Read Byrd, D., & Crawford, T. (2002). Problems of Music Information Retrieval in the Real World. Information Processing and Management 38, pp. 249-272; available at http://mypage.iu.edu/~donbyrd/Papers/RealWorldMusicIR35TR.pdf . Have ready one question the article raises that you can answer and one question it raises that you can't; I will attempt to answer the questions you can't :-) .

Homework: Our class includes people majoring in Informatics, in Music, and in Telecommunications. What is a realistic way in which people in different areas might use music searching by content in their career, not just for fun? Write a paragraph for yourself and a paragraph for someone in a different field. Which of the three basic forms--audio, MIDI, notation--would be most useful for that person (the answer can be two or even all three)? Obviously you know more about your own field than anyone else's; just do your best for the other one.

Note this is a written assignment, to be handed in, but I'll ask some of you to read one of your answers, and we will discuss them.

Homework: Go to Rainer Typke's web site, "MIR Systems: A Survey of Music Information Retrieval Systems", at http://mirsystems.info/ . It has descriptions of and links to dozens of music-IR systems, a few of which can be used over the Web for free. Choose two or three of them, and use each to try to retrieve something you'd expect (from the description of the system and its database) it to be able to find. If you don't find it immediately, try to figure out why: i.e., is it really not there, or did a problem with your query or something else result in it not being found. As you do, take notes--the verbal, not the musical, kind :-) --and whether you were eventually successful or not, be prepared to discuss what happened. (Your notes are just for the discussion, not to be handed in.)

Bring in a recording of one of your favorite pieces of music. No holds barred; the wilder the variety, the better! To do this as efficiently as possible, get me an audio file in a standard format--MP3, if possible; AAC should be okay too--by noon Thursday. You can get it to me by e-mail, if it's not too large a file; by putting it on the Web or an FTP server in a place I can get to, and telling me how to get it; by giving me a CDROM containing it; or, if it's in the Music Library's Variations2 system, by telling me how to find it there. Also give me some information about the music, including (1) a title; (2) one or more (style) genres you personally think it belongs to, and the genre(s) you think or know it belongs to according to either All Music Guide (http://www.allmusic.com/) or iTunes (http://www.itunes.com/); and, if applicable and the information is available, (3) the composer(s), date of composition, performer(s), date of performance, and performance medium (e.g., solo kazoo, vocal quartet, turntabilism, software synthesis, etc.). Of course a lot of those items don't make sense for a lot of music. We'll listen to a minute or (hopefully) two of everyone's selection (we can start in the middle if you feel the beginning of the track doesn't get the idea across), and we'll briefly discuss each of them in terms of what characteristics it has that you could plausibly get a computer to recognize. This is required of those registered for credit, but everyone is encouraged to participate, including people sitting in, the TA, and the instructor.

Homework: examine my Similarity Scale for Content-Based Music IR, available from my home page. Then write (to be turned in) 300 to 500 words or so about how differences in the desired task--say, find instances in relationship category #1 vs. #3 or #4--might affect a music-IR system trying to identify some of the music we listened to on Friday. You'll recall we had a brief discussion of just this point with respect to applause. That is, a certain feature of a recording might make things easier for a system trying to identify exactly the same recording in its database (category #1), whereas that same feature might make things more difficult for a system trying to identify ANY recording of the same music (category #3) or a system trying to identify any recording, MIDI file, or score of the same or very similar music (category #4). (A slightly-improved version of what I handed out on Friday is now linked at the I590 web page.) Also read AKoff Sound Labs' What is Music Recognition?; WAV and MIDI Formats, at http://www.akoff.com/about.html; and take at least a quick look at Castan's Music Notation Links: Musical notation codes, at http://www.music-notation.info/en/compmus/notationformats.html . The Akoff web page is short and simple, and you can read it straight through. The Castan page is long and complex and includes numerous links. You won't be able to follow all of them unless you have an extra week in the next few days, but even a few minutes with it should give you some feeling for the number and variety of music encodings there are.

Prepare for a quiz on what we've covered so far.

We'll probably spend most of the class going over the recent homework and quiz.

Read Selfridge-Field, Eleanor (1997). Describing Musical Information. In Selfridge-Field, E., ed. (1997) Beyond MIDI: The Handbook of Musical Codes, pp. 3-38. MIT Press. On reserve in the Music Library, ML74 .B49. (We will probably be reading material in this book again, and it's a useful reference; you might want to buy a copy. However, be aware of its age: it predates the invention of XML, much less XML encodings of music.)

(No assignment.)

Read more about MIDI, e.g., in the IU Center for Electronic and Computer Music's Introduction to Computer Music (http://www.indiana.edu/~emusic/etext/toc.shtml), Chapter Three (this has some rough edges, but it's useful anyway); in the Selfridge-Field "Beyond MIDI" book; or in F. Richard Moore (1988), The dysfunctions of MIDI, Computer Music Journal, 12 no.1, pp. 19-28 (available in the Music Library). Write (to turn in) a few hundred words on the specific limitations of MIDI for a specific kind of music you're interested in. Would this kind of music need more channels? Better control of the characteristics of notes (timbre, dynamics, microtonal pitch, etc.) while they're being played? Additional structural features like note spelling or expression marks? Or would MIDI actually be completely adequate for it? (The next hands-on assignment is going to have to wait.)

Do this hands-on MIDI assignment. It shouldn't take too long, but it requires some hardware you may not have at home! Also look at the list of Domains and Topics in the Call for Submissions for ISMIR 2006; think about how these topics relate to topics we've talked about in class and topics in our Syllabus, and what you'd especially like to see us cover. A lot of the ISMIR list probably won't make sense to you, but much of it will. This is just as background for discussion in class. (I'm adding this to the assignment rather late--it's almost noon Sunday now--so if you don't get around to it, don't worry.)

Read the Wikipedia article Musical Acoustics, and the IU Center for Electronic and Computer Music's An Acoustics Primer. There'll be nothing to turn in.

Read in The MP3 and Internet Audio Handbook, Chapters 11 (A Digital Audio Primer) and 12 (Digital Audio Formats). (Caveat: this was published in 2000, so it's somewhat out of date; the "systems for music delivery" they list are all dead, as far as I know, and if they were writing now, surely they'd mention iTunes, etc. But the technical information is still quite accurate.) There'll be nothing to turn in.

Do this additive-synthesis assignment. As I said in class, it requires a Mac running OS X, definitely at least v. 10.3 and quite possibly 10.4. Also read in The MP3 and Internet Audio Handbook, Chapter 13 (MPEG Audio). (Caveat: again, this is a bit old. MPEG-7 was released several years ago.)

Prepare for the quiz. No other assignment.

Read Hall, Gary (2002, April). Cramped Quarters. Electronic Musician 18,5, pp. 58-74. (This is on my "personal reserve". Go to the Music Library circulation desk and say it's on personal reserve--I'm not sure if they'll want my name or the N364/N564 course number or both. If it's in use, the Music Library probably has their own copy, but I haven't checked.)

Write up a description of (a.k.a. proposal for) your project of two pages or so. Be as specific as possible! Talk about why you're going to do it (my requiring you do a project isn't a good enough reason, sorry) as well as what you're going to do. If you're doing an experiment of some kind, you should probably also say something about what you expect and/or hope the outcome will be. If it takes you three or even four pages, that's OK. And look at the revised "Guidelines for Writing and Rubric for Grading": they now discuss proposals like this one.

Read Byrd, Donald, & Isaacson, Eric (2003). A Music Representation Requirement Specification for Academia. Computer Music Journal 27(4), pp. 43-57; revised (2005) version (PDF, 228K). You can just skim the huge table that constitutes the last half of the article, though I think it's pretty interesting, myself. Also read Byrd, Donald (1994). Music Notation Software and Intelligence. Computer Music Journal 18(1), pp. 17-20; TIFF scans of the individual pages are p. 17, p. 18, p. 19, and p. 20.

Read The GUIDO Music Notation Format - A Novel Approach for Adequately Representing Score-level Music, available from the GUIDO Documentation website. Also spend a few minutes trying out the remarkably simple (for simple music, that is!) GUIDO language with the GUIDO NoteServer.

Choose some music you're interested in and encode a "reasonable" version of part of it, either in MIDI form (with a sequencer) or CMN (with a notation program). Then, if you encoded it as MIDI, convert to CMN, then back; if you encoded it as CMN, convert to MIDI, then back. The CMN version should be about a page. Print out all three forms and bring them in, plus brief comments on how accurate or inaccurate your first version is. On the printout of the third version, add comments on what happened since the first, circling things, etc. For MIDI versions, print the sequencer's pianoroll view; if it doesn't have a print command, a screenshot of part of it will do. Please put "1", "2", or "3" at the top of each printout so I don't have to guess.

While I say "choose" music, I'd really prefer you use the music you brought in for the "our favorite music" session. So, if you choose something else, also write me an explanation of why. I'm well aware that some of "our favorite music" would be very difficult to encode accurately; that's why I'm asking you to encode just a "reasonable" version--in some cases, I'm sure that will be hugely simplified from the original. I'm also aware that some of you don't have much background for this kind of thing. Don't worry :-) .

Revise your project proposal. Find ways to be more specific than your first version. Any numbers (approximate, of course) would help; you could also mention music or IR programs you might use. Look again at the re-revised "Guidelines for Writing and Rubric for Grading". Every proposal should have a title; most, if not all, should have one or more references.

Read a late draft of Byrd & Schindele (2005), Prospects for Improving OMR with Multiple Recognizers.

Scan in to a computer a page of music each from two different publications from different publishers; alternatively, scan in one page yourself and get an image of a second page from Variations2, CD Sheet Music, etc. Either way, I recommend a resolution of about 300 dpi. Then use OMR to get the two pages into a notation program. Unless you have access to another scanner and know what to do, use the scanner in the computer cluster on the 3rd floor of the Music Library. Instructions for getting music into Finale or Sibelius are posted on the wall of the carrel with the scanner. (One misleading statement in the instructions is about cropping: instead of what they say, select the area you want, then give the Image>Crop command.) If you're using Finale, once you've saved the scanned file, use its "Scanning: SmartScore" or similar command. Whatever notation program you use, print the results from it; turn in the printouts, plus photocopies of the two original pages, with some comments on what went wrong in the process. NB: Use any music you want (scores, parts, lead sheets, etc.), but the better the graphic quality, the better. No matter how good the graphic quality is, you shouldn't expect great results :-) ; you can get full credit regardless of how good the results are.

Study for the quiz. I didn't have to tell you that, but I do have to tell you what to study specifically: lossy compression, XML encodings of music, tuplets in music notation, converting between encodings, and OMR. That's all you'll need to know. There may be a question for which you'll have a choice of one or more of those topics and something else we've studied since the last quiz--maybe defining your choice of several terms.

No assignment.

Do whatever makes you happy.

Read Wang, Avery (2003). An Industrial-Strength Audio Search Algorithm. In Proceedings of the 4th International Conference on Music Information Retrieval (ISMIR 2003), Baltimore, Maryland, pp. 7-14. Here is a text-only version. This is about Shazam and how it works; Wang's presentation version of the paper, with the audio examples is available (at http://ismir2003.ismir.net/presentations/Wang.PDF), but it's over 5 MB!--but don't worry; I'll play some of the examples in class. Also, if you have a chance (this isn't required for Monday but probably will be later on), check out one or more music-recommender sites: Last.fm, Pandora, FOAFing the Music, MusicStrands, etc.

No assignment.

Read Pickens, J., et al (2002), Polyphonic Score Retrieval Using Polyphonic Audio Queries: A Harmonic Modeling Approach. Also, write down what you think of as two musical ideas in the music you gave me for the "our favorite music" day. Describe the ideas as clearly as as you can (feel free to use CMN or guitar tab), and say where in the piece each first occurs. If you've forgotten what the music was, the list is on the course web page.

Read Voorhees, Ellen (2002), Whither Music IR Evaluation Infrastructure: Lessons to be Learned from TREC and read Downie, J. Stephen et al, The 2005 Music Information Retrieval Evaluation Exchange (MIREX 2005): Preliminary Overview.

Look at the MIREX 2005 Contest Results web site pages, especially for Audio Artist Identification; try to figure out what a "Confusion Matrix" means.

Study for the quiz:

• Searching and filtering by content, metadata, and collaboration
• Shazam and audio fingerprinting
• Relevance, queries, and information needs; luck in searching
• Musical ideas, relevance, and relevance judgments
• Music-IR evaluation: precision & recall; the Cranfield model; MIREX
• Finding themes by letter name or contour
• Efficiency & indexing in music searches

Read Typke, Rainer, Wiering, Frans, & Veltkamp, Remco C. (2005). A Survey of Music Information Retrieval Systems, available at http://ismir2005.ismir.net/proceedings/1020.pdf . (You'll recall that early in the semester we looked at Typke's related web site, "MIR Systems: A Survey of Music Information Retrieval Systems"; see above for the URL.) This is an overview of 17 existing systems for content-based retrieval of music in both audio and symbolic forms. It includes a "map" of the systems showing the tasks and users for which each system seems most appropriate; note that the horizontal axis, showing tasks, has much the same idea as my "Similarity Scale for Content-Based Music IR". (a) Print out the page that contains the "map" (p. 158); then write as clearly as you can a number from 1 through 7 next to each system on the "map", saying which of my categories you think it fits best. For example, if you think Shazam fits category 6 (Music in same genre, style, etc.) best, put a "6" next to Shazam. (b) The "artist (performer/composer)" label on their horizontal axis doesn't have an obvious equivalent on my Similarity Scale. Write a paragraph about this difference arguing that their way makes more sense; then write a paragraph about it arguing that my way makes more sense. See if you can make a strong argument both ways!

No assignment.

Try out two or more music-recommender sites--for example, FOAFing the Music, Last.fm, MusicIP Mixer, Musicmatch Jukebox, MusicStrands, Pandora, etc.--and write a comparison of them. (Note: since Justin's guest talk is on exactly this subject, there's a good chance I'll add some specifics to the assignment after we hear him.)

No assignment.

The "Music as different as possible" assignment. There will be three teams of three people, each with one MIME major, one music major, and one computer science/informatics major. For purposes of this assignment, we'll pretend Will is a music major. Initially, let's make the teams:
Team A: Kurt Kneller, Greg McCracken, John Palmer
Team B: Charlie Hoyt, Kris Lou, Kyung Ae Lim
Team C: Kris Kurtis, Will Pierce, Joey Morwick
...but you can switch around; just send me an e-mail when/if you do. Each team will choose six pieces of music, each of which is as different as possible from the other five. You can use things from "Our Favorite Music", but for no more than two of your team's six. Then create a table showing the differences among all six. You might also want to write paragraphs about some or all of the pieces, clarifying the ways in which each is different from all the others. Turn in what you've written and recordings of all the pieces. (Here's some background information: "Music As Different As Possible", including helpful ideas on where to find strange music.)

No assignment.

No assignment.

No assignment.

No assignment.

Everyone, send me a title for your presentation. Also, each team, send me a version of the "Music As Different As Possible" spreadsheet in which -- for your rows -- you've filled in the blanks and corrected anything that needs correcting. (And speaking of genre classification, as we were Wednesday, I'm having trouble believing that Keibairaku No Kyu is reggae! Where did that label come from, iTunes?? Oh well.)

No assignment.

No assignment.

Send me your final written projects. Please e-mail them well before midnight. I know I said I wanted them in printed as well as electronic form, but let's forget the printed bit and save everyone the trouble. IMPORTANT: as you may know, the IU e-mail system doesn't accept .zip files as attachments; if you want to send a .zip, you can get around that just by changing the extension to (say) .zap before sending. (On OS X, I strongly recommend doing that from a terminal window; if you do it via the Finder, it'll be "helpful" in a way you probably won't like.) Also, if what you have is really big -- more than 4000-5000K -- it'd be better to get it to me some other way; slashtmp is probably the easiest. Or you could give me a CDROM. Regardless, you're responsible for making arrangements to get your stuff to me by midnight today.