I590/N564/N364 QUIZ #3 - 8 March 2006	(PERFECT SCORE: 40 POINTS)

                                                 Name _____________________________
													
		
1. I argued that information is usually lost or distorted in going from
	one music program to another. One reason is conversion from one encoding to
	another one that might not be able to represent the information. But that's not
	the only reason; in fact, information is often lost or distorted even if the
	programs use exactly the same representation.
6>	(a) Why might that happen?

	BECAUSE EVEN IF SOME KIND OF INFORMATION IS IN THE FILE AND A PROGRAM CAN HANDLE
	IT, IT MIGHT NOT DO IT CORRECTLY--OR MIGHT NOT TRY. (LOSSY COMPRESSION ISN'T
	EXACTLY THIS PHENOMENON...UNLESS SOMEONE CONVINCES ME IT IS :-) .)


4>	(b) Regardless of the reason, assuming both programs use the same basic
	representation, for which one (audio, time-stamped events, notation) is
	information getting lost or distorted most likely?
		
	NOTATION. IT'S SO MUCH MORE COMPLEX IN TERMS OF EXPLICIT STRUCTURE, IT'S MOST
	LIKELY THAT A PROGRAM DEALING WITH NOTATION WILL EITHER HAVE BUGS OPENING OR
	SAVING INTERCHANGE (NB!) FILES OR WON'T IMPLEMENT ALL THE FEATURES OF THE
	ENCODING. (WITH LOSSY COMPRESSION, CHANGES IN THE INFORMATION WILL ALWAYS BE
	SUBTLE; WITH NOTATION, OFTEN NOT.)


	
10>	2. Lossy compression employs many techiques, but all are based on "perceptual
	coding", i.e., on principles of psychoacoustics. Name and describe one technique
	that was mentioned in class.

	* MASKING ("SHADOWS"), VIA AMPLITUDE WITHIN CRITICAL BANDS
	* THRESHHOLD OF AUDIBILITY (VARIES WITH FREQUENCY)
	* REDUNDANCY AMONG CHANNELS (5.1 EVEN MORE THAN STEREO!)
	* ANY OTHERS?



5>	3. Here are excerpts from two encodings of music.

		Encoding A:		<staff def="s2">
								<chord dur="8" dots="1">
									<note pname="a" />
									<note pname="f" oct="5" />
								</chord>

		Encoding B:		GS, K2$K, 3=4, WMFW, 8C+, 8D+, 8E+, 8D+, /, 4B,	

	One of the above is an XML encoding; the other is not. Which one is XML, and
	how can you tell?

		ENCODING A IS XML: IT USES MATCHING OPENING AND CLOSING TAGS (LIKE
		XHTML, BUT *NOT* REGULAR HTML).


	
10>	4. The Byrd and Isaacson "Music Representation Requirements Specification"
	says that tuplets (triplets, quintuplets, etc.) are exceptionally complex
	symbols in music notation, and I argued in class that they're a major source
	of problems in representing CMN, citing examples in music by Bach and Schubert.
	How and why do they cause problems?
	
	TUPLETS CAN BE UNMARKED ("INVISIBLE")--THEY'RE PRESENT, ALTHOUGH NOT VISIBLE
	ON THE PAGE. BUT IT'S SUBTLE: UNMARKED TRIPLETS ARE OBVIOUS IN THE SCHUBERT
	IMPROMTU, BUT NOT ALWAYS OBVIOUS AT ALL. ANOTHER REASON OCCURS IN THE BACH
	EXAMPLES: THE NOTATION CHANGES TIME SIGNATURE (AND TEMPO) TO AVOID TRIPLETS,
	BUT THE RESULT IS EXACTLY LIKE TRIPLETS.



5>	5. I said that every OMR program ignores some types of symbols--or, more
	precisely, each program _tries_ to ignore some types of musical symbols. What
	did I mean by that?
	
	THE PROGRAM MIGHT MISTAKE THE SYMBOL IT WANTS TO IGNORE FOR ONE (OR PART OF
	ONE!) IT WANTS TO PAY ATTENTION TO.