The Dual SP0256 vocal synthesizer is the descendant of an earlier circuit of mine, Analog Controlled Speech Synthesizer. I crammed in all the features I’d want from a singing robot. It was a good challenge because it inspired some analytical thinking about the relationship between speech and singing. The features of this circuit are:
- 2 trigger inputs to step through pitch and vocal sequences independently
- speech data stored cartridge-style on EPROM
- 2-oscillator richness using dual SP0256-AL2 speech chips with crossfade mixer
- optimized low-pass filtering for less strident tone
- SP0256 crystal oscillators replaced with VCOs for pitch control
- each chip has an independent / linkable 4-step analog step sequencer for pitch
- lower step sequencer has switchable clock divider (divide/2 to divide/16)
- glide and LFO functions
- speech rate (slur) control
- built-in realtime programmable sequencers for pitch and speech trigger
- IR Sync input for wireless infrared tempo matching
And as a bonus, it makes a hair-raising screeching sound when the power is turned off!
The Dual SP0256 vocal synthesizer is now a member of Bodytronix. It begins to speak at 1:05 in this video clip:
To see how to replace the SP0256’s crystal oscillator with a VCO, check out the schematic for analog controlled speech synthesizer.
To compose the speech data I begin with the free PC app ChipTalk. ChipTalk converts plain text into SP0256-AL2 phonetic data. I paste that data into Excel, and compute some binary math thats necessary to add the ‘wait for trigger’ and ‘loop’ flags where desired. Then I use a hex editor such as Hexplorer to organize the contents of the speech EPROM; its organized as 16 phrases of 256 bytes. Once the EPROM image is composed, I burn it onto a 27C256 EPROM with a TrueUSB Willem programmer.
Sorry, the vocal synth is not for sale. No kit version is available. Schematic is private.