OPL2/3 Music on a NEW Yamaha FM Chip (YMF825)

Yamaha still make one FM chip, the YMF825 (a.k.a. SD-1), which has some striking similarities to the OPL3 used in many PC soundcards in the 90s. I wondered if the SD-1 might be able to play OPL music, and if so, what it would sound like.
I built a USB interface and OPL to SD-1 translator, and tested it with various PC game and tracker music - here are the results.
Schematics and firmware for a device similar to the one used in this video are available here: github.com/dan...
Music:
00:05 - Mind the Lemming (Lemmings)
00:31 - Intro (Flashback)
00:43 - Intro (Monkey Island 2: LeChuck's Revenge)
00:59 - Gorbi (KGB)
01:45 - Jamulations - Louis Gorenfeld / Extent of the Jam
02:45 - Monkey Island 2: LeChuck's Revenge
04:16 - Spice Opera (Dune)
05:38 - Samsara - Diode Milliampere
07:10 - Wacky (Oh No! More Lemmings)
08:21 - High Tech Environment III - Benjamin Gerardin
09:51 - Spiral - Phandral

[@ZElTGElST]

I would say that sd 1 sounds a bit more fuller, with more body. In any case both sounds phenomenal. Its a pity that yamaha is not democratizing this chip. The community could do wonders with it. The diode miliampere song is a masterpiece. Thanks for taking the time to develop this video

[@johnmoser3594]

"Democratizing" there's another word that just gets thrown around but has no meaning.

[@CoLD.SToRAGE]

Cheers! First time I’ve heard my Lemmings music coming out of an old OPL 😅

[@danielrfry]

Amazing - Lemmings was huge for me growing up and the music was a massive part of that (we were a PC household but I knew and loved the Amiga version too). So cheers to you!

[@pc-sound-legacy]

This is very interesting to me - great comparison! And great to see Yamaha still offers OPL3 compatible IC today, and a USB FM device with DoSbox support sure will find it's supporters 👍

[@danielrfry]

Thanks - glad you liked it!

[@hagen-p]

The only real SD1 drawback is the lack of stereo (The OPL3 Samsara made me shiver listening to it -- nice, never heard this song before!). Maybe two SD1 chips could fix this? Pretty, pretty, please, with sugar on top? ;-) PS: The amount of work that went into this is already insane. Respect!

[@danielrfry]

I only have the breakout board which has the SD-1 clock oscillator integrated, so my initial assumption was that stereo would not be possible as the chips could not be perfectly synchronised (as they’d each have their own independent clock). But if I connect them to the same SPI bus and send data to both simultaneously (except when the channels need to differ), perhaps the timing would be close enough to work. A few people have asked so I’m going to try it when I have time.

[@hagen-p]

@@danielrfry Unless we can find a stereo SD-1 version. But I guess there is none....

[@frixyg2050]

I was surprised how many of the tracks, even when played on the OPL3, didn't seem to be making use of that chip's stereo anyway. You didn't artificially flatten them to mono for the comparison, right?

[@607]

That might be because the OPL3 doesn't allow soft panning; a sound is sent either to the left channel, to the right channel, or to both.@@frixyg2050 You can use tricks to allow soft panning (like sending the same sound to the left and the right channel separately, and varying the volume of both), but I'm sure this wasn't used in games much. Game Boy music also rarely uses stereo, probably for that reason (on Game Boy you also can rarely use the same trick, as you don't have the channels to spare).

[@DiodeMilliampere]

I am the composer of that work, thank you for the compliment! Made my day

[@bradallen8909]

Wow. I always thought the OPL2/3 chips sucked because of how many games used them so poorly. These tracker tunes really do show what they were actually capable of. Very impressive.

[@biffrapper]

Zeliard in DOS on adlib = phat tunes, some of best late 80s FM synth made.

[@bradallen8909]

@@biffrapper Many arcade games made incredible use of FM synth as well. I'm pretty sure the chips they generally used were more sophisticated than the OPL series, though.

[@biffrapper]

@@bradallen8909 They were. The common Arcade YM2203/YM2612 chips could do four operators per channel of additive synthesis versus the common OPL2 chip used in Adlib and even the original Sound Blaster Pro which only could use 2 operators per channel. Thus the arcade chips could produce richer sound. The OPL3 fixed this in later Sound Blaster Pro cards and beyond.

[@ChrisSTARS2407]

That's because most PC games makers that time back mostly relied on devices like Roland MT-32, the most advanced sound generator at that time, everyone who had $700 could afford. Pretty expensive, though. While the rest sound generators were utilized as residual principle, like "it does buzz something and enough, they'll pay more if they want higher quality sytnthesis"

[@biffrapper]

@@ChrisSTARS2407 I did not know anyone directly who had a MT-32 in the 80s. They were exceptionally rare due to their price. Same goes for the SC-55. Both devices sound brilliant, though. We made due with Adlib and later on Sound Blaster.

[@Cherrystarrscollide]

Very interesting video, thank you! It sounds like the SD1 is played back louder in this video - for future reference, make sure your comparisons are at the same level. Our ears/minds tend to always prefer the loudest source, so if you want to compare qualities of different sound sources, having the sources at same volume level is essential.

[@danielrfry]

Thanks for the feedback. I was careful to ensure the comparison was as fair as possible. All recordings were made using a Tascam DR-07X linear PCM recorder at 24-bit resolution with a fixed recording level, then normalised later in an audio editor to ensure they were equal. I think what you are hearing might be a difference in dynamic range (rendering quieter parts louder, relative to the same maximum level). This could be because I had to reduce the master volume level of the SD-1 to stop it clipping while more voices were playing, this seems to be happening in the digital domain so there is more noise too. My guess is that there is some difference in how the mixers work in the two chips.

[@danielrfry]

Having said that, I have checked the audio track from the video and it does seem to have inconsistent levels - my apologies. I do remember applying normalisation, so I’m wondering if I applied it to all of the recordings together rather than individually. Thanks for letting me know.

[@Aidan_Lawrence]

Super cool! I've always wanted one of these little YMF825 demo boards, but I haven't been able to find them anywhere. Great work putting it to good use with your custom driver!

[@danielrfry]

They’re so hard to find! I got mine from Switch Science in Japan, I couldn’t find anywhere else that sells them. I’ve only ever seen it in stock once - I’ve never hit Buy on anything so quickly before! I’d love to know what product these are used in, on Yamaha’s web site it just says “Home Appliances”. Maybe there’s a Chinese washing machine out there that plays OPL tracker music when the laundry is done…

[@Aidan_Lawrence]

@@danielrfry I’ve even asked Yamaha directly for samples and/or the ability to purchase some of the YMF825 chips and they refused, stating that they only sell to their approved vendors. So disappointing since it seems like a fun chip that they wanted “makers” to work with for a while there.

[@danielrfry]

@@Aidan_Lawrence Oh wow, good effort! Some time ago I tried contacting Adafruit using their ‘suggest a product’ form, thinking maybe they’d have the buying power or contacts to get the chip and either sell it bare or on their own breakout - I got a reply that they’d consider it, but nothing more so I guess it didn’t work out. I agree, it seems like a missed opportunity - such a unique device that’d surely be perfect for a little homebrew handheld retro console or something!

[@RWL2012]

@@danielrfry don't forget the "office equipment products"

[@wuqi3691]

你需要一个中国的朋友，可以搞定芯片。😁

[@MaximilianRehkopf]

Awesome work! It's a shame most games just used GM emulation back then and didn't really explore the OPL3's capabilities. As a result they sounded bland and tinny. The tracker tunes are an excellent showcase of what could have been.

[@stevenclark2188]

It's always amazed me the extent people get hung up on the more minor differences between FM chips. The math is still the math, the same waveforms combined the same way will get the same harmonics.

[@danielrfry]

I guess for many people, part of the fun of using vintage hardware is to replicate an experience as accurately as possible - emulation is much more convenient after all!

[@darak2]

Watch the video, they are not using the same waveforms.

[@ZILtoid1991]

Often there's a pretty good overlap between FM implementations. Even my own implementation can get pretty close to Yamaha's, except mine has been augmented with various extra functionality (more waveforms - including custom ones, more algorithms in 4 operator mode, every operator now has a feedback + there's often one for multiple operators in series, some algorithms can be switched to a "resonant" mode similar to what Casio's own take had, etc.)

[@TomAnderson7]

Hardware FM sounds more chunky. I like it better.

[@The_Mister_E]

@@ZILtoid1991 Sounds kinda interesting. Where's it implemented?

[@BuildYourOwnBass]

What I hear as the difference with the YMF825, is the timbre. The snap or transient in songs like HiTech Environment 3 or Samsara seem to be faster. Not sure if it's the extra FM Operators giving it that crunch or the Active EQ but it suits some tracks more than others. The Tracker tunes seem to bang harder at the cost of the Stereo image being narrowed. Stereo tracks still require proper mixing/programming to get those engineer smooth pans, something OPL being good at isnt always a guarantee. The Glissando on the Tracker tunes also sounds to be smoother with the YMF825, definitely a "modern" tone goal 👍

[@danielrfry]

One difference between the two chips is sample rate, and it was necessary to apply a frequency multiplier to ensure notes were played with the correct pitch on the YMF825 when given the same frequency values. So it’s possible other timing-related functions of the chip run at a different rate too, including the envelope generators - that might explain some of the differences you heard. The YMF825 sample rate is lower so I would expect A/D/R phases to be longer for the same parameters, if they are affected at all.

[@RWL2012]

I was struggling to think of a modern use for this chip that's still being made, until I read this on the datasheet: "YMF825 is a sound generator/controller/amplifier device with high quality melodies and keypad/alert tones, designed for sophisticated audio user interface on home appliances and office equipment products. Yamaha's FM synthesizer produce high fidelity clean vivid sound, in no way comparable to the traditional beeps and chimes used in such products, yet all the control required is just a few tens of bytes of setup parameters." I wonder what exact "home appliances and office equipment products" this chip is actually used in?

[@danielrfry]

There’s a tiny bit more information in this article about the release of the demo board: www-dtmstation-com.translate.goog/archives/52000321.html?_x_tr_sl=ja&_x_tr_tl=en&_x_tr_hl=en&_x_tr_pto=sc “YMF825 itself is also used in Japan, but the main is Chinese home appliances.” It’s been around since 2011 when perhaps it might have made more sense from a cost perspective. I guess now it’s probably close to the end of production.

[@Domarius64]

The DOS games don't sound any different, but the tracker music which is pushing the boundaries of what the chip can do, sounds way better on the OPL3.

[@computer_toucher]

Scream Tracker 3.21 had a pretty nice Adlib mode, I loved it :)

[@OpenGL4ever]

4:21 Worm sign or also known as Sign of the Worm from Dune 1 is one of my favorite music themes from that era. What I also really like is the music from One must fall, which was released a short time later. BTW your Yamaha YMF825 solution works great. I like it.

[@danielrfry]

Yeah that whole soundtrack is a work of genius - thanks for watching!

[@dycedargselderbrother5353]

4:16 My favorite artist, StÚphane Picq

[@danielrfry]

The dark times before Unicode.

[@illustriouschin]

All sounds extremely nice, great job!

[@SiD3WiNDR]

Just stumbled onto this - super cool 🙂It's not quite OPL3 in all cases but your middleware is amazing and clearly it works!

[@CameronMcCaffrey-e7p]

Thank you for your experimentation! It is great to hear this chip in action; it's quite powerful and I love the 'dirty' 4op sound.

[@DiodeMilliampere]

Lol cool my music is in here!

[@danielrfry]

No OPL3 related demo would be complete without it in my opinion, I’m in awe of your work - thanks for watching!

[@needfortweed8734]

Very interesting comparision. One takeaway from this video for me was that even though me and my friends did play around with tracker music, we never knew that there was an adlib tracker available. Could have made interesting things with that...

[@dwindeyer]

You could even do mixed mode with a GUS and Adlib in the same computer kzbin.info/www/bejne/Y52pf4yAbKdraJI

[@YanLipavsky]

Great job bro - it sounds identical - very little difference probably because of slightly different caps/resistors.

[@GreySectoid]

Amazing work, bravo!

[@thedrunkmonkshow]

Great video! Since both chips share equivalent waveforms they theoretically should have been 1:1 but I think if I were to guess maybe it's the DAC that's causing the slight differences. The old OPL sounds more warm and muted while the YMF825 sounds more open with treble. But to see that the YMF has 16 channels each with 4 operators with 28 waveforms means this thing can potentially make some incredibly complex sounds...almost to the point of blurring the lines between traditional sinewave synthesis and the kind of results you hear back with wavetable synthesis. I guess I'll need to look for videos of music composed specifically for the YMF. 😇

[@cloud_s_97]

And the question is: Is it compatible with the YM2151?

[@thedrunkmonkshow]

@@cloud_s_97That's a really good question and I would say overall, no because the YM2151 is based on OPM and derives it's sounds differently than the YMF which is based in OPL2/3. Now, it's not impossible if someone had the knowledge and patience could re-engineer copycat sounds in the YMF using only sine since the OPM is sine wave only. It's an interesting rabbit hole and thought experiment to go down but that's my best theory to a fantastic question. 😀

[@cloud_s_97]

@@thedrunkmonkshow I initially thought that it was compatible with the OPN/OPM family of chips and it turns out that it's compatible with the OPL family...

[@DOSStorm]

This is cool as balls! Great work.

[@danielrfry]

Thanks! Glad it was of interest.

[@dominikschutz6300]

Wow, amazing, a kickstarter would be worth trying :)

[@chilly058]

I like the sound of the DS1 even better, high pitch tones are clearer, sometimes OPL3 is almost a bit scratchy

[@mesterak]

That’s amazing work!

[@jimbotron70]

It would be interesting to model the OPL on a FPGA and compare the results.

[@kveemusic]

Oh damn, that's pretty close!

[@danielrfry]

Yeah, I was surprised. I think the biggest difference I can hear is in dynamic range, especially in the Dune soundtrack in this video. Otherwise I find it hard to tell apart from the OPL3...

[@randy7894]

FM. Hell yes.

[@thetechdudemc]

The8BitGuy is having issues procuring Yamaha YM2151 chips for the Commander X16, I wonder if this chip could be a possible replacement since it's still made

[@customsongmaker]

He would have preferred dual SID chips

[@GigAHerZ64]

No Doom E1M1? :o

[@danielrfry]

🤦‍♂️ What was I thinking?

[@OpenGL4ever]

He doesn't need DOOM, he has worm sign in his video.

[@enginerdy]

Ah yeah we needed to hear what chainsaw farts (lead guitar patch) sounds like on this chip lol

[@aster.ækes]

The sound of the newer one makes it so much more immersive, like how the devs intended their game to be. But perhaps a little late.

[@ValTek_Armory]

Now we just need to get these slapped onto an ISA card and programmed to work with some sort of OPL translation layer

[@Ojref1]

Being mono crushes it in the first round. Haven't looked at the spec sheet but I'm assuming it can work with another SD-1 to provide stereo?

[@danielrfry]

It should be possible to run two SD-1 chips from the same clock source, as has been pointed out in another comment. I think later OPL2 cards used to do something similar. I only have the pre-built breakout board though, which has the clock oscillator integrated.

[@AK-vx4dy]

Great job!

[@gammaboost]

To me it sounds like there's less audio filtering. Perhaps you need to put some find of low pass filter?

[@ichemnutcracker]

Honestly, I do not miss FM synthesis in even the slightest bit. The very first upgrade I bought for my 486 PC was a wavetable daughterboard for my SB16. I couldn't stand the sound of wet farts then, and I don't feel any nostalgia for them now. This certainly sounds like a great chip for people who do miss, it, though.

[@danielrfry]

I also found it disappointing at the time and bought an AWE64 as soon as I could afford it. I wouldn’t say I miss it, but now I don’t *have* to use it, it’s kind of fun to see what can be done by pushing these chips to their limits - not much software made full use of the OPL3 back when it was current.

[@OpenGL4ever]

It just depends on whether you want to recreate real natural instruments or really need electronic synthesizer music. The OPL2 and OPL3 chips can definitely be used for the latter.

[@ichemnutcracker]

@@OpenGL4ever It's not that I don't like synthesized music. I love the John Carpenter soundtracks and a lot of the synthesized music from the 70's up until ~'83-'84 (whenever the DX-7 came out and ruined music for a decade+). I am specifically saying that I hate **FM** synthesis. To me, it just sounds... bad. It reminds me of that terrible 'muzak' knock-off smooth-jazz cover music you'd hear in the waiting room to the dentist office (and, yes, I'm saying even the Pantera covers in the Doom soundtrack remind me of the dentist's office). If it were subtractive synthesis on like a Prophet-5 analogue synth, that would have been awesome! But it wasn't. FM is low-rent Kenny G backup band with a few wet fart noises thrown in. I'd rather just have square, triangle, and sawtooth waves at the point.

[@OpenGL4ever]

@@ichemnutcracker Okay I understand. But in the end, good FM in theory only depends on how many operators and functions you have available. From a purely mathematical point of view and if you chain them together, any sound should be able to be reproduced. So it's more just the limitation of the OPL 2 and OPL 3 chips that sounds bad and if you don't like the DX-7, this one too. And of course what is also required is a good understanding or luck on the part of the person who wants to recreate a certain sound with it.

[@ichemnutcracker]

@@OpenGL4ever No argument there. But, when it comes to amateur or less talented composers just trying to bang out a soundtrack on limited hardware just so they can eat for a few months, I enjoy what they put out better when they are given access to sample-based audio more than crossing my fingers and hoping they come up with something so amazing at not only composition but also programming their own patches, that it overcomes the limitations of OPL/N/M/Z or whatever.

[@Adam_Lyskawa]

It seems to sound slightly better than OPL3, despite it shouldn't - because I'd expect it to sound different only on songs written specifically for it. TIL that you could get a pretty decent sounding music on OPL-3, but for a weird reason most of the PC game music for OPL-3 sucked being way, way worse than your average C64 music.

[@MWTD]

Cool vid

[@shoopdawhoop]

I want to see the comparison with Wacky Wheels by Apogee because they've used distinctly sounding cymbal patches with high-pitched modulation that can sound dramatically different if other DAC sampling rate is used because of aliasing.

[@konradb1645]

Dan, when you have spare minute, would you be able to try "Fly In Spider's Web!" by JO/Vibrants and "Il Numitare Et Sulfo" by Metal/Vibrants (both are Edlib zaks) ? I am curious how SD-1 would play them. Maybe I would need to improve my OPL to MA-3 converter ;)

[@danielrfry]

Here you go: kzbin.info/www/bejne/kIWnh5Vmm8iifLs I hope you find it interesting!

[@Fine_Mouche]

if i'm not wrong the YMF825 is mono, so is this a way we can do stereo with it ? (little DIY / hacking / Tricks ?)

[@Alexsey_Kyznecov]

1:19 Where did you get this documentation? Is there a similar one for the sega genesis chip?

[@danielrfry]

Yamaha released a GitHub repository containing documentation and sample code. They have since removed it, but mirrors can be found by searching GitHub for “ymf825board”. The YM2612 datasheet was never made publicly available, but I remember seeing a scan of the Japanese language version uploaded at some point - you might try searching for “YM2612 Application Manual”.

[@RustLighter]

well its definetly better than pc-speaker :D

[@danielrfry]

Yes. At the time I generally considered total silence to be better than PC speaker.

[@audguy]

OMG I had that exact case!

[@GoldenPickaxe]

A shame that this likely will get discontinued eventually because it would be so cool if cheap retro PCs and keyboards sprung up using this thing.

[@calvinh4374]

Very cool stuff! Just curious, what are your specs on the PC you're using?

[@danielrfry]

Thanks! It has a 133MHz Pentium CPU, 24MB RAM and 1GB HDD (and the Vibra 16S sound card shown at the start, of course).

[@logipilot]

❤Samsara❤

[@arnetes]

can you test the APL-5?

[@marisakirisame867]

I think it was better if OPL3 was also manufactured with DIP interface

[@mrwensveen]

Wow, amazing. That must have been a lot of work. It's hard to hear but how does the stereo output of the OPL3 compare to the mono output of the SD-1?

[@petroidau]

Whaaaat no doom ^⁠_⁠^

[@danielrfry]

I will fix this...

[@607]

That's really cool! It's a shame it doesn't have stereo, though.

[@edbrito-swdev]

I wish I could have a simple synth like module with this chip I could connect a midi keyboard and play around with it

[@TonimanGalvez]

Just get a Commodore Amiga, got better sound at that time than most PC sound cards.

[@mypkamax]

The track "High Tech Environment III" reminds me of _Mega Man_ music a bit.

[@danielrfry]

Yes, it does have something of that Japanese arcade sound to it I think!

[@Voidsworn]

I mean, if it only has mono, couldn't it be paired with a second one? Of course it would need the hardware (and likely the drivers) be done but still.

[@danielrfry]

Maybe. If there’s even a very small variance in timing, such as a delay sending data to one of the two, sounds intended for both channels might play very slightly out of sync, introducing an unintended sort of chorus effect. I don’t know for sure though, and clearly it’s been done successfully with other chips.

[@nickwallette6201]

@@danielrfry Don't forget -- there's precedent for this approach with the first-gen SB Pro and Pro Audio Spectrum cards. Those used a pair of OPL2 ICs for stereo FM. Similarly, if you're able to clock them from the same master clock source, you might be able to get clever with chip-select pins to control them both at once. Panning or variance commands would be sent to one or the other, just like the dual-OPL2 boards. The tiny timing discrepancies for those changes should be minimal, while all the note-on events would be sent to both at the same time, preventing any phasing issues. In theory. :-)

[@danielrfry]

@@nickwallette6201 Ah yes, running both chips on the same clock sounds like a neat way of doing it. I guess I was thinking about what might be possible using these boards that have the clock oscillator integrated - having the bare chip available separately would certainly open up more possibilities like this.

[@in_the_pines]

What kind of computer speakers are those exactly? I tried searching online everywhere but could find nothing. Also are there any modern day soundblaster cards or audio cards in general that can actually work on modern hardware?

[@SianaGearz]

These look vaguely like clones of Roland PC speaker just made to look vaguely different enough. Haven't seen these particular ones back in the day though, so can't say too much more. Soundcards which work in which regard? Sure you can buy them, you can buy CMI8738 for a couple dollars, you'll receive a generic audio in out device, it still just works, but the internal FM synthesizer and Soundblaster emulation logic that is physically present won't be accessible. Actually maybe i should buy one of those to experiment with. There's apparently a large new old stock or maybe just landfill refresh, anyway the world isn't exactly running out of them, it's a chip from around year 1999 fundamentally. A couple dollars extra gets you a version with a PCIe to PCI bridge chip on board, so it runs in modern mainboards. Internal PC soundcards currently in production include ASUS Xonar (SE, DX) and Sound Blaster (AE, Z, Audigy Rx, Fx). There's many external ones as well. They don't have any interesting hardware features to speak of but promise better quality than onboard PC audio. The ASUS branded chip on the Xonar is actually just a C-Media CMI8788 which is from around 2005 and they have a PLX chip as a PCI-e to PCI adapter. The onboard DAC of the CMI isn't used, instead it's driving i2s devices which are also anything but fresh but usually they pick fairly nice ones. Sound Blaster Audigy Fx is a custom controller that drives one of those usual and very cheap mainboard HD Audio soundchips, with some careful external circuit design and an output amplifier MAX97220 which helps it regain decent performance when driving headphones, which is normally an issue otherwise. I haven't really looked in depth at other soundcards. Musicians use an external audio interface, those are abundant from specialised sorts of audio companies. Focusrite, Presonus, MOTU, Steinberg, Behringer, ESI, TASCAM. These have microphone preamps for professional XLR microphones and MIDI ports to connect external synths, as well as some signal level LEDs and headphone monitoring of the input signal. They also have ASIO compatibility for professional software.

[@mRahman92]

So where does one source the chip?

[@temet_nosce]

why no flashback

[@arnetes]

what is the song at 0:59?

[@danielrfry]

“Gorbi” from the game KGB. The full track list is in the video description.

[@arnetes]

@@danielrfry oh ok thanks

[@saricubra2867]

OPL3 is better because of 4 audio outputs.

[@cooliofoolio]

I think the OPL is better simply because it is stereo

[@SeanKerrygames]

so megadrive-y

[@SamPinchesMakerShop]

Do you know if it is possible to buy this chip anywhere?

[@danielrfry]

The breakout board I’m using in the video has been discontinued, and I don’t know of anywhere that sells the bare chip unfortunately.

[@SamPinchesMakerShop]

@@danielrfry thanks Dan. Wonder if it got permanently canned, or just affected by shortages. 😔 It didn’t have a long run!

[@danielrfry]

@@SamPinchesMakerShop The board was launched in 2017, the chip in 2011 I believe, so it was around a little while. Yamaha deleted their GitHub repository with the documentation and sample code, so I’m pretty sure that’s the end of it. As far as I know the chip itself was never widely available outside China which is apparently its main market.

[@SamPinchesMakerShop]

@@danielrfry Ahhh, that's such a shame. I reckon it for sure would have been a hit for makers -- could easily have ended up in things like the Commander X16 project! Do you know what the retail price of the breakout board was? Thanks for a great video showcase too! :)

[@danielrfry]

@@SamPinchesMakerShop I imported mine from Switch Science in Japan for $30 USD and shipping was nearly as much again, so it wasn't super cheap. The board was indeed intended for makers so they must have seen some potential there, and it seems like it maybe sold OK in Japan - I think it was just not as well known elsewhere.

[@92trdman]

What speaker in use ?

[@danielrfry]

Just some generic ones from a late 90s PC bundle. You don’t hear them in this video though, the audio was recorded directly from the output using a Tascam PCM recorder.

[@videooblivion]

No offense, Dan, but ... they both sound terrible to an ear that was primed by SID and Paula. ;)

[@danielrfry]

None taken, no one is presenting this as the pinnacle of synth technology, that’s not the point - it’s a sound chip designed for embedding into washing machines.

[@Galova]

LOVE

[@Synthematix]

This chip was in the sega megadrive

[@danielrfry]

Not quite, that was a different Yamaha FM chip: the OPN2 (YM2612) or in later models, the OPN2C (YM3438).

[@RWL2012]

My YM2612 and YM3438 still *are* in my Mega Drives 😝

[@NotAGeySer]

Huh

[@boardernut]

For me OPL3 always sounded like garbage and el_cheapo-soundcard, compared to wavetable synthesis

[@jimbotron70]

"Like garbage" = FM synthesis... the Yamaha DX7 did use it.

[@boardernut]

@@jimbotron70 OPL3 = 3 miserable oscilators per voice, to a maximum of 6 voices simultaneously, compare that to a DX 7, by the way have you ever had a DX 7?, because I have and I know it very well.

[@jimbotron70]

@@boardernut The underlying technology is essentially the same, the OPL is a scaled-down DX7's core on-a-chip.

[@boardernut]

@@jimbotron70 Have you ever connected a midi keyboard to a FM syntesis sound card? or have you ever tried to produce music with Cubase or Cakewalk using a SoundBlaster 16 or any SB clone for that matter ? it really is a shitty as it can get, in particular the percussion instruments, if you haven't please do, and then try something like the Yamaha DB50, or Roland SC.

[@jimbotron70]

@@boardernut As Jobs would have said, it's not a bug, it's a feature 😀 FM synthesis sounded like that, it's part of the charm, you can't compare it to later, more advanced synthesis techniques.

[@elmariachi5133]

Sounds as terrible as the orginal xD

[@marcsmithsonian9773]

Crap von crap.