the use of more color should be used even in other
cases, like in tables, but they are not essential (it's like to have a programming editor where
you have code completion or not: you can program anyway, but in the first case you have
more help into your work).
● MK4 did not sound onto it. Even if MK4 (linux) driver has some similitude with MK3, maybe
40
===== page 41 =====
the 2 sids support into MK4 has break the sound if it is generated in the mode GT2 manage
the virtual drive port. However GT2 did not support yet MK4, so if it works this will be only a
side effect of MK3 support. This is actually the features I miss, as I have to save the tune as
Sid and going to listen it into MK4 with a sidplayer for having the real chip sound.
● The save function is very dangerous: it is very easy to overwrite an existing song file as it not
asks for overwriting a already present file (this come out when you press F11 instead of F10
for loading another tune)
● The Undo for commands is a features that in a PC based program could be implemented
easily and that can be useful in some cases (sometimes I have to reload a previous version
of the tune as when I listen that the last steps I made was not what I want, trying to going
back to the last operations could not so simple).
Conclusion
I have found the tables (wavetable, pulsetable and filtertable) very powerful as now you can have
very complex instruments that you can control accurately. The other point that GT2 has over other
editors, is that his user/control interface is very intuitive and you have all under you keys (or even
under the mouse, as now you can even use it). As soon as MK4 will work onto it, I'm sure to have all
I need from the editor.
The other winning point of GT2 is the accurate documentation that follow the program (and even
in the F12 help screen) as all the features are describe in all details.
Relative on my cover, I can only says that maybe I use very strong drums and a sort of old-school
sound in some part. But as I like it, I think that my task is completed for me.
You can download the tune here:
http://digilander.iol.it/ice00/download/what.zip
41
===== page 42 =====
42
===== page 43 =====
SIDin 10 end
43