SSL Scat-Talker Puts Old-School Speech Synthesis Into An MU Module

The Synthetic Sound Labs Scat-Talker is a dotcom-style MU module that puts old-school style speech synthesis under voltage control.

They note:

The Scat-Talker can be made to ‘speak’ by sequencing its phonemes in specific orders, but be warned that the process is quite time consuming. If that is your goal we have some helpful publications on our website, but perhaps the real joy is to use the Scat-Talker’s speech sounds as a new tool in your arsenal of musical sound sources.  Imagine – “Gibberish in C Minor”.

In the video demo above, John L Rice explores the capabilities of the Scat-Talker. Video notes:

0:00 – The Scat-Talker is a new limited edition module that contains a library of phonemes which can be selected by CV and Gate inputs with 1v/Octave pitch tracking. Words are formed by sequencing the phonemes.

0:21 – The SSL Scat-Talker is a single MU width 5U tall module for modular synthesizers

0:33 – Contained within the Scat-Talker is a lookup table of human speech ‘phonemes’, the smallest unit of speech that makes words sound different from each other.

0:47 – These phonemes can be selected under voltage control and be sequenced in an order that produces recognizable words or ‘invented’ and gibberish words

1:01 – The top of the module has a backlit display that shows which phoneme is currently selected including a usage example word and list number for handy reference.

1:12 – The PHONEME selection knob selects the phoneme but the output won’t change until a gate input is received or ‘HAL’ button pushed.

1:25 – The PITCH knob sets the initial pitch/frequency and the PITCH AMT knob sets the amount of PITCH CV input.

1:37 – The HAL 9000 button activates the currently selected phoneme and flashes red when pushed or a Gate input is received if the selection has changed.

1:48 – This intro section has effects on it provided by Strymon DIG and BigSky pedals. While the dry output of the SSL Scat-Talker is interesting and fun, the module really shines with external processing.

2:03 – When first powered on and all controls counterclockwise, the output will sound like this. Most phonemes are constantly output except for certain short and more percussive ones.

2:12 – The pitch range is quite wide. The upper limit is reached a little past the 8 mark, normal sounding pitches are had somewhere between the 6 and 7 mark.

2:37 – There are 63 available phonemes including silence and the 64th position is the factory demo.

3:00 – The output won’t change until the ‘HAL’ button gets pushed or a gate input is received and there is a new phoneme selected.

3:35 – Applying a control voltage to the PHON CV input selects phonemes.

4:04 – The display can’t keep up with fast modulation sources.

4:24 – Although, chimpanzees could probably keep up! https://youtu.be/zsXP8qeFF6A

5:03 – Applying a wide range modulation CV to the 1V/OCT input probably isn’t that useful since it will create out of range pitches.

5:45 – Using the PITCH CV input makes more sense with modulation sources since this input is attenuated by the PITCH AMT control pot.

6:03 – Nice ‘singing’ vibrato from a sine or triangle LFO.

6:37 – Here I’m taking the CV and Gate output from one 8 step row of a Moon 569 sequencer to get the Scat-Talker to say its own name.

6:59 – Fun and useful to shift recognizable phrases into something more interesting using the PHONEME control.

7:30 – The ACTIVE output is not a parallel output of the GATE input jack. A gate is only output when the phoneme selection has been changed since the last GATE input or button press.

7:52 – Using the ACTIVE output to fire an ADSR that is controlling a VCF that is filtering the Scat-Talker’s audio output.

8:26 – Using the sequencer to ‘speak’ but using a keyboard to play the notes I want it to ‘sing’ using the 1V/OCT input.

9:04 – Disconnected the sequencer and using the CV and gate output of the keyboard to control the pitch and gating of the Scat-Talker set to a static phoneme.

9:19 – Scat-Talker responds nicely to pitch bend.

9:32 – Interesting to mult the keyboard CV to both the PHON CV and 1V/OCT inputs.

9:52 – Experimenting using a keyboard with arpeggiator to sequence some short words. Worked well for quickly finding recognizable or interesting phoneme groups.

10:08 – The COTK C951 arpeggiator module worked well to sound phonemes in the order played and also easily reverse the order.

10:24 – The Scat-Talker contains English language phonemes but there are many common phonemes shared between different languages.

10:37 – I didn’t attempt other languages but many words may be possible? English spoken with various accents are doable.

10:55 – Shush, Hush? A moment of silence then!

11:31 – For the next minute and a half some singing with vibrato experiments.

12:38 – A high frequency artifact becomes apparent when the pitch is brought down to an unnatural range.

12:50 – This tone is interesting as part of the ‘sound’ but could be low pass filtered out.

13:07 – Picturing Popeye or other cartoon character being punched like a speed bag here!

13:17 – Mostly some sound effect experiments for the next minute or so.

14:32 – Some droning stuff with Scat-Talker processed through Strymon pedals again and random video clips for fun.

17:02 – This last section is the output of the Scat-Talker through a COTK C1660 Phase Processor

17:47 – The limited edition Synthetic Sound Labs Scat-Talker is available now!

Controls

  • Phoneme: Manual phoneme selection (one of 63), plus Demo mode.
  • Pitch: Manual initial pitch
  • Pitch Amount: Attenuator for external Pitch CV
  • Gate Button (with LED)

Inputs

  • Phoneme (PHON) CV: 0 to +5V
  • Gate: Logic ~1.2V threshold
  • 1V/Oct: Approx. 0 to +5V. Roughly exponential.
  • Attenuated Pitch CV: 0 to 100%

Indicators

  • 2 Line high contrast Blue LCD Display
  • Active LED (in Gate pushbutton)

Power

  • Synthesizers.com (DOTCOM) 6 pin Molex standard
  • +15vdc @ 73 ma, -15vdc @ 16 ma

The Scat-Talker MU – Model 1420 is available from the SSL site for US $345.

7 thoughts on “SSL Scat-Talker Puts Old-School Speech Synthesis Into An MU Module

    1. Anybody involved in large format modulars will understand this.

      Just about all 5U modules now on the market use the dotcom MU format – which is an advancement and improvement on the original Moog format. Moog reintroduced its old modules last year, though, which complicates things a bit.

      1. Those are actually built input the SC-01 chip and blend automatically between phonemes, although there’s no way to control them individually. They’re a part of speech called diphthongs, and without them speech would be far less intelligible. Good thought though. Wish there was a way to control them.

  1. Wow – kudos for this! This is proper implementation of my homebrew “Ursonator” module that I built a few years ago, but I could never get the pause / interrupt system to work properly. There is a huge range of sounds available by this method and the chants are excellent. The best thing mine does is to chant “beer” repeatedly! 🙂

    I hope this does well.

Leave a Reply