Uploaded June 2020 | Updated November 2026, 15 hours ago
(Reuploaded due to display issues with the original video settings and Google Chrome's browser.)
A Tandy 1000 TX, along with a 1000 TL and 1000 TL/2, has an 8MHz 286 processor. These systems were released in 1987-89 and were fairly competitive price and performance-wise for the time, especially at the start of the period. But how well did they perform with games of the era and can their performance be improved?
Sierra's adventure games from this period were some of the largest and most demanding games on the IBM PC platform. The SCI engine was introduced in 1988 and offered full support for 320x200 graphics in 16 colors, sound cards and hard drive installation. But they were large games, often requiring 8-9 5.25" 360KiB floppy disks. I have chosen to use Leisure Suit Larry 3 : Passionate Patti in Pursuit of the Pulsating Pectorals from 1989 as a popular game which demands much of a system of the time. Sierra was still very much a Tandy 1000 supporter at this time, it even gave a way a TX in a contest published in its magazine.
I have divided this video into four parts. The first part compares the game booting and running on two different hard drive interfaces. The second compares an unmodified game vs. a modified version of the game.
Hard Drive Comparison - In the first and second segments of the video, I am running the game (unmodified and modified) off a Seagate ST-225 20MiB MFM hard drive. The hard drive interface is a WD-1002A-WX1 compatible with the ST-506 hard drive interface. This is a typical storage device of the time, the ST-225 was a reliable drive, 20MiB was still a good size by 1988's standards and the hard drive interface was also a decent performer supporting at least a 3:1 interleave. (A more aggressive interleave could have been imposed on an AT class system, but I pulled the drive and interface from my IBM PC XT.) Despite the size of the ST-225, it can fit into one of the 1000 TX's bays and the system can power both it and a floppy drive.
In the third and fourth segments of the video, I am using a SanDisk Ultra II 1GiB Compact Flash card for the "hard drive" and an ADP-50L with a 2.18T BIOS for the interface. The performance benefits of solid state memory are enormous, even with the relatively modest Compact Flash I am using. Moreover, the ADP-50L is an IDE compatible interface but is not a typical hard drive interface because it relies on memory mapped I/O instead of port mapped I/O of the ST-506 interface. The use of memory mapped I/O improves the speed on an 8-bit expansion bus compared to the port mapped I/O bus of the ST-506 interface cards.
Software Comparison - I am using the latest English version of LSL3, Version 1.021, Interpreter 0.000.572, released on January 27, 1990. In the first and third segments of this video, this software is untouched and run as if I had just installed it from the floppy disks.
In the second and fourth segments of this video, I am using a modified version of the program. My friend Cloudschatze supplied me with a modified MT32.DRV file which removes the wait time between sending System Exclusive MIDI Messages to the MT-32. This should improve booting speed as it will take the game less time to transmit its custom instrument patches to the MT-32. At 286 speeds, the SCI engine does not send the messages fast enough to cause the dreaded "Exec. Buffer Overflow" messages on an old revision MT-32 like the one I am using that faster systems might.
My friend NewRisingSun additionally supplied me with a version of LSL3 which has a single decompressed resource file instead of the multiple LZW-compressed files of the unmodified game. This should also decrease not only boot times but the transition times as you move from screen to screen. The uncompressed version is approximately 25% larger than the compressed version, but this game was made in the era where games were expected to be run completely off floppies, making any savings in the number of disks to be used significant.
In all four segments, my pattern has been nearly identical, I get to the point where the introductory story begins, then enter the game by bypassing the age check, then going to the house and to the advertising agency, finally ending up at the exterior bathrooms. In the fourth segment I additionally used a pair of "speed-boosting" utilities called "FAST.COM" and "SPEED.COM". FAST.COM is supposed to speed up hard drive and floppy drive transfers, but I doubt it did anything. "SPEED.COM" is supposed to reduce DRAM refresh on the 1000 TX, so it may have offered a minor boost.
When the screen shakes at the phone booth, the Datapath loses sync for a moment, hence the black screen.
Segment Timings :
0:00 - MFM Drive & Original
6:08 - MFM Drive & Decompressed
11:39 - CF Drive & Original
17:19 - CF Drive & Decompressed
(Reuploaded due to display issues with the original video settings and Google Chrome's browser.)
A Tandy 1000 TX, along with a 1000 TL and 1000 TL/2, has an 8MHz 286 processor. These systems were released in 1987-89 and were fairly competitive price and performance-wise for the time, especially at the start of the period. But how well did they perform with games of the era and can their performance be improved?
Sierra's adventure games from this period were some of the largest and most demanding games on the IBM PC platform. The SCI engine was introduced in 1988 and offered full support for 320x200 graphics in 16 colors, sound cards and hard drive installation. But they were large games, often requiring 8-9 5.25" 360KiB floppy disks. I have chosen to use Leisure Suit Larry 3 : Passionate Patti in Pursuit of the Pulsating Pectorals from 1989 as a popular game which demands much of a system of the time. Sierra was still very much a Tandy 1000 supporter at this time, it even gave a way a TX in a contest published in its magazine.
I have divided this video into four parts. The first part compares the game booting and running on two different hard drive interfaces. The second compares an unmodified game vs. a modified version of the game.
Hard Drive Comparison - In the first and second segments of the video, I am running the game (unmodified and modified) off a Seagate ST-225 20MiB MFM hard drive. The hard drive interface is a WD-1002A-WX1 compatible with the ST-506 hard drive interface. This is a typical storage device of the time, the ST-225 was a reliable drive, 20MiB was still a good size by 1988's standards and the hard drive interface was also a decent performer supporting at least a 3:1 interleave. (A more aggressive interleave could have been imposed on an AT class system, but I pulled the drive and interface from my IBM PC XT.) Despite the size of the ST-225, it can fit into one of the 1000 TX's bays and the system can power both it and a floppy drive.
In the third and fourth segments of the video, I am using a SanDisk Ultra II 1GiB Compact Flash card for the "hard drive" and an ADP-50L with a 2.18T BIOS for the interface. The performance benefits of solid state memory are enormous, even with the relatively modest Compact Flash I am using. Moreover, the ADP-50L is an IDE compatible interface but is not a typical hard drive interface because it relies on memory mapped I/O instead of port mapped I/O of the ST-506 interface. The use of memory mapped I/O improves the speed on an 8-bit expansion bus compared to the port mapped I/O bus of the ST-506 interface cards.
Software Comparison - I am using the latest English version of LSL3, Version 1.021, Interpreter 0.000.572, released on January 27, 1990. In the first and third segments of this video, this software is untouched and run as if I had just installed it from the floppy disks.
In the second and fourth segments of this video, I am using a modified version of the program. My friend Cloudschatze supplied me with a modified MT32.DRV file which removes the wait time between sending System Exclusive MIDI Messages to the MT-32. This should improve booting speed as it will take the game less time to transmit its custom instrument patches to the MT-32. At 286 speeds, the SCI engine does not send the messages fast enough to cause the dreaded "Exec. Buffer Overflow" messages on an old revision MT-32 like the one I am using that faster systems might.
My friend NewRisingSun additionally supplied me with a version of LSL3 which has a single decompressed resource file instead of the multiple LZW-compressed files of the unmodified game. This should also decrease not only boot times but the transition times as you move from screen to screen. The uncompressed version is approximately 25% larger than the compressed version, but this game was made in the era where games were expected to be run completely off floppies, making any savings in the number of disks to be used significant.
In all four segments, my pattern has been nearly identical, I get to the point where the introductory story begins, then enter the game by bypassing the age check, then going to the house and to the advertising agency, finally ending up at the exterior bathrooms. In the fourth segment I additionally used a pair of "speed-boosting" utilities called "FAST.COM" and "SPEED.COM". FAST.COM is supposed to speed up hard drive and floppy drive transfers, but I doubt it did anything. "SPEED.COM" is supposed to reduce DRAM refresh on the 1000 TX, so it may have offered a minor boost.
When the screen shakes at the phone booth, the Datapath loses sync for a moment, hence the black screen.
Segment Timings :
0:00 - MFM Drive & Original
6:08 - MFM Drive & Decompressed
11:39 - CF Drive & Original
17:19 - CF Drive & Decompressed






![Star Wars: Dark Forces Complete General MIDI Roland SC-55 Soundtrack
The complete General MIDI soundtrack to Star Wars: Dark Forces, composed by Clint Bajakian using motifs by John Williams, played via NukedSC55 on a Roland SC-55mkII. Notwithstanding the dynamic nature of iMUSEs playback, the MIDI files have been edited such that all non-repeated snippets that could play are included in this recording. Each level has a Stalk and a Fight piece, and Levels 7-12 use the same two pieces each as Levels 1-6. This recording therefore presents the music for levels 1-6 using the respective Stalk and the music for levels 7-12 using the respective Fight piece, including one set of the several dynamic transitions between the two in each. The three boss pieces seem to be unused in the game according to all playthrough videos, though some people insist they have heard them in-game back in the day.
0:00:00 - LucasArts Logo
0:00:09 - Main Title (Star Wars Main Theme)
0:01:49 - Opening Credits (Dark Trooper Theme)
0:04:51 - The Death Star Plans
0:07:16 - Launching the Dark Troopers
0:09:38 - Mission Briefing
0:11:57 - After the Massacre
0:16:34 - The Subterranean Hideout (Anoat City)
0:21:50 - Imperial Weapons Research Facility
0:26:03 - Approaching Gromas 16
0:26:16 - The Blood Moon (Gromas Mines)
0:28:51 - Facing the Dark Trooper [unused]
0:30:55 - Gromas Mines Destroyed
0:31:02 - Crix Madines Fate
0:32:21 - Orinackra Detention Facility
0:36:49 - Deadly Cargo (Ramsees Hed Docking Port)
0:39:44 - Approaching Planet Anteevy
0:39:55 - Ice Station Beta
0:42:01 - Robotics Facility Destroyed
0:42:09 - The Death Mark (Nar Shaddaa, The Vertical City)
0:44:31 - Jabbas Revenge
0:46:19 - Aboard the Star Jewel (Jabba the Hutts Ship)
0:48:49 - The Imperial Mask (Coruscant)
0:51:06 - Facing Boba Fett [unused]
0:53:54 - Smugglers Hijack (Imperial Fuel Station)
0:56:28 - The Executor
0:59:45 - The Stowaway
1:00:33 - The Dark Awakening (The Arc Hammer)
1:06:19 - Final Confrontation with General Mohc [unused]
1:09:22 - Finale
1:11:25 - Closing Credits
1:14:05 - Kyle Katarns Theme [unused] Star Wars: Dark Forces Complete General MIDI Roland SC-55 Soundtrack](https://i.ytimg.com/vi/RWCLDlyev4I/mqdefault.jpg)


![Jill of the Jungle MS-DOS Adlib Complete Soundtrack Digital Capture
In this video, I have captured Jill of the Jungles complete soundtrack from an Adlib card. However, for these captures I did not record from the headphone output from the Adlib card, instead I captured the audio digitally. Lets start with a track listing and then Ill describe my capturing method :
00:00 - Funky
01:47 - Zeppelin
03:49 - Spiders
05:12 - Ominous
06:36 - Dan
08:20 - Stealth
09:52 - It [Song]
11:50 - Jupiter
13:13 - Seven
15:28 - Peaks
17:25 - Cubist
19:19 - Action
20:52 - Dungo
22:30 - Bonzo
24:08 - Mobydick
26:14 - Undrgrnd
28:05 - Wild
30:03 - Neoclass
31:36 - Kuntamon
33:38 - Desert
35:08 - Alien
36:43 - Deepevil
The track names are their MS-DOS names. Jill Goes Underground and Jill Saves the Prince begin with Peaks and Wild, respectively. Each track is a Creative Music Format (.CMF) file with a renamed extension. CMF was Creatives answer to AdLibs Visual Composer .ROL format. Both formats were used to store FM Synthesis music for the YM3812 found on the Adlib and Sound Blaster cards.
The Yamaha YM3812 OPL2 chip outputs samples in a digital serial format. The output is sent to a small 8-pin DAC chip, the Y3014. The Y3014 converts a data stream of 16-bit signed numbers using a sample rate of 49,715.90...Hz to analog voltage levels and sends that to the amplifier on the Adlib or the mixing stage of a Sound Blaster.
With the aid of a hardware device called Digi-Snap and associated software, one can capture this data stream from the YM3812. The Digi-Snap device is a parallel port dongle with a few logic chips on a small PCB. This board connects via wires to an 8-pin socket. This socket is inserted into a socket on the Adlib or Sound Blaster card where the Y3014 DAC chip resides. The DAC is piggybacked onto this socket.
The Digi-Snap is not a common device. First, there are no schematic files you can take to a company like PCBWay to print up a board. The board I borrowed has hand-made circuit trace etching on one side and used wire-wrapping to make the remaining connections. Additionally, the DAC chip on all Adlibs and many Sound Blasters and clones is soldered to the board and needs to be desoldered and replaced with a socket. Without a motorized desoldering tool the process will not be painless.
This method of capturing requires two PCs. The first PC is where the Adlib is being employed for playback. The second PC is where the parallel port is being employed for recording. The parallel port must have EPP (Enhanced Parallel Port) capabilities and the system should be fast enough to support the required 32MiB of RAM. The software must be run in real mode DOS to avoid skipping or interrupts. The software will record 9 minutes of audio before replacing the oldest samples with the newest samples. Of the three methods of digital FM Synthesis capture using original hardware I descrive here, https://nerdlypleasures.blogspot.com/2015/02/digital-opl2-and-opl3-recording.html, this is by far the most complex method.
I used my 486DX266 for playback and my Pentium III 600EB for recording. I used Creatives PLAYCMF utility with a batch file to play the tracks in an order which would likely be encountered in a full playthrough of the Jill of the Jungle Trilogy. The Digi-Snap software crashed until I disabled the Sound Blaster Emulation of my Vortex 2 sound card. After that the program ran and did its job. Sometimes the first part of a song was cut off and I had to re-record. Once I had captured all the audio, then I imported the raw capture to Audacity and edited it into its final form. Even in the final Youtube upload the audio is still a 16-bit WAV file with a sample rate of 49716Hz. I hope Youtube treats it kindly.
Now you might remark that this recording may not square exactly with your memories of playing through the game. Jill of the Jungles soundtrack may sound different depending on which version of the game you are playing because different versions of Jill embed different versions of Creatives SBFMDRV.COM driver. This driver will not work if it does not detect a Sound Blaster card (and the game will not play music with an Adlib card until v1.2(d)), but NewRisingSun hacked the one used for this playback to support a bare Adlib card. The SBFMDRV.COM driver used is based on v1.22, which was used in Jill of the Jungle v1.0 and v1.2b. The driver found in v1.2d does not support certain music events causing a loss of fidelity in the music playback. Additionally, this driver resets the OPL2 rhythm instruments each time a new music track is played, without the reset the rhythm instruments may not sound as intended.
Even with all this, the soundtrack is not ideal. Jills composer, Dan Froelich, ironically composed his music using Adlibs Visual Composer. The conversion from the original .ROL-formatted files to CMF-formatted files was not without issue. As those originals are not available, these are the best recordings anyone other than he can offer at this time. Jill of the Jungle MS-DOS Adlib Complete Soundtrack Digital Capture](https://i.ytimg.com/vi/RqRFnHzAmCk/mqdefault.jpg)
