How does speech recognition turn a voice into text?
Short answers to 8 questions from The Encyclopedia of How. Each answer is the opening of the full entry in the book.
How does speech recognition turn a voice into text?
Modern speech recognition converts an audio waveform into features or learned representations that capture useful patterns in speech. Neural networks then estimate likely speech units, characters, word pieces, or words from the changing sound.
How does text-to-speech technology create a humanlike voice?
Modern text-to-speech systems first analyze written text so abbreviations, numbers, punctuation, and words can be pronounced appropriately. The system represents linguistic content and predicts timing, emphasis, pitch, and other features that shape natural prosody.
How does a recommendation system predict what someone might like?
The system collects data about user behavior. It finds patterns and similarities. It may use collaborative filtering: users who liked similar things. Or content-based filtering: items similar to what the user liked. It combines these to make recommendations. Feedback improves future suggestions.
How does encryption protect private information?
Encryption transforms readable data into ciphertext using a cryptographic algorithm and a key. Properly designed encryption makes the protected information computationally impractical to recover without the appropriate key, even if an attacker obtains the ciphertext.
How does two-factor authentication improve account security?
Multi-factor authentication requires evidence from more than one independent category, such as something a person knows, something they possess, or a biometric characteristic. A password plus a one-time code from an authenticator app is a common example.
How does ransomware prevent someone from accessing their files?
Ransomware prevents access mainly by encrypting valuable files with cryptographic keys controlled by the attacker. Some variants also lock systems, delete accessible backups, or steal data before encryption so attackers can threaten to publish it.
How does a television station broadcast pictures and sound?
A TV station combines video and audio signals. It modulates a carrier wave. The signal is broadcast from a tower. TV antennas receive it. The TV demodulates the signal and displays the picture and sound. Digital TV uses more efficient encoding.
How does a video game respond almost instantly to a player's actions?
The game loop runs continuously. It reads input, updates the game state, and renders the frame. This happens many times per second, often 60 or more. Low latency is crucial. Controllers and displays are designed for fast response. The result feels immediate.
These are short versions. The full entries, and the rest of the book, are in The Encyclopedia of How.