tr-909 - sample based?

[solaristica]

I just saw some Roland executive of the nineties saying on Modulations that great electronic music documentary that the TR 909 is sample based... Doesn't that disqualify it as "analogue" ?

[griffin avid]

I always thought (like most drum machines) that certain elements like clap and snap were sample based.
Where's the link to that video? I'd like to see it too.

[kuroichi]

The hi hats and cymbals are samples, all the other sounds are analogue.

[roland4387]

Kick, snare, and toms are analog. All the rest are 8 bit samples.

[Syn303]

The 707 is all sample based and it's the cheaper version of the 909, but the 707 is sound of chicago acid house though.

[solaristica]

So it's sort of a hybrid... analogue and digital sample... Thanks for clarifying.

Someone posted the entire movie on YouTube. "Modulations"...



About 8:30 minutes the Roland guy comes in... The whole movie is worth checking it out though...

[micahjonhughes]

There are three 8-bit samples used by the 909: one for the hi hats, one for the crash, and one for the ride. All of the other sounds are fully analog, including the clap.

[kuroichi]

Kick, snare, and toms are analog. All the rest are 8 bit samples.

looking at your equipment list and reading this just confused me.....

[nathanscribe]

There are three 8-bit samples used by the 909: one for the hi hats, one for the crash, and one for the ride. All of the other sounds are fully analog, including the clap.

Top marks. Extra house-points could have been obtained by explaining that the noise source for the snare, claps, etc, is digitally generated using CMOS, unlike the 808, for example, which uses a transistor.

[micahjonhughes]

I'll admit that I don't understand how CMOS chips generate "digital" noise. We could debate whether noise generated by CMOS chips or amplified circuit noise is better. Beats the DCO v. VCO battle.

[nathanscribe]

While a true analogue noise source is a voltage fully variable over a wide range, CMOS-generated noise is either up or down. On a scope it looks like a pulse wave gone apeshit - a seemingly random mixture of wide and narrow pulses and gaps changing fast. Sonically, I'm guessing here to some degree, but I expect this has a similar effect to mixing a bunch of randomly varying high-frequency square waves - which is similar to how the noise is generated for 606/808 cymbals (they use something like 6 simultaneous square waves at fixed frequencies).

It's not truly random though - the noise is a cycle that will repeat over time. How long it takes depends on the configuration of the circuit - the number of stages in the shift register that forms part of it. The 909 uses two 4006 (14-stage I think), with multiple tap-off points, which is about as close to random as I've seen a circuit like this get.

Some synths used an all-in-one chip for noise, which I can't recall the name of right now - it's in the Moog Rogue, for example - and that I suspect is digital, as the noise audibly cycles when you listen to it in isolation. Not a massive problem in most patches though, I find.

[Yoozer]

Top marks.

I thought 6-bit, not 8-bit. At least, that's what http://www.colinfraser.com/tr909/my909.htm says.

[micahjonhughes]

The samples could be 6-bit.

[nathanscribe]

Oops, yeah I'd forgotten about that. They come out through 6-bit DACs. I lose.

[Computer Controlled]

Modulations, while fun to watch, isn't exactly the place to go to for true facts. As already stated, the HHs and cymbals are 6 bit samples. The rest is all analogue.