Canyon.mid
canyonmid.com
canyonmid.com
I had to look up the Tandy 1000 RSX, because it seemed very wrong to have 16-color VGA graphics coming out of a computer labeled as "Tandy 1000".
Tandy 1000 RSX was the last model from 1991, and it had Super VGA rather than the famous "Tandy graphics" that originated with the IBM PCJr. It did not come with an Adlib or Sound Blaster card, which is what was depicted in the YouTube video. But the computer did have one ISA slot, and an Adlib or Sound Blaster compatible card could have been installed.
It also had a 386 processor rather than the 286 normally found on Tandy 1000 computers, and 1MB of RAM.
My, where did the years go...
Good times.
https://assets.superhivemarket.com/store/product/190602/imag...
It was one of those moments where computers blew my 8 year old mind, I was never the same.
You may enjoy chiptunes. Wikipedia has samples: https://en.wikipedia.org/wiki/Chiptune (see: Contemporary chiptune music)
There is no "correct" sound for a MIDI file as it's just note and tempo data. But many people probably associate them with the OPL2 synthesizer chip on AdLib and early SoundBlaster cards. [1]
Now that we have high fidelity digital sound output on even the cheapest computers/devices, at least 44 kHz, at least two channels, and at least 16 bits per sample, we can emulate (or play a recording of) anything.
[1] Personally I remember this midi file sounding different/better. Maybe because I'm remembering using a Sound Blaster AWE64 while playing these things in Windows?
There's also a "correct" sound insofar that the tracks were usually arranged using one sound module or another. Even when devices are compatible in the sense that they have adopted the same standard, hearing music on another device than the one the soundtrack was originally arranged for will cause some degradation, because the standards are only loosely specified in terms of timbre, volume levels, envelopes etc.
Some DOS games have specific arrangements for a variety of different kinds of MIDI devices for these reasons, with different mix levels and instrument setup, sometimes scaled down arrangements, adjustments for the instrument patches and even loading entirely new patches onto the devices.
General MIDI as a format basically died for serious computer use in the early 2000's. It's still a feature of some home keyboards and things, but it is a case where they genuinely do not “make them like they used to”. If you want the greatest possible GM support, you have to buy old hardware from e.g. the Roland Sound Canvas or Yamaha MU series, or obtain old software (no longer sold) like Roland Sound Canvas VA or Yamaha S-YXG50.
It seems like a missing market for vendors to directly support better sound fonts for Windows, maybe we'd see more files written for it again. But also, there's some sense that most of the best written General MIDI files for direct consumption were always the ones for videogames and it has been left as a matter for emulation for that reason as videogames themselves moved on to different tech.
Still though, there's something about loading Monkey Island 2 with a Roland MT-32 era sound font in DREAMM and experiencing what it does with "live" MIDI instrumentation that modern games wish they could do, and sounds nearly as good doing it. For certain values of nearly, of course. Redbook audio gave us all the lovely little "mistakes" of live band things like jazz performances that MIDI could never quite replicate in the same way. I wouldn't necessarily trade modern orchestration for MIDI in all cases, but it is easy for me to imagine that it might be great to see both side-by-side and the best of both in interactive media, best tool for each job style.
I also want to point to Nuked SC-55, but that is also not based on SoundFont®.
Although I have to admit the cheesy versions do sound more familiar to me because I certainly had the most cheapo-grade sound hardware in any 90s PC I had access to.
In the case of canyon.mid there is because it was composed for a specific midi instrument with a particular set of timbres.
Or to put it another way, it’s music and therefore complicated.
QRS is still in business and some of its product lines use Midi conversions of piano rolls and last month I landed a mixed lot with a bunch of those Midi files on 3.5” floppy disks.
I had George Gershwin rocking a Yamaha XG clavichord and Liberace on a Sitar and similar shenanigans.
In particular, think of how this would play out if you had never heard that song before, would that be what you were after?
I was thinking about retelling the story I wanted to retell.
Thinking about writing.
The banjo was just an excuse.
I didn’t hear it.
I am reminded of Mean Streets and Martian Memorandum, which let the PC speaker output something beyond just bleeps and bloops.
If you've heard music from the Sega Genesis, that's the same Yamaha FM Synthesis technology (details differ, of course).
So, games would often support different audio hardware. Here's a favorite:
https://www.youtube.com/watch?v=1IOL4q5tDDQ (Secret of Monkey Island theme, PC Speaker)
https://www.youtube.com/watch?v=8qpB-HAZsRs (Secret of Monkey Island theme, Adlib/OPL2/FM synthesis, common option for PC sound hardware)
https://www.youtube.com/watch?v=i3dB0qEcG20 (Secret of Monkey Island theme, Roland MT-32/Linear Arithmetic synthesis, premium hardware, likely what the music was originally composed for)
Finally, “high fidelity” is not a synonym for “musical.”
The great thing about MIDI is that it is easily routable to any number of things (physical instruments, samplers, etc.).
Being able to listen to Sonic 2 - Chemical Zone with a combination of a Minimoog Model D and a Jun-6 (basically a Juno-6) is unbelievably fun.
It’s human nature to think of familiar things from our youth as the height of achievement. That was the time of the best music, the best movies, the best culture, the best sports, the best everything. No matter if you were born in 1950 or 1990.
You can argue that software does much more than before, sure I agree but no one asked for so much bloat and features in every day apps. My note taking app doesn't need AI.
Are you talking about the 90s or now? Because those were all at least as true then as now. Everything took forever. You needed more RAM every month. Everything crashed constantly. I had to reinstall Win98SE so many fucking times that I can still type F73WT-WHD3J-CD4VR-2GWKD-T38YD from memory.
The amount of suck in commercial software is constant. Companies always prioritize adding the shiny-looking features that sell software to rubes over improving things like memory use, response time, and general quality of life until the quality of life is actually bad enough to drive customers to another vendor, so it's perpetually bad enough to keep the average customer right on the edge of "oh fuck this, I'm switching to something else."
My whole machine reboots in less than 10 seconds. I haven’t seen a blue screen of death in a decade. I haven’t had significant data loss from a failed drive or a corrupt machine in.. I can’t remember. Even DaVinci Resolve is ready to run in a few short seconds.
This is all on a machine I bought 6-8 years ago. I reboot my phone and watch and laptop when I think to, not because I have to. I run half a dozen browsers and hundreds of tabs and play YouTube while waiting for a remote machine to deploy to an immutable temp instance that gets destroyed after every test cycle.
I speak to my AIs and I can live and work anywhere on this planet that legally allows me.
There are problems in our world and on our machines and in our governments but apps don’t take 10 seconds to load.
Except ServiceNow. I’ll give you that one.
As if anything done on a machine going through 10_000_000_000 (10 Giga) instructions per seconds should be anything but perceived instantaneous, for it's finite lifetime human user.
(Though I do think it takes significantly longer to start on my 2024 MBP…)
Kicad loads in 2 seconds, and FreeCAD in 7 seconds though.
Part of my job when I was a youngn, was rebooting the Windows NT server running the software router because packets stopped forwarding for the entire net cafe.
Every generation has its hype cycle; it's nothing new.
Is it? I think there's a common tendency to "stop exploring" cultural artifacts very deeply as we age, but not everyone shares this trait. Some people continue to value novelty in those areas well into old age.
For my part, treasured artifacts of my youth don't impede my ability to appreciate new things. And indeed, I think many videogames I loved dearly have aged poorly.
And to me the best desktop experience in term of software has been gnome 3 after it had time to hone its jump forward from its previous major release. So, not the newest hot thing out there, but not my first crush.
Regarding forward, augmented reality on glasses seems to have great potentials, but I don't have much hope foe the default systems they will come with. A future where most people wear those stuff filled with signal tailored by the ad industry and whatever governments is just not letting much room escaping the obvious various dystopian scenarios.
U.S. debt in the 1950s was well below $1T, in the 1990s it was something around $3T, now it's $36T.
Having said that, I wasn’t the only person deeply upset when Greenspan gave the green light for Bush’s tax cuts. Under Clinton we were on track to wipe out the debt in another 10 years.
Things in the 90s were more straightforward because supply chains and business processes were much shorter and less complex. What people interpret as nostalgia might actually reflect a recognition that systems/products genuinely were more efficient before they became increasingly layered with intermediaries and dependencies.
An illustration of these dependencies and layers is debt - the mounting complexity parallels the mounting debt levels.
It explicitly excludes the values of the assets that we hold in security of the debt, and is frequently trotted out to remind the little people how terribly large all of the problems in the world are and how hopeless it would be for any of us to try to solve them.
But!
No where in the world is the adage about, "If you owe the bank $100,000 you have a problem, but if you owe the bank $100,000,000 the bank has a problem" more evident than in the US National debt.
America could pay off this debt at any time, but doing so would cause our economy to massively deflate and cause a lot of harm and unhappiness for people all over the world, including Americans.
Not telling us these things gives America a way to inflict fear and control on its populace with no measurable downsides.
Had they? I gamed in the 90s and I game now. And boy, its not even remotely the same and iam thrilled to see what comes next (hello, gta6)
No words for how this game blew my mind back in those days...
I am happy with the potential that we have available today to do things that we couldn't in the past. And it's always possible to improve software on top of more capable hardware and OSes.
Software in the '90s was mostly driven by altruism, software in '20s starts with an A-round.
(… although there are sometimes "seed" rounds that precede a series A, or even pre-seed rounds … like everything else, it's complicated & messy. But hopefully you see the metaphor the parent was trying to paint.)
echo "https://news.ycombinator.com/reply?id=44286820" | cut -d'/' -f3
Nearly 19000 games were released on Steam in 2024. A lot of the most interesting stuff that came out simply wouldn't have existed 10 or 20 years ago. I think it's great that those things can exist now and potentially find an audience.
It will never be enough until we can manipulate the fabric of space and time directly as gods and create entirely new universes and physics and live forever for an infinity infinities.
The ratio of our infinitesimal, geologically small existence to the whole of the light cone and the observable universe - it is just a glimpse at the fractal of what will be enough to satiate our curiosity and desire.
The same drive for betterment that made our species “kill such beautiful minimalism” was the one that lifted billions out of subsistence farming and 50% infant death rate, and will be the one to escape the destruction of planet earth by sun’s evolution. You cannot have one without the other.
I'd say try Linux.
Curiosity and desire can be focused on minimalism and elegance of the smallest most essential cores of whatever is at stake.
Also because everyone seems too scared to practice adversarial interoperability.
Also because SoCs are now a thing which allows unhackable secure boot and other DRM-like functionality that prevents people from modifying their own devices to act in their own interest, or, as is the case with Android devices, allows it but punishes the user for having gained full access to their own device.
- A Windows 3.1 window manager theme
- The Windows 3.1 fonts with font hinting/antialiasing disabled
- Windows 3.1 icons
- A matching cursor theme
- Lower your display resolution
The audio sounds like it sounds because Microsoft Licensed Roland’s GS Wavetable. Without that you lose timbral information.
And any OPL is almost certainly emulated, not an actual chip.
Perhaps somewhat surprisingly, most video codecs aren't tremendously efficient with static screens.
* Emu Proteus 1: https://www.youtube.com/watch?v=K5FffG_0sqw
* Emu Proteus 2: https://www.youtube.com/watch?v=O4KW9uWCY3A
* Roland MT-32: https://www.youtube.com/watch?v=jdSKg5G9MPc&t=22s
* Roland D-10: https://www.youtube.com/watch?v=zXGdyp7Ml-Y
* Roland SC-33: https://www.youtube.com/watch?v=as_jVNIvleI (complete with MIDI animation)
* Yamaha MU100: https://www.youtube.com/watch?v=6BL_RzeWDxg (1 hour!)
Etc. Don't know if they still do, but it was a thing when these "romplers" came on the market.
1. generate audio signal
2. reduce volume of that signal, losing information because it's quantised
3. take that volume-reduced signal and boost it right back up again, but now with the lower bits destroyed
You can make this effect as bad as you like, e.g. turn it down to 1% and then amplify by 100x... but why?
https://www.izotope.com/en/learn/what-is-dithering-in-audio....
More information: https://www.youtube.com/watch?v=iuEtQqC-Sqo
I'm aware of ReplayGain and this processing is important for per-track overall gain, but what I'm getting at is lower level: instead of there being two lossy/rounded stages of dimming and amplification, you want to communicate to the OS a log2 "dimming factor", so that this can be subtracted from a later log2 amplification factor such that we ideally waste no processing time if the sum is zero, and otherwise don't suffer the twice-quantised signal degradation (at most one accurate scaling pass, instead of two arbitrarily precision-reducing ones). It's maybe a minor point / imperceptible as others have noted, but IMO this seems like the Correct (TM) approach.
When properly understood.
And provide a consistent interface for managing.
When thoughtfully placed in a signal chain.
I'd be much more worried about 44.1khz sources being resampled to 48khz if that's the OS playback rate. I mean you won't be able to hear that either in practice but at least it's not negligible.
Truly first-class audio with sublime control plane ergonomics.
That'll never happen since any random app developer can just multiply audio volume by a float in whatever API and attach their own unique take on a slider. I'll merit the cases where you need to have individual level and channel controls, such as editing software and professional music tools, but most apps are not these.
It's times like this when I do appreciate Apple's dictatorial take on things, though even they could not win this fight.
Clearly there is a need to give different volumes to different apps, so you can have quiet background music while a timer app is louder, or Zoom is louder.
Ideally there would be an OS-level mixer to independently set the volume of each app. I believe Windows has this, Mac definitely doesn't. And for convenience, an app's local volume control would exist, but set it at the OS mixer level, so you don't have them competing with each other.
But without this, an app does have to have local volume controls.
Also, it's important to be able to set gain as well, i.e. turn the volume "above 100%". For those YouTube videos that for some reason are only 5% as loud as other videos. Even better is if you can set the gain per-video so that it won't be absurdly loud and clipping when you move on to the next video.
Bonus points if an OS or media player ever gives the option of a dynamic compressor, so you can actually listen to those amateur podcasts where one speaker's microphone is 10x quieter than another's. Or listen to the quiet parts of classical music recordings even in the presence of background noise.
https://github.com/kyleneideck/BackgroundMusic and others.
Clicks for the click God!
There are no controls to indicate that you can pause and restart, but this just-click-anywhere-to-play/pause has been standard on all video players everywhere for a long time.
I've been on Linux for so long now, that being able to just play a MIDI file without making a bunch of decisions about soundfonts and synthesizers [1] just seems mind-blowing to me now.
Part of me wishes that just by default, mpv or something would just pick a softsynth and just play it (like WMP here) rather than have me install a separate program, pick a sound font, invoke it in some weird way to let it know what soundfont I want, and not even be able to seek back and forth.
[1] https://wiki.archlinux.org/title/MIDI#List_of_SoundFonts
All right. Apropos the retro theme, here's a Web page dedicated to my Apple II program "Electric Duet": https://arachnoid.com/electric_duet/
The gory details: the Apple II had a simple TTL output connected to a speaker -- crude, not meant for music, and certainly not with two voices. Did this stop me? Read on.
I created an assembly-language player that switched the TTL speaker driver's output at 8 kilohertz, then created two musical voices by controlling the pulse width of the 8 KHz clock. So two voices, two notes at once, from a TTL driver.
Here's a sample of the music Electric Duet created, on an Apple II, in 1981: https://arachnoid.com/electric_duet/music_tracks/prelude.mp3 -- press the play button.
That's retro defined.
What a step up it was over the bleeps that came out of the PC. Like going from B&W to color TV. Advances like that were commonplace back then.
Maybe we'll have fun again with quantum computing.
"The Qamiga has 16 qubit sound dude!"
"Yeah, but it uses quantum annealing! You can hear the entropy loss in the higher frequencies!"
Canyon.mid played using modern real instruments: https://www.youtube.com/watch?v=1QN3ToFCv_s
The monitor is the standard Tandy VGA monitor of the system's era. The styling on the speakers feel newer than the RSX's 1991 launch, they're more what I'd expect from the mid to late 90s.
You had to upgrade the VGA chip's BIOS to use Win 95 on it:
> The ACUMOS VGA graphics can be software-updated with Cirrus Logic BIOS (via MS-DOS driver) to allow VESA/SVGA to function in Windows 95, as the Windows 3.xx Tandy VGA drivers are insufficient for Windows 95.
ref: https://gunkies.org/wiki/Tandy_1000_R-Series
I think the background image is probably authentic, it has the feel of a mid-late 90s digital camera picture. It reads to me as the desk of someone who is trying to keep that system alive long past its prime years. Which were arguably over before they started, given 486 systems had been available for a bit when this launched. We end up with an early 90s system with a handful of mid-late 90s peripherals.
The bigger problem to me is this sounds like MIDI played back on a sound card with FM synthesis. The 1000 RSX had the poorly supported Tandy 3 tones + DAC sound hardware. You could install an AdLib, Sound Blaster, or other card to give it MIDI FM synth capability, but the base system can't do it. Alas, we can't see the back to see if it has such an upgrade...
When I was a kid, I was using Windows 95 for a while when Me was already a thing - newer versions could technically run, but the experience wasn't great on that hardware. You could even still find Windows 3.11 computers at my school at the time. Computers don't go die at the exact moment a successor becomes available on the market.
* III.C.1. Can I run Windows on my 1000?
...
The RLX's can run Windows 3.1 in standard mode only, if they have the RAM upgraded to 1M. The RLX just barely meets the minimum hardware require- ments for Windows 3.1, however, and performance will be poor. Windows will not recognize the built-in mouse (see section II.G.2.). One user says of Windows 3.1 on the 1000RLX:
Windows' performance is tolerable on a 486DX2/66. I like it on the RLX because I can start a program, go to the bathroom, and when I come back only have to wait a few minutes before I can actually use the #@$% thing.
The RSX's can run Windows 3.1 (or 3.11) in 386 enhanced mode if the memory has been upgraded to 2M or more. There is a Windows sound driver for the RSX's built-in sound at my ftp/WWW site and at Tandy's support WWW site (see sections IV.B.1. and IV.B.2.).
Tandy does not officially support the use of Windows on any model of the 1000-series. The RSX's could theoretically run Windows 95, but Microsoft does not recommend Win95 for 386's.
https://www.youtube.com/watch?v=gRGOm071sE0
And I still love that tune.
Looking at it now, I love how they used a scroll bar as the UI widget to represent playback progress.
https://www.brianorr.com/blog/2010/01/14/windows-95-easter-e...
So that particular track never gelled with me.
I listened the heck out of some of the other included sound and MIDI sample material however, back then. As one did, with sources being so comparatively limited in the pre- and early- Web years.