FOCUS: MTS to present AI-voiced audiobooks soon to save cash, time - News Archive - PRIME Business News Agency - All News Politics Economy Business Wire Financial Wire Oil Gas Chemical Industry Power Industry Metals Mining Pulp Paper Agro Commodities Transport Automobile Construction Real Estate Telecommunications Engineering Hi-Tech Consumer Goods Retail Calendar Our Features Interviews Opinions Press Releases

FOCUS: MTS to present AI-voiced audiobooks soon to save cash, time

By Yekaterina Yezhova

MOSCOW, Jan 30 (PRIME) -- Major Russian mobile operator MTS follows Apple with its forthcoming launch of audiobooks voiced by artificial intelligence (AI) as the technology is gaining momentum, the companies and experts said.

MTS launched e-books service Stroki (or Lines) only in September 2022. Later we discussed using the research of MTS AI, also part of MTS Group, for a robotized voice playback of books. We have recently announced the start of the project, about 1,000 books are in progress now, an operator spokesperson told PRIME.

AI is simulation of human intelligence processes by machines, especially computer systems.

MTS AI will transform texts into audiobooks on the basis of its own platform of speech recognition and synthesis, called Audiogram.

The product uses neural networks, machine learning (ML), and the natural language processing technologies, the spokesperson said.

The platform records audiobooks in the popular EPUB format. It can not only synthesize speech, but also stress and pause correctly, reproduce intonations specific for texts, including interrogative and illocutionary.

Stroki will offer the books voiced with the speech technologies soon, the spokesperson said.

The company picks popular but not mass books in the text format, which cannot be dubbed with the help of actors any time soon, including self-help and self-development editions.

MTS did not disclose its investment in the project, but said that the technology is popular because it saves time and money on turning texts into audiobooks.

Yevgeny Selivanov, head for content development at LitRes, the countrys leader of the audiobooks market, said earlier that the prime cost of making one audiobook via AI amounts to 400700 rubles. A narrators work costs from 50,000 rubles per book, according to sources.

The Guardian reported in early January that U.S. tech giant Apple had quietly launched a catalogue of books narrated by AI, the move that may mark the end of human narrators, yet it is an attempt to enjoy the lucrative and fast-growing audiobook market, which may reach U.S. $35 billion by 2030.

Deputy Prime Minister Dmitry Chernyshenko said in mid-January that Russias AI technologies market should at least double to 20 billion rubles by 2030 as more than 52% of large companies in the country implement AI in their operations.

From 2023, the countrys state agencies will also report to the government on implementation of AI-based solutions as well. Experts said earlier that the government needs AI to make fact-based decisions, while neural networks will speed up provision of state and municipal services.

President Vladimir Putin said earlier that the importance of breakthroughs in AI is colossal, and the results achieved in this industry will define the countrys place in the world, its independence, security, and potential to reach the goals of industrial and social development on an entirely new level.

Elena Belobrova, head of the group of development of AI and ML services at Yandex Cloud, part of Internet giant Yandex, told PRIME that the company sees growing interest to AI technologies to voice all sorts of texts.

Our clients voiced more than 100,000 hours of audiobooks with our service of speech synthesis SpeechKit, she said.

The use of speech synthesis helps companies speed up voicing of books by several times. To use SpeechKit, one needs to open an account in Yandex Cloud and contact the service via an application programming interface. One can turn a text online into a speech and save it as an audio file. Now Yandex Cloud offers seven voices in Russian, besides, the company can create a unique voice with the Brand Voice technology.

AI is improving all the time to make speech sound more naturally. We use transformers when teaching synthesis models in Yandex SpeechKit. This architecture of neural networks produces perfect results in various applied tasks related to AI, Belobrova told PRIME.

(69.3372 rubles U.S. $1)


30.01.2023 09:30
Share |
To report an error select text and press Ctrl+Enter
Central Bank Official Rate
1W 1M 1Y
EUR 86.5009 -0.0110 01 jun
USD 80.9942 +0.3070 01 jun
Stock Market Indices
1D 1W 1M 1Y
micex 2718.27 +0.72 18:50 01 jun
Stock Quotes in RUR
1D 1W 1M 1Y
gazp 163.09 +0.74 18:49 31 may
lkoh 5575.00 +0.81 18:49 31 may
rosn 447.15 +2.91 18:49 31 may
sber 245.91 +0.52 18:49 31 may
MICEX Ruble Trading
1D 1W 1M 1Y
EURTD 86.6400 -0.1425 14:59 31 may
USDTD 81.0625 +0.1450 17:44 31 may