Speech Recognition (Technology, Market)

Market in Russia

2024

10 neural networks for voicing texts and generating voices that work in Russia

Modern neural networks are able to generate high-quality speech, which in some cases can be difficult to distinguish from human. Such artificial intelligence-based tools are in demand in many areas. They are used when creating voice chat bots, in text reading programs and navigation systems, in applications for people with disabilities and in programs for professional voice acting. TAdviser has prepared a list of ten neural networks for generating votes that work in Russia.

1. SteosVoice

This tool provides ample opportunities for creativity and content creation. Each user gets free limited access to high-quality voice AI with more than 400 voices of all kinds. Speech synthesis in the form of a Telegram bot provides a convenient and fast way to convert text messages into an audio format. Key features of the service:

• Voice acting of books, articles, presentations, short videos, etc.;
• Hundreds of quality votes;
• Sound at studio recording level (44.1 kHz);
• Free bot in Telegram;
• Voice monetization capability;
• Variable narrative speed.

2. Lovo.ai

It's a highly realistic AI-based voice generator. Over 500 votes are available in 100 languages. The service allows you to create voice-over videos for marketing, training, social networks and other purposes. The easy-to-use user interface makes voice-over generation an easy task even for those with no audio experience. A free tariff is available with the ability to voice five minutes a month. Platform functions:

• Synchronization of audio and video;
• Automatic subtitle generator;
• Voice cloning;
• Creating images without copyright;
• Universal API;
• Genny's AI assistant for scripting.

3. Zvukogram

The service supports more than 10 languages, including Russian. It is possible to voice texts for video platforms, advertising, presentations, news and other scenarios. The service has an understandable user interface with hints and works on a token system. Opportunities:

• Creating dialogue and multilingual voice acting;
• Over 1,000 votes;
• Processing of long texts;
• Generation of a series of files;
• Compatibility with the installation software.

4. FreeTTS

This fully free resource provides an unlimited number of generation attempts. The system is extremely easy to use - just insert the source text into the dialog box and select the preferred voice of the announcer. At the same time, the developers admit, the service is somewhat inferior in quality to paid commercial platforms. Functions:

• About 30 Russian-speaking votes;
• The ability to play the result on the site;
• Loading the final file in MR3 format;
• Processing up to 2000 characters;
• No payment.

5. Robivox

A service for voicing text with a realistic voice created on the basis of a real recorded speech of the announcer. Without registration, you can process text up to 100 characters long. In paid mode, about 100 minutes of voice acting with a regular voice and 100 minutes of Pro voice are offered for 20 rubles. Opportunities:

• Support for more than 100 languages;
• Speed adjustment;
• Simple user interface;
• Support for the formats of generated MR3 and WAV files;
• Work with and without registration.

6. PlayHT

The extensive library of AI voices of this service covers all major languages ​ ​ (including Russian) and accents of the world. You can save the speaker's voice and his native accent when translating and dubbing into other languages. Emotional styles of speech are available.

• More than 800 natural-sounding voices;
• Context-sensitive, emotional and expressive models of text-to-speech conversion;
• Voice cloning;
• Multilingual speech synthesis;
• Online text-to-voice studio.

7. Deepgram

The voice AI platform provides a program interface (API) for converting speech into text as well as text into speech. A free loan for $200 is available, and paid services are provided using the Pay As You Go model (payment as consumed).

• Support for several AI models;
• Unified API;
• The ability to integrate voice AI into your own applications;
• Several dozen languages, including Russian.

8. Murf.ai

The platform provides a choice of more than 200 votes in several dozen languages, including Russian. The service offers the ability to fine-tune various aspects of the generated voice, including pitch, speed, pronunciation, pauses and accent, making it more natural. Opportunities:

• Voice cloning;
• Voicing of video materials;
• Creating multiple versions of voice-over;
• API for integration into various applications, websites or other services;
• User-friendly interface.

9. Speechify

This system is capable of working with various documents, including PDF files. Using a mobile application, you can take a picture of any page, and then convert the text into speech. About 60 languages ​ ​ are supported, including Russian.

• Naturally sounding human voices;
• Over 200 votes;
• Ability to adjust speech speed;
• Integration with Google Drive and Dropbox.

10. Synthesys

The service offers realistic synthetic voices in more than 140 languages. There is free access, and the cost of paid subscriptions starts at $20 per month. You can use the system for professional voice acting and video.

• Voice cloning with AI;
• Avatars for converting text into speech;
• Support for more than 400 voice options;
• Intuitive user interface.

In Russia, developed AI for simultaneous speech translation

Russian developers presented on November 11, 2024 a new artificial intelligence technology for simultaneous translation between four languages: Russian, English, Chinese and French. The system will be first applied at the IV Congress of Young Scientists in the federal territory "Sirius" on November 27-29, 2024.

According to "Наука.рф," access to translation will be carried out through a system of QR codes in the halls of the business program, which will allow participants to quickly choose the necessary translation language in real time.

In Russia, developed AI for simultaneous speech translation. It will be used at the congress of scientists in Sirius

Advisor to the President of Russia Anton Kobyakov stressed that the successful practices of the Congress of Young Scientists will be scaled up to other events for the convenience of our foreign guests.

More than 500 applications were submitted to the congress from representatives of the scientific community from the BRICS countries, including Brazil, China, India, South Africa, as well as from Germany, France, Switzerland and other states.

In addition to the artificial intelligence system, the event will feature 250 volunteer translators from leading Russian universities who speak English, Arabic, Bengali, Chinese, Portuguese and other languages.

The exhibition stands of the congress will be equipped with QR codes with links to the description of expositions in Russian, English, Arabic, Chinese and Portuguese. The event website is already available in Russian, English and Chinese versions.

The technology is designed to facilitate international communication and make the exchange of scientific knowledge between specialists from different countries more accessible. The development is part of a program to develop domestic artificial intelligence technologies.

In the future, it is planned to expand the functionality of the system and add new languages, including Arabic and Portuguese, which will reach a wider international audience.

The Congress is the main annual event of the Decade of Science and Technology in Russia, announced by Russian President Vladimir Putin for the period 2022-2031.^[1]

2020: The volume of the conversational AI market in Russia is $44 million ($76 million, including government orders)

Just AI, a company specializing in technologies for conversational artificial intelligence, machine learning and understanding of natural language, on August 16, 2021 presented its forecasts for the development of the conversational AI market until 2025, compiled based on the results of the^[2].

Analytics covers the tools and platforms of conversational AI - technologies for speech synthesis and recognition, voice cloning, speech biometrics, voice activation, natural language understanding and generation platforms, tools for visual development of dialog scripts in voice or text channels, speech analysis platforms, as well as solutions for outgoing calls and in the field of custom voice assistants for business, skills for smart devices and meta-assistants (Alice, Marusya, etc.), incoming telephony and smart IVR, development of bespoke chatbots.

The volume of the Russian market in 2020 amounted to $44 million or $76 million, taking into account government orders. The industry adds 46-93% of the year, the total growth since 2015 was 1288%. According to Just AI forecasts, by the end of 2021, the market volume will reach $80 million or $120 million, taking into account government orders. In the next five years, the industry will maintain growth dynamics from 38% to 81% and in 2025 will reach $561 million (excluding government orders).

"More than 100 companies operate in the conversational AI market in Russia, many of them grow by 200-400% per year. They do not always compete with each other: a significant part of the players specializes in individual industries, types of customers and technologies and can dominate their segments, even having a small share in the market as a whole, "said Just AI Managing Director Kirill Petrov.

In the revenue structure of the CST group of companies of 1 billion + rubles per year, more than 80% is occupied by income from government contracts. Just AI with revenue of 500 million + rubles. focuses on the segments NLP (Natural Language Processing )/NLU (Natural Language Understanding )/DM (Dialog Management) -playforms, No-code/Low-code designers and custom voice assistants. In the group of companies with revenue of 200 million + rubles. per year presented Yandex.Cloud (speech technologies), 3iTech (solutions for the public sector, speech technologies and speech analytics platforms) and Aero PBX (solutions for the public sector, outgoing telephone communications).

The largest segments on the market in 2020 were speech technologies (speech synthesis and recognition, voice cloning, speech biometrics, voice activation) and NLP platforms (natural language processing). Business and NLP platform solutions grew fastest in 2020.

According to Just AI forecasts, in five years half of the entire Russian market will be occupied by conversational AI solutions targeted at certain business tasks and industries, such as voice directory search for retail, virtual assistants for housing and communal services, chatbots for hotels. They will add 100-120% annually, medicine, HoReCa, e-commerce, tourism, the beauty industry, etc. are already showing interest in them.

According to analysts, requests for NLP platforms from large businesses will continue to grow for several more years. This will be due to the inclusion of new industries and the expansion of the scope of natural language processing. Visual designers for the development of bots with increasing interest from SMB companies will begin to actively grow and specialize in narrow tasks and the provision of ready-made templates and tools. The growth of Custom Assistants, Customer Support Solutions, Assistant Skills, Incoming IVR, Recruiting and HR Solutions will accelerate with the introduction of new developers and the engagement of new categories of SMB customers, and the growth of the Assistant and Smart Speaker market will be an additional incentive.

Outgoing telephone communications will continue to grow rapidly until 2022. Next, we should expect the introduction of legal regulation aimed at combating spam, and the widespread use of anti-spam technologies, which will lead to a fall in the market. After adapting to the new restrictions, it will be possible to grow the segment, possibly in new areas and industries, according to Just AI. Speech analytics in the coming years will face moderate growth, which may slow down with the development of NLP technologies and the refusal of contact centers from the staff. Speech technologies are experiencing an increase in consumption, but with an increase in the availability of models and data sets and the emergence of new players and inhouse developments, they will face significant price pressure.

Global market

2024: Live interpreter released for video conferencing. There is support for the Russian language

In mid-November 2024, the German company DeepL announced the introduction of the DeepL Voice function in its online translation system. It allows you to translate spoken language from one language to another in real time, which will be useful for personal communication and video calls. Read more here.

2023: Global Speech Translation Software Market Growth by 14% to $1.38 Billion

In 2023, the global market for speech translation software reached $1.38 billion. For comparison, in 2022, costs in this area were estimated at $1.21 billion. Growth was recorded at 14%, as stated in the Market Research Future study, the results of which are presented in mid-November 2024.

The sector under consideration is actively developing due to several factors. As businesses and organizations expand around the world, there is an increasing need for effective communication in different languages. This demand is further fueled by the development of a model of remote work and international cooperation. Companies invest heavily in translation technologies to improve customer engagement and user experience. The use of such tools in mobile applications, conference systems and support platforms is becoming more and more relevant.

Another driver of the industry is the achievements in the field of artificial intelligence and machine learning. The integration of such algorithms promotes more accurate speech recognition and translation, allowing systems to learn from previous interactions and improve over time. As machine learning algorithms become more complex, the ability to process slang and regional dialects improves, making the technology more convenient and reliable.

An increase in the number of voice-controlled devices and the development of the smart home concept are having a positive impact on the industry. Consumers are increasingly looking for solutions that integrate easily into their daily lives. At the same time, the popularity of social networks creates a demand for real-time translation to improve communication between users from different countries.

By the type of analytics systems, machine learning tools, natural language processing tools, cloud platforms and local solutions are distinguished. In 2023, the first type of products accounted for $0.45 billion. Natural language processing software brought in $0.38 billion. Revenue from cloud and local solutions is estimated at $0.32 billion and $0.23 billion, respectively. Significant players in the industry are named:

• IBM;
• Amazon;
• Google;
• Nuance Communications;
• Baidu;
• Unbabel;
• Lingmo;
• SAP;
• iFlytek;
• Apple;
• Microsoft.

Geographically, North America is the leader, providing revenue of $0.55 billion: the dominance of the region is due to the active introduction of advanced technologies and the need for communications in different languages. This is followed by Europe with costs of $0.4 billion, and the Asia-Pacific region closes the top three with $0.25 billion. South America brought in $0.1 billion, the Middle East and Africa - $0.08 billion.

The study states that companies recognize the importance of effective communication to interact with international customers, partners and customers. This contributes to the integration of speech translation technologies into different systems. Organizations that implement such solutions gain a competitive advantage, and therefore can increase sales.

At the end of 2024, revenue on the global software market for translating speech into another language is estimated at $1.58 billion. Market Research Future analysts believe that in the future, the CAGR (average annual growth rate in complex percentages) will be 14.29%. As a result, by 2032, costs on a global scale could rise to $4.6 billion.^[3]

See also

• Physician speech recognition
• Speech technology: On the path from recognition to understanding
• Speech synthesis

• Biometric Market Today
• Biometric identification (Russian market)
• National Biometric Platform (NBP)
• Unified Biometric System (UBS) of Bank Customer Data
• Biometric Identification (Global Market)
• Biometric identification technologies
• Facial Recognition Systems
• Biometrics myths
• A brief history of biometrics
• Biometrics and Border Control
• Biometric Time and Attendance Systems

Notes