As specialists in Speech To Text – not to be confused with text to speech – at Authôt we pay a great deal of attention to vocoders and electronic devices for processing sound signals. In fact, today we are more than ever in an era that favours audio. It’s very simple, audio, voice recognition, is everywhere and developments on this subject are multiplying all the time! What’s more, it requires real technical expertise. So, after having presented the Soundfilter tool, we will now discuss physical machines. Tools or devices that process sound, audio, sound signals. Are you ready to be amazed? Here we go!
To begin with, what is a vocoder? According to the Larousse.fr, this term was born from the combination of “vocal” and “coder”. In concrete terms, it is an “electronic device allowing the synthesis of vocal responses by a computer system”. This synthesis is done by “words, phonemes or coded diphonemes. It is electronic circuits that have these codings as inputs that allow these vocal responses to be rendered.
As Wikipedia states, the vocoder “analyses the main spectral components of the voice – or of another sound – and produces a synthetic sound from the result of this analysis”. In this theme of sound signal processing, it is therefore a question of creating a synthetic result thanks to decoders.
In fact, whether applied in the world of music, video production or content, the use and ‘electronic’ processing of a voice is crucial. As a result, there is a whole range of different equipment available. Indeed, as sonovente.com points out, although it is often “confused with a talk box, a vocoder is actually a device that allows you to sing into a microphone and translate the timbre of your voice into an instrument – most often a keyboard. This vocal effect is mostly used in pop music. An example of this phenomenon is the music of the French band Daft Punk.
Vocoders have their origins in the 1920s. They were a device used by the military to encode voice messages. They converted their voices into coded analogue signals. It was the receivers that were able to decode these signals and understand secret military information.
Then, around the 1960s, engineers built on this principle of alienating the natural voice to develop devices whose sound generation is different… Sound sources such as other audio signals or the human voice.
The operation of vocoders is generally simple. Such a sound signal processing device consists of two inputs:
– The carrier input: for an instrument;
– The modulator or analysed input: for a microphone or other signal source.
The sound sources are varied.
For vocoders and other electronic devices, there are many challenges in sound signal processing.
According to metisse-demo.irisa.fr, the sound signals “that we produce and that we perceive around us form a sound environment in which we have learned to distinguish sounds and to associate them with objects, situations, words, music or other concepts.
For the machine, however, the process is different! It has no awareness or knowledge of its environment. The challenge is therefore to analyse, decompose, compare and recognise sounds electronically and automatically.
For researchers in this sector, the aim is to build systems that draw on many other sectors: linguistics, acoustics, artificial intelligence and computing. By alternating theory and experiments, the processing of sound signals is becoming increasingly sophisticated.
In addition, there are other electronic devices that will process sound signals in their own way. These include talk boxes and auto-tune technology.
A talk box is an effect that allows a musician to link the sound of his instrument to a voice. This means that he can create a new sound, for example, a guitar or a singing synthesizer. From the sound of the instrument, words are then projected onto it from the musician’s mouth. The effect shapes a sung phrasing, words. This converts the sound of the instrument into a voice in a special way.
This effect created from a talk box is called “morphing“. It consists of modifying the original audio signal by using a mouth as a “filter” to shape a sound.
Auto-tune, on the other hand, has been the ideal ally for over twenty years for those who can’t sing in tune…! It is a tool, a software that allows one to have a clear voice, to improve the accuracy of the voice with a correction of the tones in real time.
In fact, auto-tune is an audio processor that uses a device to measure and correct the pitch of notes during recording. Since 1997, this type of device adjusts the voice, the soundtrack, to the nearest semitone!
Whether in the form of a plug-in or stand-alone hardware, the use of auto-tune is very common in professional recording studios today.
Basically, a vocoder mixes two sounds – instrument and voice – to achieve specific effects. The Talk Box creates an artificial voice from instruments. Auto-tune is a process that uses algorithms to automatically correct the tone of a sound – voice and instrument.
Who knows what the next sound development will be! In any case, we’ll be watching closely…!
We hope you have enjoyed this article about sound signals! Don’t hesitate to tell us if you want more technical articles! To benefit from our expertise in sound signal processing and much more, contact us!
Authôt. You speak. We write.