Multimedia Search Tools

1
Multimedia Search Tools

• David Berry, Poppy Johnson-Renvall, Megan Meyer
  Kristi Mulhern

2
What is multimedia?

• Using, involving, or encompassing several media.
• Examples
• Audio (Music, sound effects)
• Radio
• Images
• Video

3
Searching Multimedia

• It is unlike searching for text on the web.
• Cannot search the full text of an image or song
  or video.
• At the very least you can search a file name.
• Content of the file cannot be searched unless a
  transcript is supplied.
• Difficult to find an exact spot in a multimedia
  file.

4
What is streaming media?

• Audio or video sent in compressed form over the
  Internet and displayed by the viewer as they
  arrive.
• With streaming media, a Web user does not have to
  wait to download a large file before seeing the
  Video or hearing the sound.
• The user needs a player, which is a special
  program that uncompresses and sends Video data to
  the display and audio data to the speakers.
• From All the Web

5
Audio

• Types of audio files
• .au
• .aif
• .lal, .lqt, .lsl
• .mid
• .mp3
• .ra, .ram, .rm
• .wav
• .wma
• Since these formats are always evolving and the
  files are so big, makers are always inventing new
  files. Currently there are more than 40 different
  kinds.

6
Audio

• Where you can find audio on the web
• FindSounds (sound effects) http//www.findsounds.c
  om
• Lycos Pictures and Sounds http//multimedia.lycos.
  com/
• MIDI Explorer (MIDI sound files only)
  http//www.musicrobot.com/
• The Music Finder (artists songs, not files)
  http//www.music-finder.net/
• Singingfish (audio/video only) http//www.singingf
  ish.com/

7
Singingfish

• Company began in 1999. Head office in Seattle.
• Designed with audio video metadata rather than
  textual metadata.
• Automated systems locate streams, extract
  internal metadata, then add their own metadata
  for indexing.
• Can locate MP3 music files, as well as QuickTime,
  RealMedia and Windows Media files.

8
Singingfish
9
Singingfish

• Advantages
• Comprehensive streaming audio available
• Great help pages
• Advance search options
• A few audio clips require user to go off site to
  locate them
• Disadvantages
• Cant get to Advance Search from home page. Must
  enter terms first, then options appear
• Pop-ups appear after playing a song

10
Lycos Multimedia

• Lycos was launched in 1994. In 1999 it stopped
  doing its own indexing and began using FAST's Web
  index
• March 2004, Lycos switched to Yahoo!'s Inktomi
  database

11
Lycos Multimedia
12
Lycos Multimedia

• Advantages
• Can search MP3s right from the home page.
  Singingfish requires going to the Advanced Search
  page
• Most search results usefully include suggestions
  for further search phrases
• Has recently developed MP3 only search engine
  http//mp3.lycos.com/ Better access to format
• Disadvantages
• Smaller index, fewer hits (eminem and mosh 1
  hit searching all multimedia. 0 hits searching
  for audio alone. Same query 1 558 574 hits in
  Singing Fish)
• No Advance help page. No truncation, or wildcards
  mentioned, but sometimes works
• Later pages frequently contain dead links

13
Findsounds

• Limited to simple sounds and short samples
• Used mostly for replacing the built-in event
  sounds in operating system effects.
• Sound types are grouped by category
• Most sounds use Quick Time or Windows Media
  Player
• Sounds like search feature copies the
  characteristics of a digital audio recording
  using sound matching technology called
  Comparisonics

14
Findsounds
15
Findsounds

• Advantages
• Sounds like search capability
• No blatant advertising or pop-ups
• Simple search interface
• Most sounds are in the site
• Disadvantages
• Only basic searching, with some limiters
• Most users probably rely on the types of sounds
  you can hear box.

16
Radio

• Most radio search engines are geared to talk
  radio, although a few include some musical
  programs.
• They usually link directly to streaming audio,
  rather than to downloadable files.
• Although some radio shows or stations have their
  own searchable websites (CBC, NPR), radio search
  engines are usually a third-party indexer of the
  content available on a variety of websites, much
  as standard text search engines are.

17
Common audio players for radio streams

• RealPlayer
• .ra, .ram, .rm, .rmm, .smi, .smil
• Winamp, other MP3 players
• .mp3, .pls, .m3u
• Windows Media
• .asx, .wma, .wmv, .wax
• Winamp 5
• .ogg, .pls
• Quicktime
• .mov, .qt

18
What do radio search engines do?

• Generally, they provide a way to locate specific
  programs, stations, guests or musical genre.
• Often more of an elaborate portal than a true
  search engine.
• Examples PublicRadioFan, Radio-Locator
• Notable exception SpeechBot, which creates
  transcripts and then searches the resulting text
  for search terms.

19
PublicRadioFan

• Contains links to the official web pages and
  audio streams of public radio around the world.
• Customizable time zone, favourite
  stations/shows.
• Created and maintained by a math and computer
  science teacher in his spare time.

20
PublicRadioFan
21
PublicRadioFan

• Advantages
• Uses extensive pull-down menus to locate radio
  shows
• Allows known-item search or browsing by genre,
  etc.
• Users can customize the site to reflect their
  location and favourite stations.
• Disadvantages
• Crowded site looks messy and is overwhelming.
• Does not allow user to input search terms, merely
  to select from predetermined menus.
• Vast grids and lists not user-friendly.

22
Radio-Locator

• Formerly the MIT list of Radio Stations on the
  Internet.
• Links to over 10,000 radio station websites.
• Links to over 2500 radio audio streams.
• Allows searching by call letters, location, or
  genre.
• Promises some global radio, but U.S.-centred.

23
Radio-Locator
24
Radio-Locator

• Advantages
• Clean, visually appealing design.
• Nice descriptions of genres/searches for similar
  stations.
• Information is generally very up-to-date
• Databases used include information from the FCC.
• Disadvantages
• Canadian content advertised but not delivered!
• Much information too technical (i.e. the science
  of radio broadcasts) and not useful to the
  listener.

25
SpeechBot

• Uses speech recognition technology to index the
  content of popular U.S. radio shows.
• Creates a rough transcript that is indexed and
  fully keyword searchable, although searching can
  sometimes be tricky due mis-transcribed words.
• Offered by HP/Compaq.
• Currently indexes 15316 hours of content.

26
SpeechBot
27
SpeechBot

• Advantages
• Fully searchable, using Boolean or guided search.
• Allows limiting by topics, dates, or stations.
• Clean, well-designed interface for basic and
  advanced searching (Power Search
• Shows snippet of text with bold keywords in
  results page to allow users to check relevance.
• Disadvantages
• Video search doesnt seem to work right.
• Searchable transcripts contain errors.

28
Images

• There are numerous websites and search engines
  that specialize in images, but most of them have
  small collections.
• Types of image files
• jpeg
• tif
• gif
• bmp

29
Image Search Engines

• Search engines do not own the images they find.
  These images are usually available over the
  internet.
• For obscure subjects a general search engine is
  best Google alone indexes 880 million pictures.
• Picsearch (www.picsearch.com) specializes in
  searching for images and animations.

30
Picsearch

• Interface available in English, German, Swedish,
  Norwegian and Danish.
• Limits by image/animation, color/bw, picture
  size
• Default AND, accepts NOT and quotation marks, no
  OR operator

31
Picsearch
32
Image Collections

• Images that are held by a single institution or
  consortium of institutions.
• May or may not be searchable in any organized
  fashion.

33
Getty Images

• Getty Images (www.gettyimages.com) sells
  professionally produced pictures and film.
• Limits by a number of different image
  characteristics (orientation, subject, style,
  concept, etc.)
• Clarification feature automatically requests
  limits on very broad searches.
• Default AND, accepts OR, NOT, ,

34
Getty Images
35
Images Canada

• Images Canada (www.imagescanada.ca) offers
  pictures from the holdings of libraries, archives
  and museums across Canada.
• Each image is catalogued using standard library
  metadata.
• Various special collections on topics in Canadian
  history.

36
Images Canada
37
Video

• Types of video files
• .avi
• Audio Video Interleave.
• Standard file format for Microsofts Video for
  Windows.
• .mov, .movie, .qt
• Quicktime files.
• .mpeg, .mpg
• Standard video platform for the Web.
• .rm
• Real Audio Video file.
• .rv
• Real Video file.

38
Video

• Where to find videos on the web
• IFILM (www.ifilm.com)
• Singingfish (www.singingfish.com)
• Lycos Multimedia (multimedia.lycos.com)
• Windows Media (www.windowsmedia.com)
• Altavista (www.altavista.com)
• AlltheWeb (www.alltheweb.com)

39
IFILM

• IFILM is a privately held media company based in
  Hollywood, CA.
• IFILM contains many videos divided into various
  categories.
• 56k 200k streams available / 500k available
  with membership.

40
IFILM

• Can search by
• Title
• Main theme or gist
• Name of a person associated with the clip
• Admit that searching the site can be clunky
• Point users to the Top 100 Channel if search
  produces unsatisfactory results.

41
IFILM
42
IFILM

• Advantages
• Many types of clips available
• Choice of players to use
• Encourage filmmakers to submit their independent
  films.
• They dont take themselves too seriously.
• Disadvantages
• Annoying Advertisements.
• Search function bites (their words)
• Can be difficult to find something specific
• No advanced search screen
• Video quality is not always good.

43
Windows Media

• Powered by Singingfish
• A portal for online streaming media using the
  Windows Media format
• Contains Movie trailers and clips and music
  videos.
• Can sort results by
• All media
• Music
• Movies
• Entertainment
• Current events

44
Windows Media
45
Windows Media

• Advantages
• 56k 100k 300k some 500k
• No ads at the beginning of clips
• International media guides
• Disadvantages
• There is no advanced search screen
• There is no help page in case you get stuck.
• Links to outside sites to get media.
• Does not do Boolean searching.

46
Altavista

• After typing in a query it retrieves video files
  that contain the search words in
• the file names
• ALT text
• nearby text
• the page metatags and/or are embedded in the file.

47
Altavista

• Advantages
• Can limit by file type or duration
• More info page (when theres actually info in
  it)
• Disadvantages
• There is no advanced search screen.
• No Boolean searching
• Takes you to a host site the user has to find
  the file on that site it may look differently
  than the thumbnail presented on Altavista. User
  has to remember the name of the video file.
• Does not contain a lot of trailers clips for
  movies coming out in the near future.

48
Resources

• Alltheweb Help pages http//www.alltheweb.com/hel
  p/faqs/video_search
• Altavista Help pages http//www.altavista.com/hel
  p/search/help_vid
• Compaq. Technical Whitepaper SpeechBot audio
  search using speech recognition. Available
  http//speechbot.research.compaq.com/SpeechBotWhit
  ePaper.pdf
• Getty Images Help pages http//creative.gettyimag
  es.com/source/home/homeCreative.aspx
• Harney, John. The Grammar of Sound New software
  lets you index and search audio much faster than
  in the past. Technology Review. April 30, 2003.
  Available http//www.technologyreview.com/article
  s/03/04/wo_harney043003.asp
• IFILM Help pages http//www.ifilm.com/help/
• Lycos Multimedia Help pages http//multimedia.lyc
  os.com/help/help_video.asp
• Notess, Greg R. Searching Beyond Text Issues
  with Multimedia Searching. Online 24. Sept./Oct.
  2000, p. 61-63. Retrieved Nov. 2, 2003 from
  Academic Search Premier.
• Sullivan, Danny. Multimedia Search Engines
  Image, Audio Video Searching Search Engine
  Watch. September 5, 2003 http//searchenginewatch
  .com/links/article.php/2156251