Recording of DJ Culture and Was It Worth It?

[markus]

When with some proper AI software I plan to create a mashup of Dj culture using the backing track of Where the streets have no name, taking it up to well deserved new heights. Maybe next tour they'll remember this underrated moderate hit once again.

[102dB]

Haha well go on then, enlighten me - how was it done? The vowel sounds seem somehow unpitched rather than sounding too high or too low which I'd expect if they had been sampled from an actual disc that was being spun by hand - I just don't know enough about music tech from that period to understand what they've done. In other words, how did they get the words to have the grain of a record scratch and yet still be identifiable as speech?

Back in 1991, there were quite a few different ways to create that robotic "DJ Culture" voice. It's far from rocket science and actually quite easy to create.

#1) It could have been any sound effects processor (a hardware device typically used in recording studios) which supports pitch shifting, for example Lexicon, Eventide (the H3000). Even a cheaper effects processor such as the Yamaha SPX90 (released in 1986) featured pitch shifting.

#2) The graininess of the robotic voice can be an indicator it was created on an Akai S900 or S1000 sampler. These devices were famous for their grainy time-stretching capabilities. Two Akai S1000 samplers feature quite prominently on the cover sleeve of DJ Culture: the white/grey machines at arm's height on the left next to Chris, on the right next to Neil.

#3) Back in those days, the Atari ST computer was a very popular studio computer, used in many production studios around the world to program/sequence synthesizers and samplers. There was a software for Atari called "Speech synthesizer" which allowed you to type the words and the Atari created the sound of a speaking voice. It sounded very robotic and grainy, especially at 8-bit and 4-bit digital audio resolution. For example the robotic vocals in the 1991 techno song "Das Boot" by U96.
https://www.youtube.com/watch?v=YVxXbTk-zsQ

Here are a few examples how robotic voices were created in the early '90s on the Atari computer:
https://www.youtube.com/watch?v=DhExwdZEyLk
https://www.youtube.com/watch?v=xx1qXzaKBdA

[Ramzy]

Thank you, 102db, this is really interesting. It wasn't exactly the speech synthesis itself I was surprised by, so much as the fact they made it sound like a record (rather than a robot) "talking". I guess it was just about trying different things until it sounded right, probably all the things you've helpfully listed above!