Meta platforms technologies, llc (20240112687). GENERATING AUDIO FILES FROM TEXT INPUT simplified abstract

GENERATING AUDIO FILES FROM TEXT INPUT

Organization Name

meta platforms technologies, llc

Inventor(s)

Yaniv Nechemia Taigman of Raanana (IL)

Felix Kruk of Rehovot (IL)

Yossef Mordechay Adi of Rishon Le Zion (IL)

Gabriel Synnaeve of Paris (FR)

Adam Polyak of Tel Aviv (IL)

Uriel Singer of Harish (IL)

Devi Niru Parikh of San Francisco CA (US)

[[:Category:Alexandre D�fossez of Paris (FR)|Alexandre D�fossez of Paris (FR)]][[Category:Alexandre D�fossez of Paris (FR)]]

Jade Copet of Paris (FR)

GENERATING AUDIO FILES FROM TEXT INPUT - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240112687 titled 'GENERATING AUDIO FILES FROM TEXT INPUT

Simplified Explanation

The patent application describes methods, systems, and storage media for generating audio data by encoding text input and representative audio sources into audio tokens and text representations, respectively. The relationship between audio tokens and text representations is determined to reconstruct audio sources.

• Receiving text input and representative audio sources
• Encoding audio sources into audio tokens and text input into text representations
• Mapping audio tokens to text representations to determine relationship scores
• Decoding audio tokens to reconstruct audio sources

Potential Applications

This technology can be applied in speech recognition, language translation, audio editing, and voice synthesis.

Problems Solved

This technology solves the problem of accurately mapping audio tokens to text representations to reconstruct audio sources effectively.

Benefits

The benefits of this technology include improved accuracy in audio data generation, enhanced speech recognition capabilities, and efficient language translation processes.

Potential Commercial Applications

The potential commercial applications of this technology include speech-to-text software, language translation services, audio editing tools, and voice-controlled devices.

Possible Prior Art

Prior art in this field includes speech recognition software, language translation algorithms, and audio editing tools that may have similar functionalities.

How does this technology compare to existing speech recognition systems?

This technology offers improved accuracy in audio data generation by mapping audio tokens to text representations, which may result in more precise speech recognition compared to traditional systems.

What are the potential limitations of this technology in real-world applications?

One potential limitation of this technology could be the processing power required to map audio tokens to text representations in real-time applications, which may impact the efficiency of the system.

Original Abstract Submitted

methods, systems, and storage media for generating audio data includes receiving a text input. the method also includes receiving a plurality of representative audio sources and encoding the plurality of representative audio sources into a plurality of audio tokens. the method includes encoding the text input into a plurality of text representations. the method comprises mapping each audio tokens of the plurality of audio tokens to a text representation of the plurality of text representations. the method also comprises determining a relationship score based on mapping each audio tokens to the text representation, wherein the relationship score identifies a distribution of audio tokens from the plurality of audio tokens. the method and systems can also comprise decoding the subgroup of audio tokens to yield a reconstructed audio source.