Scali wrote:

However, I would like to know the other side of the story as well... What kind of deal did AdLib and Yamaha have, and how was Microsoft able to override that?

I think Beegle would need to add Martin Prével, Rich Heimlich, and/or Karen Collins to his list of interviewees to get definitive answers.

Whatever original arrangement Ad Lib had with Yamaha regarding the OPL2 was likely tenuous at best, given the obfuscation of the chips.

An MIT document I'd linked to several months ago describes the Ad Lib/Yamaha interaction as follows:

The Sound of One Chip Clapping: Yamaha and FM Synthesis wrote:

The chip in question was a low-priced version of FM synthesis that Yamaha had designed to provide audio for an ill-fated (Japanese government sponsored) teletext terminal project. The chip was not on the market, and when As Lib (sic) first contacted Yamaha, the Japanese company was very nervous about the idea of selling it. The reason for this reluctance was that Yamaha was concerned that FM chips might end up in cheap knock-off keyboards made by Korean or other East Asian firms. But eventually Prevel presuaded Yamaha to supply him with the chips, and Ad Lib's FM-based sound board made its debut at the Chicago Consumer Electronics Show in June 1987.

Microsoft isn't contextually mentioned in that document, suggesting that any involvement they may have had was either not widely publicized, or was of lesser importance than the "Game Sound" reference would seem to suggest, to the extent that they may have simply asked/persuaded Yamaha to add the YM3812 to their chip catalog or somesuch. Another possibility is that Microsoft's stated involvement is complete supposition, erroneously presented as fact.

While I like to err on the side of believability, where evidence and plausibility is concerned, I will say that the statement in Karen Collins' book bothered me when I first read it a few years back, and bothers me yet still, for several reasons:

• I've encountered zero additional references relating to Microsoft's involvement.
• And, perhaps most importantly, Karen Collins' book describes Roland's MT-32 as having been "designed as a MIDI soundcard," thereby casting doubt upon the entire body of work. 😀

Cloudschatze wrote:

The Sound of One Chip Clapping: Yamaha and FM Synthesis wrote:

The chip in question was a low-priced version of FM synthesis that Yamaha had designed to provide audio for an ill-fated (Japanese government sponsored) teletext terminal project. The chip was not on the market, and when As Lib (sic) first contact Yamaha, the Japanese company was very nervous about the idea of selling it. The reason for this reluctance was that Yamaha was concerned that FM chips might end up in cheap knock-off keyboards made by Korean or other East Asian firms. But eventually Prevel presuaded Yamaha to supply him with the chips, and Ad Lib's FM-based sound board made its debut at the Chicago Consumer Electronics Show in June 1987.

Microsoft isn't contextually mentioned in that document, suggesting that any involvement they may have had was either not widely publicized, or was of lesser importance than the "Game Sound" reference would seem to suggest, to the extent that they may have simply asked/persuaded Yamaha to add the YM3812 to their chip catalog or somesuch. Another possibility is that Microsoft's stated involvement is complete supposition, erroneously presented as fact.

Well, I am not entirely sure about this story either.
I would like to have some timeline.
Namely, according to this: https://www.youtube.com/watch?v=QLJSdNYcdpk
The YM3812 has been around since 1986 at least (the PSR-11 was launched in 1986, and powered by an YM3812 chip).
Various sources, such as this and this claim the chip was released in 1985.
So I am not sure when the teletext terminal project was around... was that before 1986 even, and did Yamaha just decide to 'recycle' the chip in a synthesizer line? Doesn't sound very plausible.
And how long had AdLib been asking for these YM3812 chips until Yamaha started selling them? The AdLib was launched in 1987, so it didn't take Yamaha long to be convinced then...

I think it's more likely that Yamaha always wanted to build synthesizers with the YM3812, and was well aware of the potential to apply the chips in home computers, consoles, arcade machines and such as well.
In fact, IBM's Music Feature card is from 1987 as well, and used a slightly different Yamaha chip, the YM2164, another chip that was also used in Yamaha's synthesizers.

If I were in Yamaha's place, the only deal I'd make with companies that want to buy YM3812s is that they don't use them to build those cheap knock-off keyboards.
That is, unless Yamaha was considering to go into the PC soundcard business themselves, in which case it would make sense not to sell them to board makers, but only to companies that build entire machines (home/personal computers, consoles, arcade machines), so that they wouldn't compete with Yamaha's sound cards.

Scali wrote:

Cloudschatze wrote:

The Sound of One Chip Clapping: Yamaha and FM Synthesis wrote:

The chip in question was a low-priced version of FM synthesis that Yamaha had designed to provide audio for an ill-fated (Japanese government sponsored) teletext terminal project. The chip was not on the market, and when As Lib (sic) first contacted Yamaha, the Japanese company was very nervous about the idea of selling it. The reason for this reluctance was that Yamaha was concerned that FM chips might end up in cheap knock-off keyboards made by Korean or other East Asian firms. But eventually Prevel presuaded Yamaha to supply him with the chips, and Ad Lib's FM-based sound board made its debut at the Chicago Consumer Electronics Show in June 1987.

Well, I am not entirely sure about this story either.

The sources for that publication include interviews with Yamaha executives and Martin Prével of Ad Lib. As far as known information goes, I don't know if it gets much more authoritative than that.

I would like to have some timeline.

Well, here's my poor stab at it, concerning the PC side of things:

• 1984/5(?) - Yamaha develops the YM3526, "OPL" (FM Operator Type-L) sound generator for Japanese CAPTAIN/teletext systems. (1)
• 1985 - Yamaha develops the incremental-upgrade YM3812, "OPLII" (FM Operator Type-LII) sound generator for "computer apparatus, teletext instruments, etc." (2), (3)
• 1986(?) - Martin Prével persuades Yamaha to supply Ad Lib with YM3812 chips (4), (5)
• 1987, August - Ad Lib releases the Music Synthesizer Card (6)
• 1989 - Microsoft purportedly convinces Yamaha to sell the YM3812 on the open market. (7)
• 1989, November - Creative Labs releases the Sound Blaster (8)

Sources:
1. YM3526 datasheet
2. YM3812 datasheet
3. YM3812 die surface photograph
4. Johnstone, Robert. The Sound of One Chip Clapping: Yamaha and FM Synthesis. MIT JAPAN PROGRAM, 1994
5. Based on the dates of some of the earliest Ad Lib instrument files, Prével would likely have obtained access to the OPL2 sometime in 1986.
6. Ridge, Peter, et al. Sound Blaster: The Official Book, Second Edition. Osborne McGraw Hill, 1994.
7. Collins, Karen. Game Sound: An Introduction to the History, Theory, and Practice of Video Game Music and Sound Design. The MIT Press, 2008.
8. Creative Corporate Milestones

Use of the YM3812 prior to 1989 does appear to have been limited to Yamaha, Ad Lib, and a few arcade games, which gives some credence to the Microsoft intervention claim. As an update, Great Hierophant provided me with the citation for the reference in the full "Game Sound" book, which apparently tracks back to a defunct presentation by Lyndsay Williams, and is further described in a defunct archive of Williams' Microsoft research page, albeit, with more vaguery:

Lyndsay Williams wrote:

Reportedly Microsoft approached Yamaha and asked them to make their chips available on the open market which then assisted Creative Labs and their Soundblaster card (1988) to become an industry standard for audio for games and multimedia computing.

Reportedly? Oh, well. If nothing else, there's an interesting reference to a YM3526-based soundcard created for Olivetti's PC1 computer.

So I am not sure when the teletext terminal project was around... was that before 1986 even, and did Yamaha just decide to 'recycle' the chip in a synthesizer line? Doesn't sound very plausible.

Looks like the CAPTAIN system was commercially launched in 1983. Regarding the second point, and from what I understand, Yamaha's semiconductor and musical instrument divisions were fairly separate. If one assumes that the YM3812 was simply a development continuation for some failed teletext system or accessory, I can imagine Yamaha may have wanted to capitalize on the investment by putting the chip to use elsewhere internally, as per the keyboard examples, rather than market it openly (for the reasons described in the MIT publication).

I am far from persuaded of the truth of Lyndsay Williams' story that Microsoft intervened with Yamaha to make the YM3812 chip widely available. The idea that Ad Lib Inc. had the ability to enter into an exclusive supplier relationship with Yamaha is absurd. In 1986-87, Yamaha was a huge multinational corporation, Ad Lib, by comparison was not much more than a garage-startup. The boys from Quebec had no financial authority to obtain an exclusivity arrangement. If they did secure a deal to be the exclusive buyer of this chip, why bother to sand down the chip markings for their first years of manufacture?

Haunted Castle, Konami's rather frustrating conception of Castlevania for the arcades, used a YM3812 sound chip. Not many arcade machines used the YM3812, but this 1988 game shows that Ad Lib had no deal whereby Yamaha would see this chip only to Ad Lib.

A PCB designer could have determined how the Adlib card worked in a day with the proper equipment. Even without knowing the identity of the chips, he knows what they do. He knows that they use FM synthesis and who had a patent on FM synthesis? Why Yamaha of course. A perusal of the datasheets, a request for a few engineering samples and the chip's true identity would soon be discovered. I would suggest that by mid 1989, the chip Ad Lib used would have been an open secret in the industry.

While Microsoft was interested in standards, they announced the first MPC standard in November, 1990. They indicated that the Sound Blaster was the closest sound card to the standard available on the market at the time : http://www.atarimagazines.com/compute/issue14 … ster_turns_.php While the Ad Lib card was successful by 1989, I think it would have been too forward thinking on Microsoft's behalf to have considered it as the core of its MPC standard at that time.

Great Hierophant wrote:

The idea that Ad Lib Inc. had the ability to enter into an exclusive supplier relationship with Yamaha is absurd.

And yet, some kind of closed-market arrangement or agreement precluded the direct purchasing of the YM3812 by anyone other than Ad Lib and a few Japanese arcade system developers until 1989.

While the Ad Lib card was successful by 1989, I think it would have been too forward thinking on Microsoft's behalf to have considered it as the core of its MPC standard at that time.

The Multimedia PC Marketing Council was established in 1989, meaning it wouldn't have been all that forward thinking, and it's easy to see how the genericness of the Ad Lib card, and the industry support it had garnered by that time, would have made Yamaha's inexpensive OPL2 very appealing to a standards committee.

Cloudschatze wrote:

Great Hierophant wrote:

The idea that Ad Lib Inc. had the ability to enter into an exclusive supplier relationship with Yamaha is absurd.

And yet, some kind of closed-market arrangement or agreement precluded the direct purchasing of the YM3812 by anyone other than Ad Lib and a few Japanese arcade system developers until 1989.

Yamaha offered a wide range of FM chips, and whether one company chose a YM3812 or a YM2143 or a YM2151 would have depended on the company's needs and the price point it wanted to hit. Some companies would have had the market clout to ensure that the FM chips it used would not be available to third party clone makers. Thus you only tend to see the YM2614 in IBM and Korg products. The YM2612 is not often to be found outside Sega and Fujitsu. The YM2203 and 2608 were primarily the domain of NEC.

But other chips were widely available. Arcade manufacturers needed something very capable and usually chose the 4op, 8-channel YM2151. The YM2149 was a drop-in replacement for an AY-3-8910. The YM2143 was not exclusive to Sega's Japanese 8-bit consoles. Other OPL family chips like the YM3526 and Y8950 were also found in products by different companies. The cheaper chips tended to be more diversely used. But there was only so much demand for FM Synthesis chips. Ad Lib wanted something cheap yet fully programmable with several simultaneous voices and better sound than the NES, so the YM3812 was probably their cheapest option for what they wanted to accomplish. When other people wanted to accomplish that, it doesn't seem like Yamaha did anything to discourage them.

On the other hand, I do recall the story about how W.H. Sim, the President of Creative Labs, was able to very accurately predict when the Adlib Gold would pass Yamaha's QA. I believe that the board did not pass QA, despite at least a year's worth of Yamaha development, until after CL shipped the SB16. Yamaha probably had a greater degree of partner loyalty than another chip supplier, but absent an exclusivity agreement, those loyalties could shift as the market dictated.

Great Hierophant wrote:

I do recall the story about how W.H. Sim, the President of Creative Labs, was able to very accurately predict when the Adlib Gold would pass Yamaha's QA.

Yes that was said by Rich Heimlich if my memory is correct.

Great Hierophant wrote:

When other people wanted to accomplish that, it doesn't seem like Yamaha did anything to discourage them.

Not openly marketing the YM3812 until 1989, as the evidence suggests, would effectively discourage selecting it as an option. I mentioned it earlier, but perhaps that's really all it was though - you simply had to "know" about the OPL2 in order to even inquire about it. By obfuscating the labels, Ad Lib would have been doing their part to ensure that the exact chip itself remained something of a mystery, despite Yamaha being the obvious source.

Isn't speculation wonderful?

On the other hand, I do recall the story about how W.H. Sim, the President of Creative Labs, was able to very accurately predict when the Adlib Gold would pass Yamaha's QA. I believe that the board did not pass QA, despite at least a year's worth of Yamaha development, until after CL shipped the SB16.

Not to get into the weeds, but the card we know as the Sound Blaster 16 shipped sometime around December, 1992. Yamaha's MMA chipset passed QA nearly a year earlier, suggesting that Creative's "go-ahead" would have instead coincided either with an early announcement of the SB16, the finalization of the SB16 chipset, or perhaps even an earlier incarnation of the card (there are references to a Sound Blaster Pro 16, even as far as the Tandy multimedia board is concerned). I tried following-up with Rich on this discrepancy, and didn't get a response.

Cloudschatze wrote:

[*]1984/5(?) - Yamaha develops the YM3526, "OPL" (FM Operator Type-L) sound generator for Japanese CAPTAIN/teletext systems. (1)
[*]1985 - Yamaha develops the incremental-upgrade YM3812, "OPLII" (FM Operator Type-LII) sound generator for "computer apparatus, teletext instruments, etc." (2), (3)

See, that makes more sense to me already: the YM3812 wasn't designed specifically for the CAPTAIN.
The YM3526 may have been, but it's all a matter of interpretation I suppose. Yamaha develops FM synthesizer technology, so even the YM3526 is just a custom design based on technology they had developed earlier.
Likewise, you could see the YM3812 as an incremental upgrade of the YM3526, but on the other hand, it's just another FM-based product in their product line.

I mean, even the legendary SID chip used by the Commodore 64 wasn't actually developed specifically for that machine. It was a cut-down version of a synthesizer design that originally had 32 voices. The limited technology would only allow 3 voices in a single chip at the time (and some other features of the design also had to be dropped). So the designer, Bob Yannes, thought you should just use multiple SID chips to build the full synthesizer.
But then the C64 was being designed, and they used a single SID as its sound chip, because it just was a great fit. The rest is history. The full SID synthesizer was never completed.

[*]1986(?) - Martin Prével persuades Yamaha to supply Ad Lib with YM3812 chips (4), (5)

This point...
As for Yamaha not 'openly marketing' the YM3812... Well, isn't that part of negotiations? I mean, if I were to build a PC sound card in the mid-80s, I'd approach suppliers of synthesizer chips to see what my options are. Yamaha would be one of the candidates (as would be Roland, and perhaps others).
Then I'd picture it like this: I'd arrange a meeting with them, I'd explain what kind of sound card I want to develop, what my price range will be, and my projected audience... Then I'd expect them to come up with a list of chips that would fit this template (or perhaps offer me a custom design to suit my needs if they don't have anything on the shelf already).
So Yamaha would probably see the business opportunity and say: "Hey, we've just developed this new YM3812 chip, it might be exactly what you're looking for. We've just made the first synthesizer based on this new chip, here! (shows PSR-11)".

So on the one hand I don't see why the AdLib company would have to have had specific knowledge of the YM3812 in the first place. And I don't see why Yamaha would need any convincing to sell their products.

In this light, it also makes sense that IBM's Music Feature card uses a slightly different Yamaha chip.
They probably didn't use that different chip because IBM didn't know about the YM3812, or because Yamaha didn't want to sell them that chip. But probably because it was a slightly more high-end chip, and IBM went to Yamaha with slightly different requirements (more for the semi-professional musician, in a higher pricerange than the AdLib). As a result, Yamaha and IBM agreed on using a different chip.

Though the music was "durchwachsen", I liked the venyl-player program. It lookes very nice. (LGR showed it an an adlib video)

Cloudschatze wrote:

Well, here's my poor stab at it, concerning the PC side of things: […]
Show full quote

Scali wrote:

I would like to have some timeline.

Well, here's my poor stab at it, concerning the PC side of things:

• 1984/5(?) - Yamaha develops the YM3526, "OPL" (FM Operator Type-L) sound generator for Japanese CAPTAIN/teletext systems. (1)
• 1985 - Yamaha develops the incremental-upgrade YM3812, "OPLII" (FM Operator Type-LII) sound generator for "computer apparatus, teletext instruments, etc." (2), (3)
• 1986(?) - Martin Prével persuades Yamaha to supply Ad Lib with YM3812 chips (4), (5)
• 1987, August - Ad Lib releases the Music Synthesizer Card (6)
• 1989 - Microsoft purportedly convinces Yamaha to sell the YM3812 on the open market. (7)
• 1989, November - Creative Labs releases the Sound Blaster (8)

I would add this to the timeline :

1988, September - Sierra Online releases King's Quest IV, the first PC game to support the Ad Lib.

Never realised Chalifour lived across the street 🤣 Here, adlib are cheap, I've seen an adlib gold for 200$ CAN a month ago on the local adds. The HQ of adlib is now an quite luxurious apartment building.

I often wonder if Prevel hand crafted the 1987 edition of the cards with off the shelf parts there or just made a prototype and offloaded the mass production somewhere else? This is just a silly question that bug my mind time to time, must be the answer being obvious to someone else.

All in all, I'm eager to read your interview Beegle, I have seen the youtube presentation of your impressive soundcard collection, mine is far from being that amazing but the main reason I maintain vintage computers is to drive old soundcards so I can listen to hardware chip music and watch demo.

The competition was stiff between 1988 to 1993, every year was a new revolution in ISA sound cards mainly by Creative.

It would be interesting to know what strategy Adlib planned to compete with the Killer Card.
The timeline shows no new products by Adlib from 1987 to January 1992 when Adlib Gold showed up.

For 5 years Adlib did what?
Meanwhile an army of Singaporean engineers innovated every couple of months.

Scali wrote:

And I don't see why Yamaha would need any convincing to sell their products.

From my perspective, the account in the MIT publication, the Microsoft intervention reference, and the circumstantial evidence otherwise all support the idea that there was some initial reluctance on Yamaha's part to sell the YM3812. Given the lack of information, the scenarios you described could be just as likely though.

I don't know that the folks Beegle has lined-up to interview will have any additional insight at this level, but I have to say that I've very-much appreciated the discussion here, all the same.

Scali wrote:

In this light, it also makes sense that IBM's Music Feature card uses a slightly different Yamaha chip.

The IBM Music Feature Card doesn't require nearly as much introspection - it's a complete Yamaha FB-01 on a card, with additional PC interfacing - which is presumably exactly what IBM had asked to be produced. Use of one of the lower-cost FM chips would have resulted in a completely different product.

James-F wrote:

It would be interesting to know what strategy Adlib planned to compete with the Killer Card.
The timeline shows no new products by Adlib from 1987 to January 1992 when Adlib Gold showed up.

The original release date estimate for the Ad Lib Gold 1000 was July, 1991, which would suggest that the initial development probably began sometime in 1990, if not sooner. While there is some supposition that a small quantity of Gold cards shipped just before Ad Lib Inc. folded on May 1, 1992, it's more commonly known that cards shipped in quantity after the asset acquisition and rebranding, in the August/September 1992 timeframe.

James-F wrote:

The competition was stiff between 1988 to 1993, every year was a new revolution in ISA sound cards mainly by Creative.

Indeed, and what's worse... The things Creative did were pretty obvious really.
I mean, the first Sound Blaster was little more than an AdLib with a game port and ADC/DAC added.
Surely the AdLib guys could have figured out that gamers would like those features? I mean, there were already games trying to play digital sound effects on PC speaker, and the Covox was already on the market.
As for MIDI... well, could have figured that one out as well. Adding a MIDI uart is a cheap addition, so although it's not something gamers would use, you can add value for semi-professional users that way, so why not include it?

Once you have a Sound Blaster, moving to stereo with either two OPL2s or with the newer OPL3 would be quite obvious as well. Again, there were various other computers and sound cards that were already stereo, so quite an obvious step forward.

Likewise, moving to a full CD-quality 16-bit ADC/DAC was not exactly rocket-science.

You could say that Creative just picked up all the low-hanging fruit that AdLib ignored.

Also in 1987~1992 computer science in Quebec/Canada was very low profile. Very few people cares about computers and I have no difficulty to believe it when I read that adlib inc. had to fight government instead of being supported by them. It is known that the founder was a music teacher at local university so he could not cope with an army of Singaporean engineers.

ElBrunzy wrote:

It is known that the founder was a music teacher at local university so he could not cope with an army of Singaporean engineers.

No, but he was on the market a few years earlier, and the AdLib was a reasonable success. He could (and should) have hired some engineers and marketing people who could cope with competition from Creative and others.

Scali wrote:

I mean, even the legendary SID chip used by the Commodore 64 wasn't actually developed specifically for that machine. It was a cut-down version of a synthesizer design that originally had 32 voices. The limited technology would only allow 3 voices in a single chip at the time (and some other features of the design also had to be dropped). So the designer, Bob Yannes, thought you should just use multiple SID chips to build the full synthesizer.
But then the C64 was being designed, and they used a single SID as its sound chip, because it just was a great fit. The rest is history. The full SID synthesizer was never completed.

Yannes did eventually get around to developing that full 32 oscillator chip, otherwise known as the 5503DOC that powers the Ensoniq Mirage, ESQ-1, SQ-80, and Apple IIgs.

Wow, I think the most interesting subjects to hear about would be to do with the very beginning of their idea and birth of add-in PC audio, then their interactions with Yamaha, and finally Creative.

ElBrunzy wrote:

All in all, I'm eager to read your interview Beegle

Ho ho ho, you're in for a surprise my friend 😀

Hello people!
The interview was conducted last Saturday, and went smoothly!
Since I don't feel like typing a detailed story, here are some statistics about it :

Hardware --
Number of cameras : 3
Types of cameras : 2xCanon T2i, 1xNikon d7000
SD cards used : 4x64gb, 2x32gb
Number of audio recorders : 2
Number of AdLib Golds present : 3

Environment --
Room had windows : TRUE
Snowing outside window : TRUE
Quebec city ambience : ACHIEVED

Food --
Water bottles used : 6
Pain au chocolat consumed : 6
Croissants consumed : 4

Time frame --
The interview lasted from 10 AM to 2PM approximately
Yes, that's 4 hours of material overall. I expect the final result will be cut and polished to about 1.5 hour.

People --
AdLib crew : Henri Chalifour (composer, developer support), Vincent Dallaire (technician, card prototyping), Raymond Skilling (composer, documentation).
Film crew : An assistant for the 3rd camera, and myself.

The interview --
Number of questions : 64
Number of fan questions : 18

Data --
Size of unedited data on disk, video and audio : 194GB

Editing process --
I'm used to making 10~15 minute videos... This one will take a WHILE to edit so we all have to be super patient 😀 including myself!

In the next few weeks/months I'll have a few surprises to show you.
But since the editing process is very time consuming, I don't want to give everything out too quickly at once, and then have you wait for X months without communication.
My first communication with Mr. Chalifour was in Spring 2015, so this project has been in the making for a year and a half. Hadn't realized that until now.

Take care!
Beegle