Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Query By humming - Music retrieval technology

2,096 views

Published on

For slide details , visit following link
http://www.slideshare.net/shitalkr/query-by-humming-music-retrieval-technique

Published in: Technology
  • Be the first to comment

Query By humming - Music retrieval technology

  1. 1. Music retrieval technique Shital Katkar 132011005
  2. 2. Index What is QBH? Basic Architecture Application Challenges File Formats System Architecture Parson code algorithm  Benchmarking MIR System
  3. 3. •“I don’t know the name. I don’t know who does it. •But I can’t get this song out of my head.” •Well, why not just hum it. QBH System Query By Humming
  4. 4. Basic Architecture Microphone Extraction Transcription Comparison Result List DB Fig- Basic System Architecture
  5. 5. Applications Shazam •identify pre-recorded music being broadcast from any source, such as a radio, television
  6. 6. Applications Sound Hound •identify music by humming, singing or playing a recorded track
  7. 7. Applications Midomi •identify music by humming, singing or playing a recorded track
  8. 8. Applications Musipedia identify music by whistling a theme, playing it on a virtual piano keyboard, tapping the rhythm on the computer keyboard
  9. 9. Challenges • Users may not make perfect queries. • Accurately capturing pitches and notes from user hums is difficult, even if the user manages to submit a perfect query. • Similarly, accurately capturing melodic information from a pre-recorded music file is difficult.
  10. 10. File Formats Wav File Format short form of the Wave Audio File Format the most common use is to store an uncompressed audio  quite large in size first generation files of high quality
  11. 11. File Formats MIDI File Format Musical Instrument Digital Interface MIDI files are not exactly the same as the typical digital audio formats we use (like WAV, MP3, MP4 etc.) a MIDI is made up of information that describes what musical notes are to be played MIDI Files therefore do not contain any 'real world' recordings
  12. 12. System Architecture
  13. 13. Parson Code Algorithm Algorithm A note in the input is classified in one of three ways 1. U = "up," if the note is higher than the previous note 2. D = "down," if the note is lower than the previous note 3. r = "repeat," if the note is the same pitch as the previous note 4. * = first tone as reference
  14. 14. Textual Pattern C C G G GA A U r F F E E D D C D r D r D r D* rUr D 72 72 79 79 81 81 79 77 77 76 76 7274 74
  15. 15. Introduction Music Information Retrieval (MIR) efficient content-based searching retrieval of musical information should be easily operated by users should be controlled by a simple-to-use graphical 'musical' interface
  16. 16. MIR System Problem Definition Lot of MIR System All have the same task- to enable users to search for music Very few systems that are actually publicly accessible and comparable Some System works only with MIDI representation, some with transcriptions Each system has a different set of files available in its database
  17. 17. Music Information Retrieval Methods MIR Systems may be divided into two categories 1. those that search symbolic representations of music MIDI files or Common Music Notation (CMN) 2. those that search raw audio files WAV or mp3 file format
  18. 18. Symbolic representations  consist of a list of instructions as to how the piece should be played include the notes, when and for how long each is played Typical Query – Involve a search for files with a given sequence of notes List of MIDI files from database Music Information Retrieval Methods
  19. 19. raw audio files digital representations of an actual recording contain a level of complexity that is not found in the symbolic representations composition is contaminated by noise Music Information Retrieval Methods
  20. 20. Two Approaches Extraction involves finding certain features, such as the mean and variance of audio signal Transcription convert the query into a symbolic representation Music Information Retrieval Methods
  21. 21. Online MIR Systems CatFind MelDex MelodyHound ThemeFinder Music Retrieval Demo
  22. 22. CatFind  search MIDI files using either a musical transcription or a melodic profile based on the Parson’s Code It has minimal features intended primarily for demonstration Online MIR Systems
  23. 23. MelDex  MELody inDEX allows searching of the New Zealand Digital Library designed to retrieve melodies from a database on the basis of a few notes sung into a microphone It accepts acoustic input from the user, transcribes it into common music notation, then searches a database for tunes that contain the sung pattern, or patterns similar to it. Retrieval is ranked according to the closeness of the match Online MIR Systems
  24. 24. MelodyHound developed by Rainer Typke in 1997 originally known as "Tuneserver" It searches directly on the Parsons Code was designed initially for Query By Whistling return the song in the database that most closely matches Online MIR Systems
  25. 25. Themefinder created by David Huron allows one to identify common themes in Western classical music, Folksongs of the sixteenth century Online MIR Systems
  26. 26. Music Retrieval Demo  performs similarity searches on raw audio data (WAV files) No transcription of any kind is applied It works by calculating the distance between the selected file and all other files in the database Online MIR Systems
  27. 27. Comparison Of MIR Systems
  28. 28. Evaluation Issues The coverage of the collection, that is, the extent to which the system includes relevant matter. The time lag, that is, the average interval between the time the search request is made and the time an answer is given. The recall of the system, that is, the proportion of relevant material actually retrieved in answer to a search request The precision of the system, that is, the proportion of retrieved material that is actually relevant.
  29. 29. Conclusion In this work, we have laid down a framework for benchmarking of future MIR systems. There are only a handful of MIR systems available online, each of which is quite limited in scope. Still, these benchmarking techniques were applied to five online systems. Proposals were made concerning future benchmarking of full online audio retrieval systems. It is hoped that these recommendations will be considered and expanded upon as such systems become available.
  30. 30. Thank you

×