The mysterious heart of the Roland TR-808 drum machine
secretlifeofsynthesizers.com
secretlifeofsynthesizers.com
I mention all of this to make a few points. 1- I don’t doubt that these rejects contribute to the sound in a very specific way. But I don’t think it is impossible to build an 808 that would not fool even the most seasoned experts.
2- the 808 was not exactly a hit product. The “we ran out of parts” line sounds to me like a great way to revise what the real truth of the matter is- it wasn’t selling since sampled drum drum machines like the Lynndrum/Oberheim DX/sequential drumtraks were in far greater demand. Most people at the time thought the 808 was terrible sounding. It soon was widely available at thrift shops and used gear racks for 10-20% of its original 1000$ price. Ad rock of Beastie Boys fame mentions in an interview that he went to Rogue Music in NYC to get a rickenbacker guitar but instead found a used 808 for 250 bucks around 1984 or 1985. Since the tr909 was released in 84 or so with a focus on sounding more realistic, I don’t think parts availability is the real reason.
3- there are solid clones available that are 100% analog and approach if not hit 100% accuracy in the sound. The Yocto kit when assembled competently is pretty much a 1:1 copy. Behringer has a forthcoming rd-808 that is analog and from the videos sounds the business.
4- the modeled 808 in the Tr08/tr8/Roland cloud 808 is in the ballpark but the cymbals snare and clap are the weakest parts of the modeling. I would say those pass the 5 minute quick listen test but once you spend time with them you realize it’s not right in an annoying way. If anyone wants to read more about the tr-808 a me Robin Whittle has done some amazing work documenting the circuits, explaining how they work and gives instructions on modifying them. Now that these machines are 30 years old, I hope people don’t mod them and focus on preserving them at this point- especially considering that there will soon be cheap Behringer copies to mod.
Finally just want to say it’s great to see this kind of stuff on hn. I could write long winded comments on this stuff all day long.
Edit- adding link to RWs 808 mods/info http://machines.hyperreal.org/manufacturers/Roland/TR-808/mo...
This sentence hurt my head...
But I don’t think it is impossible to build an 808 that would fool even the most seasoned experts.
Because there are synts with built in imperfections that are very good.
I also believe some synths synthesize the electronic components instead of the sound.
‘I think it should be possible to build an 808...’
Whoah, double negative there
"I think it's possible to build an 808 clone that would fool even the most seasoned experts."
I'm 100% sure about that since a junction generating noise is just a junction generating noise: color is added by filtering/clipping afterwards. One transistor can be much noisier than others but other stages would compensate level differences anyway. To me the only requirement for building a 808 clone that sounds just as the original is to put it in the same original enclosure; psychology plays a role here: if you put the original 808 circuit into a modern enclosure some people will tell it's a clone and sounds different. I wish I had time and money to make a blind test because I'm totally sure most people, including experts, would be fooled both by a clone in the original case and by the original circuit put in a different enclosure.
For me, Im most interested in the clap, ch, oh, and cy sounds. Those are the most unique sounding to me among analog drum sounds. The hi hat and cymbal sounds are remarkably complex and everything else Ive heard sounds too much like a metallic pulse followed by white/pink noise. The 808 is not necessarily realistic sounding but it is very cool sounding to my ears. The bass drum, which is mainly what made the 808 so popular in the first place, is now pretty easily duplicated or sampled with other synthesis types, but the 808 kick is unique in that it is more a (near)self oscillating filter controlled by an envelope than a pulse followed by sine wave that other analog kicks use. This definitely contributes to its less static, more bouncy kick drum sound, especially at longer decay settings. Its pretty cool that you can program a pattern and on the first pass, subsequent kicks will add to each other, and then the next time around, they cancel each other. This combined with accent can create some very live feeling patterns, and this is a very underrated feature of the 808.
Especially the voicing board schematics is simpler than the 808 one. Interestingly they used a shift register for noise while cymbals are still made using dissonant oscillators with cmos gates. The schematic though is much easier to read and understand. That drum machine sounded close enough to the 808 to gain its share of popularity when a couple decades after that sound became mainstream, and of course I was stupid enough to sell mine for cheap just before that moment.
Sourcing and testing those bad transistors had to be more expensive than just buying good parts, and surely they must’ve known that their supply of bad parts wouldn’t last forever?
Success depends less on applying principles in a straightforward way and more on being able to recognize when something is compelling and doggedly experimenting until something clicks. The story of the TR-808's design feels familiar.
https://www.soundonsound.com/people/classic-tracks-dire-stra...
And Phil Collins' gated reverb drums, one of THE sounds of the 80's, were discovered through another random placement of microphones, although (un)luckily this time it was easy to reproduce:
https://www.vox.com/videos/2017/8/18/16162440/recording-stud...
A lot of early Moog synthesizer users complained about the inconsistency. They'd get a "great" sound in the studio, but on stage it sounded different. It was one of the reasons for slow adoption by general pop groups. While many musicians thought they were "cool sounding", their inconsistency kept them out of mainstream.
Kraftwerk often made their own circuits so they could control and tune them as they saw fit. It's one of the reasons why they were (arguably) first with the "80's sound" in the 70's.
808s everywhere.
> How exactly Mr Kakehashi found these specific parts is not known, [...] but you can hear him tell some of the story himself right at the end of the excellent 2015 documentary movie “808”.
But, is there even a way to produce and reproduce a consistent physical noise source that generates the same noise with each physical copy?
Here's a video that explains how to go from a white noise source to something that sounds like the TR-808 snare. https://www.youtube.com/watch?v=8LWQqDHhpKw
Once the spectrum is known, synthesizing the same sound is really just a matter of starting with white noise and shaping it with various lowpass, highpass, and/or bandpass filters whose skirts match the original slopes. It sounds like the 909 took that approach, generating noise by sending the output of a maximal-length shift register through some analog filters.
If the noise isn't stationary -- for instance, if its frequency content varies noticeably as the envelope progresses through its various phases, as the instrument warms up, or simply as it ages -- then it's no longer so easy to model and reproduce the sound. That's the sort of thing that can easily happen when the design relies on defective transistors.