I briefly listened the one track I knew from these and wow, what a difference 🤣. I'll have to do a proper listen with something beyond crappy laptop speakers to evaluate artifacts etc. later ~
EDIT: I listened through the Depeche Mode track on big speakers now and the main negative thing I noticed is that while much of the sound got a lot brighter and clearer, there often were "vibrating" noises happening near transients and a quite audible vibrato effect in many of the sounds, this sounded really offensive to me so I am not sure I would want to enable it myself, but perhaps it works better on higher quality sources that are closer to the ideal so less needs to be painted back and with that, less artifacts.
I have not listened to the other tracks, and especially since I don't know them to know what they actually have to sound like.

Thank-you for the feedback, Tiido! I know exactly what you're referring to with the artifact behaviors. Some of that can yet be tweaked; I mostly just focused on trying to match the sound of the original (CD) recordings with a single, generally-applicable effect configuration. With the entire endeavor being destructive in nature, there's quite a bit of subjectivity, compromise, and error-tolerance involved. As such, I would say that this is very much a "casual listening" experience, rather than anything too serious. 😀

That said, I'm pretty impressed with the overall results! Given a choice between the non-effected and effected examples presented, my pick would be the latter every time.

Cloudschatze wrote on 2023-05-03, 03:45:

[…]
Show full quote

• 8-bit resolution
• 22kHz playback rate
• Single-channel (mono) output
• ...and a static, 5.3kHz cutoff frequency imposed by the low-pass circuitry.

5.3kHz? Either the designers were drunk, or they chose that cutoff as a compromise so 11kHz samples also sound good. For 22kHz, you should get way better results at 8-10kHz (for simple filters with low steepness, better go 8 ). At that time, using 11kHz samples to conserve space might have been a very relevant use case, though.

The circuit designer, Frank Durda IV, has said that the 5.3kHz filter value was chosen solely for the purpose of computer-noise suppression. With the first DAC-bearing Tandy systems having been released in 1988, the effect this would have on higher playback rates likely wasn't even a consideration. That Tandy continued to use the same filtering in later systems is unfortunate, but still "better" than that of the contemporaneous Sound Blaster cards.

I hate those things. might as well chuck an alesis 3630 on it as well 🤣 there are so many exciter type devices, aka " lets add analog harmonics back into sterile digital recordings!", they can be fun to play with. I had an old silver face aphex exciter, didnt care for it much.

aphex released quite a few models of exciter. there is also the BBE Sonic Maximizer! You can also drop 2.5k on a new EL7x Fatso from empirical labs 😀 or a crappy behringer for $100 new, but since its behringer it will just make it sound like 2bit mono.

be interesting to run an exciter over the normal SB16 output.

reverb makes everything better...

I love it. Sure, this kind of processing can exhaust the ears quickly, and you won't be "improving" your lossless audio libary by pumping it through an MU100. But in the Tandy context it's a really impressive application.

I was somehow ignorant until this moment of the Yamaha's audio input and range of effects... uh oh.

I tried unsuccessfully to replicate the enhancements with the Crystalizer effect on my Sound Blaster X4. Do any sound cards have the Aphex Aural Exciter or do you have to have an external module?

Cloudschatze wrote on 2023-05-03, 18:25:

The circuit designer, Frank Durda IV, has said that the 5.3kHz filter value was chosen solely for the purpose of computer-noise suppression. With the first DAC-bearing Tandy systems having been released in 1988, the effect this would have on higher playback rates likely wasn't even a consideration.

Thanks. That makes sense. The lower the cutoff, the less "hiss" gets to the speakers. It's also a quite good (possibly slightly high) cutoff for 11kHz anti-aliasing filtering, and also works good enough to prevent aliasing at 8kHz sample rate to hit the ranges the human ear is most sensitive to (around 1kHz). At 1988, with software still being started (not just installed once) from floppy disks, optimizing the design for noise suppression and low sample rates was a good approach.

I spent an inordinate amount of time tweaking the MU settings further (because I lead a life of constant excitement). Does the following version sound any better than the first?

Depeche Mode. (1990). Enjoy the Silence [Excerpt]. On Violator. Mute Records.
♫ Tandy PSSJ output (4.8 MB FLAC)
♫ Tandy PSSJ -> Yamaha MU100 output - Version 2 (6.4 MB FLAC)

Stretch wrote on 2023-05-06, 15:24:

I tried unsuccessfully to replicate the enhancements with the Crystalizer effect on my Sound Blaster X4.

I'm afraid I don't have any experience with Creative's Crystalizer effect. It does appear to be similar in principle.

Closing the loop on this thread, the effected examples in the parent post now reference recordings that leverage the "version 2" MU100 configuration, detailed as follows:

SYSTEM
================================
AD PartLock = on
Mlt EQ Lock = on

A/D
================================
Input Preset: ST Audio (019,002)
Rcv CH = OFF
Volume = 100
Expresn = 127
Pan = C
RevSend = 000
ChoSend = 025
VarSend = 000

CHO
================================
Type = Chorus 1 (default)

INS1
================================
Type = Aural Exct
HPF Cutoff = 500
Drive = 026
Mix Level = 051
AC1Ins1Ctrl = +00
Ins1 Part = AD01

INS2
================================
Type = 3-Band EQ
Low Gain = +12
Mid Freq = 2.0k
Mid Gain = -02
Mid Width = 1.0
High Gain = +10
AC1Ins2Ctrl = +00
Ins2 Part = AD01

EQ
================================
Type = Pops

This A/D effect configuration operates near-independently of the tone generator functions of the MU, allowing for proper, and concurrent, General MIDI, GS, or "basic" (MU50-level) XG playback. Shared functions include the master EQ, which is locked to the "Pops" setting (and which I find more-appropriately matches the tone of an SC-55), and the Chorus type, which is simply based on the GM System On, GS Reset, and XG System On default. Consequently, while realtime or MIDI-based changes to the Chorus also affect the A/D part, the intended configuration is restored upon receipt any of the aforementioned reset messages.

Finally, the above A/D configuration is tailored to the 5.3kHz, lowpass-filtered output of the Tandy PSSJ-bearing systems. The need for adjustment should be assumed for any other application.