The mysterious heart of the Roland TR-808 drum machine(secretlifeofsynthesizers.com)
secretlifeofsynthesizers.com
The mysterious heart of the Roland TR-808 drum machine
http://secretlifeofsynthesizers.com/the-strange-heart-of-the-roland-tr-808/
30 comments
I’ve owned 3 tr-808s in my life. I currently own one. I can confirm that each one sounds different. I’ve samples from each one. The first 2 I owned were from later in the production run. There are 2 distinct revisions of tr-808s. The early models had a shorter higher pitched snare sound and later models were a lower pitch deeper snare. My current 808 is an early revision with the high pitched snare. My current 808 has the weakest clap of the three I’ve owned.
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...
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...
> But I don’t think it is impossible to build an 808 that would not fool even the most seasoned experts.
This sentence hurt my head...
This sentence hurt my head...
Imho it should be:
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.
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.
Cancelling out, we get
‘I think it should be possible to build an 808...’
‘I think it should be possible to build an 808...’
Sorry about that. I should proofread my posts.
> But I don’t think it is impossible to build an 808 that would not fool even the most seasoned experts.
Whoah, double negative there
Whoah, double negative there
Triple negative actually. Before some heads explode, I'll attempt to turn it into vulcan approved logic:)
"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.
"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.
I agree. I was on another site where we were uploading samples of various owners 808s/808 clones and the results were pretty conclusive that once you match for level and settings you would never know if it was an 808 or a clone. The Yocto is supposed to be as close as possible in design, but does substitute some unavailable parts for modern equivalents. Ive heard a few different Yoctos, and they ranged from dead ringer to way out of adjustment, to unreliable and difficult to build correctly. The fact that it is a kit makes it very difficult to create repeatable results. There are a few other clones out there that I was less impressed with- ACIDLAB MIAMI is in the ballpark but examples Ive heard are not as close as I would hope.
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.
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.
If you're interested in claps and cymbals you may find interesting the Boss DR110 schematics here:
http://www.freeinfosociety.com/electronics/schemview.php?id=...
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.
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.
Wasn't there a study where wine tasters couldn't distinguish red wine from white wine containing red food coloring?
I am surprised that Roland would use such an unconventional method for generating noise.
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?
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?
I worked in a recording studio for a few years, and my impression is that for pop production, a lot of sound experiments are not reproducible. There are certainly fundamentals (e.g. understanding microphone design and placement techniques) but the random factors loom large.
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.
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.
You're right. The guitar sound in Dire Straits' Money for nothing is an experiment that couldn't be reproduced in another studio in NYC, even when placing microphones and speakers in the same random locations:
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...
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...
Re: a lot of sound experiments are not reproducible.
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.
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.
I concur. Made some music with a buddy and his ARP 2600, tweaking the knobs. Suddenly we both started dancing, ecstatically. Then the pitch started to drift downwards, detroying the vibe for me. He worked the machine and told me it is the same as before, but it did not feel as great as before at all.
They didn't expect this thing would become so massively popular and probably figured they had more than enough supply of this part to produce every unit they'd ever need to make.
It doesn't sound like it was intentional, I can imagine a rep or distributor giving the guy a bag of them without any proposed purpose, and the founder or an assistant found they were well suited for this task in an experiment.
its amazing how prevalent the 808 sound is in modern music... probably 80% of current rap music is based on sounds derived from the 808 kit. The combination of smoothness, "bassiness" and chattering hats and static snare really is infectious.
Here's a good look at the then-current top twenty, versus the same week in 1998.
808s everywhere.
https://youtu.be/rn85sCVEGOw
808s everywhere.
https://youtu.be/rn85sCVEGOw
ha, that's a good one. Speaks to the lack of diversity in current hip hop and also hip hop's dominance of pop music currently. Also tho, in '98 there is all kinds of transition happening in pop ...boy/girl bands in and out of popularity, dance music resurging but moving away from the euro thing, hip hop itself coming out of a golden era early 90's beat, and transitioning to pop centric and about to be super super influenced by the beats of Timbaland..that style..which of course, is more 808 based than what was before and dominated for years and years until finally the minimal 808 sound took over with I dunno kanye and stuff in the mid 00's? No where else to go with beats and it's lucky it has survived without much innovation. Meanwhile outside of the U.S. in places like UK influences of underground dance created innovative beats in the grime scenes and you hear things that are more creative/dancey/electronic/tribal even....which only those different producers singled out in this video of the current crop seem to be delving into. This is America beat? much closer to grime/drum'n'bass style out of the UK than mainstream U.S. production....
There is a fun documentary called "808" (available on Amazon Prime) about the history and impact of the TR-808. It's mostly interviews with musicians and only a little technical information.
https://en.wikipedia.org/wiki/808_(film)
https://amzn.com/B06XT8M3TV
https://en.wikipedia.org/wiki/808_(film)
https://amzn.com/B06XT8M3TV
How ironic that an engineering mistake which produced a unique "bad" batch of transistors also produced another engineering mistake. It's bad practice to make use of not reproducible parts, especially as such a unique noise source which basicaly defines a musical instrument. One of the most iconic electronic musical instruments ever created was born by engineering mistakes!
But, is there even a way to produce and reproduce a consistent physical noise source that generates the same noise with each physical copy?
But, is there even a way to produce and reproduce a consistent physical noise source that generates the same noise with each physical copy?
I wish they had actually explained how they did the noise modeling and the actual result.
Assuming the noise is stationary (meaning its probability distribution doesn't vary over time), it can be characterized entirely by its power spectrum, recorded over a time window that's long enough to represent the lowest frequencies of interest. The amplitude-versus-frequency response curve will exhibit various slopes at different frequencies, and it's these slopes that define the noise's "color" properties.
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.
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.
Agreed. The article implied the "special" transistor was used to create something like white noise. It'd be interesting to know what that noise sounded like compared to "normal" white noise.
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
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