Done
General
- Firmly settled on Didactyl (piano fingering) as my thesis project.
- Created blog at http://rndlph.blogspot.com to track the project. (This status report is actually the very first post.)
DataCollector
- Reorganized a room in my house for prototyping.
- Configured XCode and OpenFrameworks for development.
- Got MIDI interface and PS3 Eye camera working.
- Explored some sensor options to deal with occlusion of thumbs.
Corpora
- Exchanged several emails with Richard Parncutt.
- He expressed interest in collaborating on a conference paper. (I didn't even ask.)
- Obtained data (as a set of very old Word files) from sight-reading paper (Sloboda et al., 1997).
- Translated the Word files to text.
WebService
- Obtained Parncutt source code from Parncutt himself.
Doing
- Comparing Parncutt source to my Python code.
- Implementing basic color tracking synchronized with MIDI events.
- I need (I think) to obtain an IRB waiver to use the anonymized data shared by Parncutt. Any advice on this?
- Trying to obtain evaluation corpus used in (Parncutt et al., 1997) with a mind toward quantifying IAA.
- Thinking about ways to quantify similarity between fingerings.
- Thinking about ways not to quantify any such thing: maybe we measure the quality of a model by its ability to effect change in a player's preconceived notion of how a piece should be fingered.
Red Flags
- In this preliminary IAA exercise, I want to control for hand size. At a minimum, I would like to obtain gender information for each subject as a gross approximation of hand size. Would this complicate the IRB waiver?
- Ideally, we would be able to track this back to the 28 pianists who provided these data and obtain detailed hand measurements. This may be impossible to do completely, but we could probably do it partially. (10 of these subjects could be considered annotators and not human subjects. They were just asked to mark up the scores and return them. But the sight readers were videotaped and their fingerings were subsequently transcribed. I think most people would say these people were human subjects in an experiment.) What are the IRB implications?
- If we can squeeze anything promising out of these old data, I think a similar exercise should be undertaken, leveraging our local/domestic contacts with pianists (through Bukvich, Laird, Schrader, et al.) and perhaps Dr. Parncutt's network in Austria. If we only ask for annotations (as I did for DHCS) and self-reported hand measurements, what do we need to do with the IRB? We will probably want to collect some other information about playing experience. I think we need to request a waiver. Do you agree?
- I tried to ping Dr. Wilkinson about being on my prelim committee, but I notice he has left Skytree. Then I tried to Link him In. No response so far.
- I am unclear about the requirement to have two "tenured" faculty members on the committee. Does this mean two tenured UIC faculty? Tenured UIC CS faculty? Or tenured anywhere? Would emeritus faculty count as tenured? You are the only person on my short list of committee members who has tenure in the purest (UIC CS) sense of the word. Ziebart and Forbes are junior faculty. Moher and Jordan are emeritus. Wilkinson is adjunct (?), but he was formerly tenured in the Psychology Department at UIC. Raphael and Parncutt are external.
- And would Moher, who is retired, be considered to have a 50% appointment in the CS Department? Would Wilkinson? If the answer is no in both cases, Ziebart goes to the top of my list. The majority of the committee should have at least 50% CS appointments.
- Another strange thing is that the handbook talks about what the committee composition "should" look like. (Indeed, it says the five-person committee should have five or more members.) It is not stating absolutes, though that may be the intent. Ultimately, it seems the Graduate College is responsible for approving the committee. But I can't believe the Department (DGS?) does not sign off on this. The handbook doesn't say, and the Committee Recommendation Form link on the Department's Graduate Programs page is dead.
