By Dr Laila Abdel Aal Alghalban
Give me the flute, and sing,
Immortality lies in a song
And even after we perish
The flute continues to lament
The Lebanese poet Gibran Khalil Gibran
What is in a song? Singing is magic; singing is another form of prayer. In fact, how melodies penetrate into our souls, sway us to the fullest, hit our emotions at their rawest and nourish our thought remains a mystery. What we all know is that we all sing with varying degrees of competence. Singing involves forming words, vocalising them at certain pitches and notes, all require countless neurological, physiological, psychological and physical tasks. How does singing affect us and other fellow creatures? Moreover, with the mind-blowing leaps of Artificial Intelligence (AI) taking the world by storm and the meteoric rise of deep writing, deep literature, deep journalism, deep fakes, etc., can deep or bot singing also wow us with bot pop stars, along with bot influencers and avatars assuming increasing presence day after day?
Body organs work rhythmically
How music affects the body is a very interesting and yet not fully-discovered issue. Singing has to do with sounds which are harmoniously combined to change the chemical interactions in our bodies and evoke emotions in us. Think about your voice, that miraculous musical instrument that plays, through speaking and singing, varied tunes that stir up emotions beyond comprehension. Think about how your body organs work rhythmically: your heart, feet, hands, respiratory system, etc. Music is inside us and it is part of who we are. Language also has music, and when language and music get together, they rock usmassively. Think about the emotions you experience when language and music are combined in movies, songs, opera, theatre, etc.
Singing makes wonders
Singing helps people with dementia come back to their old selves and restore part of their crumbling memories and diminishing past; they feel less isolated. During hardships, singing is always there, soothing people’s hearts and tapping softly on their troubled feelings, like a magic wand; after a severe drought and experiencing low yields, farmers simply sing for the sake of singing, for the sake of sharing asense of community. During Covid-19 lockdown, singing has been a comfort zone for all of us, it turns into a sort of talking to the Master of the universe, our God to have mercy on us and protect us. Even while protesting and facing brutality and possible detention, demonstrators turn to singingto voice their protest and vent their anger. All professions have developed a legacy of songs sung by workmen during working hours, it is folk music. Homesickness has always been a catchy theme, where people find solace in singing for home.
Since humans walked the earth, they have appreciated birdsong as one of nature’s most melodic sounds. In all cultures, birdsong is mapped on to lovely, kind, gentle, caring and loving individuals. We are not alone in being lured by singing. Animals are also drawn to singing: “A cockatiel who flew from his cage has been returned to his owner after he was heard singing The Addams Family theme,” reads a famous news story. Kulning, an old Nordic high-pitched singing used as a call for cows, sheep, and goats to come back to barns and cottages at sunset after grazing in far distances, is another amazing instance of the magic of singing. Charming snakes is another case in point.
Teaching a computer to sing like a human
The first step is turning speech into text. Thanks to speech recognition technology and voice text convertors; it has become easy to turn millions of audio files into texts though it still stumbles with a lot of problems such as converting the delicate melodies and pitch trajectories of human voice. Meanwhile, chatbots are becoming so common and sophisticated in simulating the human voice that a Russian chatbot has managed to fool an AI specialist who spent two months trying to seduce ‘a computer program!’ smart speakers like Siri and Alaxa furnish another case in point.
When it comes to teaching computers to sing, a lot of algorisms or computer formulas must be developed to simulate the human voice intricacies such as stressing the correct syllables. Recognising the different allophones of phonemes in different phonetic contexts, watching various durations of individual sounds, syllables and words. Another important step is to teach computer how to beat the rhythm trough the correct succession of stressed and unstressed syllables in stressed-timed languages such as English and Arabic, or simply giving all syllables the same duration in other languages called syllable-timed languages such as French. Of course these instructions must be machine-readable. The most hectic problem is hitting the right note, the right melody. When we sings, our voice made up of lots of frequencies, the phonation of our vocal folds and its subsequent overtones or resonances, termed formants. The combination of frequencies differs as the note changes.
Finally, let’s wait for more wonders.
By Dr Laila Abdel Aal Alghalban
Professor of linguistics
Faculty of Arts
Kafr el-sheikh University
Email: [email protected]
