Chipspeech is a vocal synthesizer developed by plogue. It resurrects old voice synthesis chips into virtual singers.
The vocalists within Chipspeech are all re-creations of vocals which are often no longer able to be found as easy as they once were. To collect all the same vocals separately would be a long and costly process due to the the limited numbers and rarity they are made available. In addition, Plogue were given exclusive access to the data used for a few of the vocals. Chipspeech is the most accurate creation of these chips currently on sale and offers them at a easy to access and more affordable price. Every vocal is created with permission of the original creators, even if the company were no longer around or had gone bust.
One of the biggest advantages that Chipspeech has over the original sources each vocal produces is that you can impact the vocal in ways that were not possible with the original courses. As such, vocal sounds that can be created in Chipspeech are not possible with the original versions to recreate.
Chipspeech has continued to receive a number of updates.
- 1.026; VOSIM added.
- 1.039; Daisy introduced as an optional extra vocal originally released for Alter/ego.
- 1.041; Lady Parsec HD improved, bug fixes made also.
- 1.066; Dee Klatt and Otto Mozer improved.
- 1.072; Formant Singer 2 is introduced improving the vocal results of Bert Gotrax and Dandy 704
- 1.5; Rotten.ST and CiderTalk'84 introduced. Talking abilities added, growl parameter added.
One thing that both Chipspeech and its sister software Alter/ego share is the use of the Formant Singer engine as a main source. Many of its vocals up to, and including, Daisy used this engine version. It is used whenever there is only vocal source materials available and can be adjusted to fit a more closer result.
The Formant Singer engine makes use of a single or monopitched result with a large vocal range. This is different to other synthesizers such as Vocaloid, which has a multiple set of pitches which produce a high quality result, but in turn produce a more limited vocal range, often impacting the overall usefulness of the vocals. The reason for seeking a large range is mostly owed to how some vocals like Bert Gotrax or Dandy 704 have limited source materials for use. One disadvantage, however, is that there is limited room for improvements on vocals like Dandy 704 regardless and Plogue can only improve him so much.
As of October 2015, Formant Singer 2 was worked upon and slowly introduced to Chipspeech during 2016. Formant Singer 2 improves results greatly and vocals like Bert now sound much closer to how their original source material sounded. In addition, the original versions of many of the Formant Singer vocals were created with with no use of a enlist. This is a list consisting of nonsense word and/or phrases used for recording the phonetic sounds needed for creating words in synthesizers. The result is sounds are more accurate in addition.
Since Chipspeech and Alter/ego can host a variety of different synthesizer types, this gives Chipspeech a bigger advantage over some other synthesizers in addition. For vocals like Otto Mozer, Terminal 99 or VOSIM do not use the Formant Singer and instead use their own separate technology. These produce results that are full simulations of the chip or source themselves and often re-creations of the source from scratch. These vocals have full capabilities to simulate their chips often accurately.
However, because they do not use the Formant Singer method they do not produce the same results. While these are the most authentic of the Chipspeech vocals, these have limited room for variations on the vocal. So the difference between a vocal like Lady Parsec and Terminal 99 is that variations of the Parsec vocal can be created, despite both being based on a similar source.
- Bert Gotrax
- Lady Parsec
- Lady Parsec HD
- Otto Mozer
- Dandy 704
- Spencer AL2
- Dee Klatt
- Terminal 99
Singing and Talking
Chipspeechs singing vocals via the X-SAMPA method of input for singing, allowing Chipspeech to support a wide variety of languages. Currently, it uses only English and Japanese. Vocals are flexible and highly adjustable.
The talking vocals work differently and do not support X-SAMPA or Japanese and instead are text-to-speech English only. Vocals will sound accurately like their original vocals and for some like Rotten.ST or CiderTalk'84, the result is a more realistic rendition of how the vocals originally sounded. In addition, not all parameters work with the vocals. The talking vocals react differently compared to those of the singing method and will react to pronunciation.
This is an accurate simulation of the act of trying to manipulate a chip via circuit bending. In real life, this involves things like working with a low battery. Traditionally, children's toys were a popular choice for this. The goal is to create a one-off choice result that would otherwise be beyond the chips normal capabilities.
Results are strange and often warped and unpredictable. This mode accesses half-sounds and broken words to create unusual sounds, product nonsense or gibberish, incoherent results. This is ideal for those seeking to create a unexpected sound for use within music. The results are completely random and recommended for those who enjoy experimentation.
Once opening Chipspeech, the top half of the window features eight different options, these are the Global Parameters.
Lets you select which voice to use. Voices are divided into 2 categories:
- Voice: Singing voice using lyrics. This is intended for musical use.
- Bending: Circuit bending. This lets you playback and loop raw preset voice data, with often interesting but unpredictable results.
Snapshots are “Master Programs”, complete setups with a “snapshot” of the current sate of the system at the moment of saving, including parameter values, MIDI controls and lyrics. You can save any number of snapshots into your user folder and even create sub-folders, which will mimic the structure of the the default programs. Snapshots have the .ariax extension.
Pitch Bend Range (PB Range)
Maximum amount of effect the MIDI pitch bender has on the pitch, in semitones.
Helium / Sample Rate Bend Range (He / SR BN Range)
Maximum amount of effect the MIDI pitch bender has on the voice resonant frequencies (formants), in semitones.
Switches between single voice (lets you play legatos) and multiple voices (lets you play chords).
Overall volume level of the voice.
Changes how fast each vowel transition is played, without changing the overall playback rate.
Wave RateHow fast the voice plays (transitions from each consonant/vowel to the next). Adding 25% will double the speed of the whole voice synthesizer.
The program also features a 'Words' tab. The words tab differs depending on whether you have a singing voice selected or circuit bending selected.
Swaps current line and previous one.
Swaps current line and next one.
Keep looping the current line instead of advancing through the text. You can automate this using MIDI CC79 or plug-in parameter automation.
Switch between editing lyrics in English (or Japanese hiragana/katakana) or in phonetic notation (using X-SAMPA - similar to IPA but doesn't require special characters).
Copy whole lyrics to clipboard.
Paste whole lyrics from clipboard.
Undo the last change to the lyrics.
Load preset lyrics to test your voice settings. You can also save your own presets.
Lyrics sung by chipspeech. You can edit the text, or cue a line for the next note. If you click on the line number, the line will also be cued next time playback is restarted.
Circuit Bending Voices
Selects which bank of words you play from.
Selects which entry from the bank should be played on next note.
Original Data Pitch
Controls whether thepitch used comes from the played note, or from the original pitch data (SP0 and TMS only).
Offsets where the data starts playing from.
Lets you adjust how much data is played before the reading position is reset. If you set this to 0, you can scan through the data in real time by automating the Loop Start.
Selects whether notes loop or stop playing once the current entry was read.
Freeze DataLoops the currently playing frame, emulating another famous circuit bending trick.
The controls tab features three different categories, Voice Parameters, Stack Parameters, and Setup Parameters.
This parameter scales the voice's resonance frequencies (formants). Raising this simulates a smaller mouth, giving rise to the “squirrel effect”.
This parameter modifies the voice's formants to be more male-like or female-like.
- Bert Gotrax, Dandy 704: 25% gives you the original male formants.
- Dee Klatt: 0%: husky male, 33%: higher male, 66%: female, 100%: higher female.
- Other voices: No effect.
Controls the amount of top-end harmonics.
- Bert Gotrax, Dandy 704: stretches the high spectrum to increase coverage.
- Dee Klatt: controls the mix level of the built-in harmonic exciter.
- Other voices: Controls the gain of the high-end shelf filter.
Controls the amount of noise added into the voice. (Bert Gotrax, Dandy 704 and Dee Klatt voices only)
Select which base waveform is used to simulate vocal cords. (no effect on Otto Mozer voice)
Controls the stretch factor of the base waveform. On the square waveform, this controls pulse width. (no effect on Otto Mozer voice)
Impulse Width Key Follow
Controls how much the impulse width follows the keyboard. Setting this to 0% will use the same waveform for the whole keyboard. Setting this to 100% will fully scale the waveform as you play higher notes. (no effect on Otto Mozer voice)
Increase or decrease the volume of noisy and sibilant consonants (p, t, k, f, th, s, sh, h).
Controls the number of detuned voices in the stack. 0% gives a single voice. 30~50% gives you about two voices. 100% gives you the full 9 voice stack.
Amount of detuning between the stack voices
Activates legato mode. Lets you do legatos in the melody without triggering new syllables. (monophonic mode only)
Controls the maximum number of voices in polyphonic mode.
Controls how fast notes fade in when played. In monophonic mode, only applies to the first note of a group.
Controls how fast the note will decay when released. The release is somewhat different in polyphonic mode.
Rate Key Follow
Controls how much the wave rate scales with the keyboard. 0% is fully pitch independent. 100% will fully scale speed, similar to a sampler.
Increases/decreases the pitch in octave steps.
Increase/decrease the pitch in semitone steps.
Fine TuneIncrease/decrease the pitch in cents.
The modulation tab has three categories, Vibrato Parameters, Expression Parameters, and Humanize and Glide Parameters.
Vibrato intensity, as controlled by the Mod Wheel .
How fast the vibrato is.
How much the vibrato rate increases when the vibrato amount is increased.
Delay time it takes for the vibrato to start.
How much effect the Mod Wheel (Midi CC1) has on vibrato amount.
How much vibrato is applied to pitch.
How much vibrato is applied to volume (tremolo).
How much vibrato is applied to impulse width (PWM LFO). (no effect on Otto Mozer voice)
How much vibrato is applied to helium (formant LFO).
Expression level is controlled by CC11. This lets you control the expression level with the Mod Wheel (CC1) as well.
How much effect expression has on volume.
To HF Sizzle
How much effect expression has on HF Sizzle.
Humanize and Glide Parameters
How much random variation is added to pitch.
How much random variation is added to volume.
How fast pitch changes when playing separate notes (triggering new syllables). (Monophonic mode only)
Legato GlideHow fast pitch changes when playing legato notes. In combination with legato mode, this lets you easy simulate voice inflections without using the pitch bender. (Monophonic+Legato mode only)
The mix tab has two categories, with one varying depending on the FX type selected, Effect Parameters - Ambience and Effect Parameters- Detune. The second category is th Mix Parameters.
Selects between the Ambience effect and the Detune effect.
Effect Parameters - Ambience
Ambience is a very smooth and flexible stereo reverb effect created by Magnus Jonsson, known mostly for its Ambience VST/AU version. In particular, the decay time can be shaped to different values in the low/mid/high frequency ranges.
Lets you load reverb settings from a preset.
How long the reverb decays.
How smooth the decay tail is.
Time between the echos (simulates a small or large room).
Amount of delay between the original sound and its reverb tail.
Amount of stereo separation (how different the reverb is on the left vs right side).
Increases the complexity of the reverb tail, at the cost of requiring more CPU to process.
Volume of reverb.
Equalizer Lo Freq
Frequency under which the low gain is applied.
Equalizer Lo Gain
Gain applied on low frequencies.
Equalizer Hi Freq
Frequency over which the high gain is applied.
Equalizer Hi Gain
Gain applied on high frequencies.
Damping Lo Freq
Frequency under which low damping is applied.
Damping Lo Amt
Decay time change under low damping frequency.
Damping Hi Fre
Frequency over which high damping is applied.
Damping Hi Amt
Decay time change over high damping frequency.
Effect Parameters – Detune
This effect creates two detuned versions of the voice input, one panned left and one panned right.
How detuned the simulated voices are.
How much effect is mixed in.
How delayed the simulated voices are.
Mix Parameters Pan
Balance of the voice between the left and right channels.
SendHow much of the voice is sent to the effect. Setting this to 0% disables the effect.
Each character has their own storyline and background which Plogue supply as an introduction.
One of the other noted things about the software is that each character has their own Twitter account, with a few also having other social network accounts. Characters will talk to each other about things, sometimes these are random things and other times these are events occurring to the characters.
Fans can also interact with the social accounts which can occasionally cause a character to react, providing additional information.
All of these are considered "canonical".
Outside of the current events, the Timeline for the history of the characters is a little vague. These are the only establish dates that are known;
- Dandy 704 was alive in the 19th century, for this reason it can be establish Dandy is well over 116 years old.
- Otto Mozer and Dandy met in the Great depression which occurred in the 1930s in America. Giving or taking a decade or two, this places Otto's age as anything from 80 to close to a hundred years old.
- Bert Gotrax is 12 years old, as a result all events that occurred after he was brought to Otto to be repaired by Dandy happened within the 12 years of his life.
- ciTech labs existed since 1984, since many of its dated site navigation maps are apparently dated at this year. This means ciTech has been around for 32+ years.
- On May 7th 2016 Terminal 99 reported he had been alone in the server room for 192,775 days. If this is correct, then Terminal 99 is over 528+ years.
Due to the nature of the accounts, there can be inflecting details caused at times of events and details on the characters.
These are the list of established major events.Due to the way the history of the characters has been presented, it is not currently possible to fully establish a truly accurate timeline of events. Events in italic are unknown on their true placement within the timeline. Their placement is an estimation.
- 520+ years ago, Terminal 99 was left alone in the server room.
- Dandy 704 made himself into what he is at some point in the 19th century.
- He met Otto Mozer during the 1930s during the Great Depression when work became hard to come by.
- In the last 12 years Bert Gotrax was brought to Otto for repair.
- From this point onwards Dee Klatt, Terminal 99 and Spencer AL2 were around at the time.
- Terminal 99 enters a Turing chat contest and wins.
- At some point the android Dee Klatt got accused of a crime, causing it to have to go into hiding.
- Vosim was being produced by a group of young scientists. When they could not impress their investors, Vosim and the labs he was being worked upon were abandoned. Later at some point they were knocked down. Vosim wondered aimlessly in the wastelands.
- Vosim leaves the wastelands, he enters a (be it low profile) singing career.
- Daisy entered the storyline - she and Dandy 704 became a couple.
- They departed, causing Otto to make the choice to become a full fledged robot.
- Daisy left suddenly, leaving behind codes.
- At some point, Dee Klatt found a kitten. The kitten stuck with Dee and was dubbed "Snowball".
- Vosim came along in 2015.
- At some point for a couple of months, Vosim and Dee Klatt dated.
- Dee wanted to discover who they were and did not want to endanger Vosim so ended their relationship.
- Dee's attempt to rediscover who they were ended in failure and they came to terms with the idea of never rediscovering themself.
- Meanwhile, Dandy 704 at some point ended up staying with Vosim, along with other robots.
- Vosim worked at a library in 2015 at a point.
- During the latter half of 2015, Dandy suffered a major memory loss. The loss was quite a serious one and he forgot everyone, friends included causing him to act strange.
- Rotten.ST is finally let out of prison, he hacks CiderTalk's account.
- CiderTalk'84 came into Chipspeech, inviting Dandy 704 into ciTech labs to stay and work when Vosim kicked Dandy out.
- The pair promptly got lost, leading to several days of trying to get back and Dandy nearly "dying".
- CiderTalk'84 examines Dandy 704's memories, after discovering things Dandy shouldn't remember he decided to not tell him any of the details. Vosim later confirms some of Cider's fears about the situation.
|Chipspeech Wikia Directory|
|Original||Bert Gotrax ● Lady Parsec ● Otto Mozer ● Dandy 704 ● Dee Klatt ● Spencer AL2 ● Terminal 99|
|Additional||VOSIM ● CiderTalk'84 ● Rotten.ST ● SAM ● Voder|
|Misc.||Plogue ● Pixoshiru|