Audio search engine

Types of audio search
– Audio search from text: User enters text into search bar, compared to database, matching results with descriptions and options to download.
– Audio search from image: Query by Example (QBE) system uses content-based image retrieval (CBIR), generates keywords from image, searches for audio files based on keywords and user preferences.
– Audio search from audio: User plays audio of a song, sound pattern derived from waveform and frequency representation, matched with patterns of sound files in database, displayed as search results.

Design and algorithms
– Audio search uses keywords found in media titles, attached text, and linked web pages.
– Some search engines can search recorded speech using phonetic search algorithms or text transcriptions.
– Applications like Munax and Shazam use ranking algorithms, acoustic fingerprinting, and databases to identify and match audio.

Notable audio search engines
– Deep audio search: Picsearch Audio Search powers image, video, and audio search for over 100 search engines.
– For smartphones: SoundHound (formerly Midomi) offers audio-based AI services and song identification capabilities. Shazam identifies songs by creating an acoustic fingerprint and comparing it to a central database.
– Doreso identifies songs by humming, singing, or direct input, providing information and options to purchase.
– Munax, a defunct company, released a multimedia search engine in 2005, allowing search for images, video, and audio in one place.
– Z39.50 is a protocol used for searching and retrieving information from databases.

Additional information on audio search
– Audio search has evolved slowly and uses various search formats, all relying on keywords.
– Some search engines can search recorded speech, while others use phonetic algorithms or full audio transcription.
– Shazam uses an audio fingerprinting technique called spectrogram to identify songs from various sources.
– Munax shut down after releasing their multimedia search engine.
– Citation needed for some information.

Notable features and capabilities of audio search engines
– Picsearch Audio Search powers image, video, and audio search for over 100 search engines.
– SoundHound (formerly Midomi) offers audio-based AI services and song identification capabilities.
– Shazam identifies songs by creating an acoustic fingerprint and comparing it to a central database.
– Doreso identifies songs by humming, singing, or direct input, providing information and options to purchase.
– Z39.50 is a protocol used for searching and retrieving information from databases.