I am surprised Yamaha ever created the DX9 given the manufacturing
economics. Did they even save a few pennies here and there to create
an inferior instrument? I guess even in the early 80s product
differentiation was something.
Thank you! I agree with you. I find it hard to believe that leaving out the sub-CPU, one RAM chip and removing the velocity sensitive keyboard would've made a big difference to the manufacturing cost. My guess is that they always planned to have a budget offering, but didn't have a good technical idea of how to implement it. My opinion is that the DX9 was too handicapped in the end to be good value. According to this contemporary review[1], at least one person thought the lower price made up for the handicap!
'Budget' was relative. £900 is almost exactly £3,000 in modern money, and the DX7 was equivalent to £4300.
80s synth enthusiasts were often young and not particularly rich, so a cut down version wasn't an entirely insane idea. And if it hadn't been cut down enough it would have cannibalised sales of the more expensive synth.
Ironically the opposite happened - many people looked at the DX9, thought "No..." and pushed themselves to their limits financially to get the DX7.
I doubt this was deliberate, but it probably wasn't a net loss for Yamaha.
It's maybe more accurate to think of it as the prototype for the much cheaper budget 4-ops that Yamaha flooded the market with, after the DX7-tier was saturated.
One thing that you have to accommodate in this assessment, is that Yamaha run the worlds largest networks of music schools, and new musicians coming into, and subsequently graduating from those schools often had a fixed budget for their instruments.
The DX9 was Yamaha's way of marketing to those schools in a fashion that would imply affordability, while giving access to the 'professional' tools that commercial musicians would use (DX7). The idea being, the student starts off on the DX9/DX100/DX21 and then - when they graduate and can eventually afford it - upgrades at a later date to the DX7.
They used this strategy very successfully for decades, to expand their market and capture devoted fans of their instruments. Even today, Yamaha's product lines are designed to be affordable to students, and eventually when those students become professionals, provide a 'guided upgrade path' to the higher end products.
> And if it hadn't been cut down enough it would have cannibalised sales of the more expensive synth.
I think you're right about this. I originally thought that maybe removing velocity sensitivity would have been enough, but lots of people would probably have been happy living without it.
> It's maybe more accurate to think of it as the prototype for the much cheaper budget 4-ops that Yamaha flooded the market with, after the DX7-tier was saturated.
I'd genuinely love to know what was going on inside Yamaha after the success of the DX7. There's a lot of information available about the development of the DX7, but I can't find much about what happened after. I'm not sure why there were so many of the 2nd generation 4-operator FM synths: DX21, DX100, DX27, FB01, TX81Z, V50, TQ5, and a bunch of other variants of these. Of these, only the TX81Z has really achieved the same kind of 'iconic' status as the DX7.
The answer is: music schools. We don't see this in the west so much, but Yamaha run the worlds largest networks of music schools, predominantly in Asia.
The DX21, DX100, DX27, FB01, TX81Z .. product range .. is designed to cover the following use cases: music schools, karaoke bars, normal bars, and hobbyist musicians. In fact, Yamaha's biggest FM-based products (in terms of sales generated) is the karaoke market, where the more 'sophisticated' synths would be used by karaoke artists to create the MIDI files that then get shipped off to be 'played back' on the karoake-friendly devices, such as TG-series.
Evidently, the answer is that the DX7 was already cheap to produce. Engineering new stuff and spinning up a new assembly line would probably cost more than they would save. Remember, market segmentation is often about price discrimination. It's wrong to assume that consumer goods are priced at cost-plus-margin.
For what it's worth, they're actually totally different machines under the hood. They share the same main CPU, sound chips, and D/A section, but those are about the only similarities. I think you and GP make great points about product differentiation/price discrimination though. I think this explains Yamaha's motivations.
> I am surprised Yamaha ever created the DX9 given the manufacturing economics.
It was quite common back then to produce the flagship instrument at high price for professionals, studios, etc, then once it became popular creating a demand, follow with a cheaper line. Happened with other brands too: Roland made a fortune with cut down versions of the D50.
I didn’t have the money for a DX7, but I did have it for a DX9 back then, so that’s what I got. It lacked velocity sensitivity but was still superior to other offerings in that price class, IMO at the time. It served me well for many years, so for some it did make sense.
If you come across a CZ230S, be aware that you can exfiltrate all 100 of the fixed voicings through SysEx. I got that far, saving those tables to disk; my plan, when I ran out of time and attention, was to SysEx in one of those settings to the last entry in the CZ101's table, which is overwriteable, so as to get 8 notes at a time. Maybe you'll find a place to drop in a ROM with them all.
Right now I'm just having fun poking around the code, investigating if there's much room for firmware modification. The CZ101 uses the NEC uPD7811 CPU, which has a really interesting architecture: It doesn't have any conditional branching, only conditional 'skipping' of instructions. I'll probably write something about the CPU at very least!
Hmm, I wonder if I could do something similar with my TX81Z, which is more like a DX11. But I don't really need to – it's great as it is, and shines as a 4-op FM synth because of its non-sine waveforms, and because of its multitimbral capability. Still cheap on ebay too.
Though a free FM softsynth like Dexed is hard to beat for cost and convenience; and on the hardware side, Korg's opsix has a great front panel with knobs and sliders for programming, while their Volca FM2 is a nice 6-voice mini-synth that can play DX7 patches.
Excellent work and very interesting to read. DX9s very occasionally pop up for sale for silly small money, and now I would seriously entertain getting one just for the fun of having a crack at installing your custom ROM.
I imagine if Yamaha ever starts going after the people eating into their FM synthesis profits they'll have a long, long, long list of people to deal with before they get to me! In 1983 this firmware would definitely have upset someone, but right now I think I'm pretty safe.
Author here: Thank you so much for posting the article! I'm happy to answer any questions about the project.
If you find this interesting, I wrote an overview of how I approach these projects: https://ajxs.me/blog/Introduction_to_Reverse-Engineering_Vin...
I'm currently messing around inside the Casio CZ101, another cool 80s digital synth.
Fantastic read! Love that you did this.
I am surprised Yamaha ever created the DX9 given the manufacturing economics. Did they even save a few pennies here and there to create an inferior instrument? I guess even in the early 80s product differentiation was something.
Thank you! I agree with you. I find it hard to believe that leaving out the sub-CPU, one RAM chip and removing the velocity sensitive keyboard would've made a big difference to the manufacturing cost. My guess is that they always planned to have a budget offering, but didn't have a good technical idea of how to implement it. My opinion is that the DX9 was too handicapped in the end to be good value. According to this contemporary review[1], at least one person thought the lower price made up for the handicap!
1: https://www.muzines.co.uk/articles/nine-times-out-of-ten/791...
'Budget' was relative. £900 is almost exactly £3,000 in modern money, and the DX7 was equivalent to £4300.
80s synth enthusiasts were often young and not particularly rich, so a cut down version wasn't an entirely insane idea. And if it hadn't been cut down enough it would have cannibalised sales of the more expensive synth.
Ironically the opposite happened - many people looked at the DX9, thought "No..." and pushed themselves to their limits financially to get the DX7.
I doubt this was deliberate, but it probably wasn't a net loss for Yamaha.
It's maybe more accurate to think of it as the prototype for the much cheaper budget 4-ops that Yamaha flooded the market with, after the DX7-tier was saturated.
One thing that you have to accommodate in this assessment, is that Yamaha run the worlds largest networks of music schools, and new musicians coming into, and subsequently graduating from those schools often had a fixed budget for their instruments.
The DX9 was Yamaha's way of marketing to those schools in a fashion that would imply affordability, while giving access to the 'professional' tools that commercial musicians would use (DX7). The idea being, the student starts off on the DX9/DX100/DX21 and then - when they graduate and can eventually afford it - upgrades at a later date to the DX7.
They used this strategy very successfully for decades, to expand their market and capture devoted fans of their instruments. Even today, Yamaha's product lines are designed to be affordable to students, and eventually when those students become professionals, provide a 'guided upgrade path' to the higher end products.
> And if it hadn't been cut down enough it would have cannibalised sales of the more expensive synth.
I think you're right about this. I originally thought that maybe removing velocity sensitivity would have been enough, but lots of people would probably have been happy living without it.
> It's maybe more accurate to think of it as the prototype for the much cheaper budget 4-ops that Yamaha flooded the market with, after the DX7-tier was saturated.
I'd genuinely love to know what was going on inside Yamaha after the success of the DX7. There's a lot of information available about the development of the DX7, but I can't find much about what happened after. I'm not sure why there were so many of the 2nd generation 4-operator FM synths: DX21, DX100, DX27, FB01, TX81Z, V50, TQ5, and a bunch of other variants of these. Of these, only the TX81Z has really achieved the same kind of 'iconic' status as the DX7.
The answer is: music schools. We don't see this in the west so much, but Yamaha run the worlds largest networks of music schools, predominantly in Asia.
The DX21, DX100, DX27, FB01, TX81Z .. product range .. is designed to cover the following use cases: music schools, karaoke bars, normal bars, and hobbyist musicians. In fact, Yamaha's biggest FM-based products (in terms of sales generated) is the karaoke market, where the more 'sophisticated' synths would be used by karaoke artists to create the MIDI files that then get shipped off to be 'played back' on the karoake-friendly devices, such as TG-series.
Evidently, the answer is that the DX7 was already cheap to produce. Engineering new stuff and spinning up a new assembly line would probably cost more than they would save. Remember, market segmentation is often about price discrimination. It's wrong to assume that consumer goods are priced at cost-plus-margin.
https://en.wikipedia.org/wiki/Price_discrimination
For what it's worth, they're actually totally different machines under the hood. They share the same main CPU, sound chips, and D/A section, but those are about the only similarities. I think you and GP make great points about product differentiation/price discrimination though. I think this explains Yamaha's motivations.
[dead]
> I am surprised Yamaha ever created the DX9 given the manufacturing economics.
It was quite common back then to produce the flagship instrument at high price for professionals, studios, etc, then once it became popular creating a demand, follow with a cheaper line. Happened with other brands too: Roland made a fortune with cut down versions of the D50.
I didn’t have the money for a DX7, but I did have it for a DX9 back then, so that’s what I got. It lacked velocity sensitivity but was still superior to other offerings in that price class, IMO at the time. It served me well for many years, so for some it did make sense.
If you come across a CZ230S, be aware that you can exfiltrate all 100 of the fixed voicings through SysEx. I got that far, saving those tables to disk; my plan, when I ran out of time and attention, was to SysEx in one of those settings to the last entry in the CZ101's table, which is overwriteable, so as to get 8 notes at a time. Maybe you'll find a place to drop in a ROM with them all.
I'd love to hear what you're doing with the CZ-101. CZ-5000 owner here.
Right now I'm just having fun poking around the code, investigating if there's much room for firmware modification. The CZ101 uses the NEC uPD7811 CPU, which has a really interesting architecture: It doesn't have any conditional branching, only conditional 'skipping' of instructions. I'll probably write something about the CPU at very least!
> only conditional 'skipping' of instructions.
Something you have to consider if you want short fixed length opcodes. I guess you could call it "Macro Op Fission."
Casio's CZ/phase distortion synths are underrated, terrific sounding instruments (perhaps because of their 12-bit DAC?)
All synths should have (at least) 8-step envelopes like the CZ series.
Full multi-timbral hardware synths also seem rare these days.
Really cool project. Hope you still have fun with your DX97!
This is right up my alley right now! I bought a busted CZ-1 as my first synth and I’m trying to get it working. I look forward to reading your blog!
I quite enjoyed this quote:
But as someone who always wishes they had had a DX7. has a Reface DX and built multiple Dexed hardware synths, I really appreciated the whole thing.Korg Volca FM is a DX7 clone in a tiny and very cheap package: https://www.korg.com/us/products/dj/volca_fm2/
If anyone is interested, there is a related meme/trend, "something that’s an absolute flex to a niche group but means absolutely nothing to most"; popular examples: https://www.instagram.com/mkbhd/reel/Cbio5sLFV81/ , https://x.com/thebunten/status/1494373600240963584
Hmm, I wonder if I could do something similar with my TX81Z, which is more like a DX11. But I don't really need to – it's great as it is, and shines as a 4-op FM synth because of its non-sine waveforms, and because of its multitimbral capability. Still cheap on ebay too.
Though a free FM softsynth like Dexed is hard to beat for cost and convenience; and on the hardware side, Korg's opsix has a great front panel with knobs and sliders for programming, while their Volca FM2 is a nice 6-voice mini-synth that can play DX7 patches.
Excellent work and very interesting to read. DX9s very occasionally pop up for sale for silly small money, and now I would seriously entertain getting one just for the fun of having a crack at installing your custom ROM.
There have been many times where companies took down the resource. Be aware that they may make a claim against you !
I imagine if Yamaha ever starts going after the people eating into their FM synthesis profits they'll have a long, long, long list of people to deal with before they get to me! In 1983 this firmware would definitely have upset someone, but right now I think I'm pretty safe.
They just hate people having fun, they don't care about profits at this point.
I was wondering what sort of motorbike it was until I realised it was about the synth