level 10 tales

Comments and questions about Absolute Pitch Painter
aruffo
Site Admin
Posts: 1694
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

level 10 tales

Postby aruffo » Sun Jun 15, 2014 12:48 am

My identification performance for the Large games of Level 10 is not so great, but I have some insight this time as to why.

Image

The red line shows a clear tendency toward height judgments, with my worst recognition in the yellow area. I wrote in the Level 9 thread that this was likely due to my mind's dividing the octave into seven ranges-- the bottom six categories individually, leaving only one range left for the entire upper part of the octave. I was actually starting to become very worried that I was locking myself into these divisions and perhaps had started to lose absolute knowledge of the various tones.

Fortunately, the problem is strategic rather than perceptual. The issue appears to be that, with 11 categories in play for the Large games, there are comparatively few border tones being presented. This meant that-- until today, when I realized what was happening-- I'd constantly get lulled into a false sense of confidence by easily identifying a string of non-border tones and then POW stumble on some tone that I didn't pay enough attention to. I wasn't able to keep myself above 4x without great difficulty, because I kept guessing, but I didn't realize that I was guessing because I was in a mode where I was actually able to correctly identify a whole bunch in a row, and therefore why should the next one be any more difficult?

Once I understood that I was being suckered into guessing, I decided on a different approach: that I would definitely not drop any egg at all until I was utterly certain I had identified the sound quality correctly. There's a certain calm confidence that I feel when I definitely know a tone, and that feeling is what I relied on in changing my strategy. This meant, for one, that I did not drop any eggs that seemed "almost certainly" to be such-and-such a category, even though it was tempting; and, for another, every time I felt that I was making a decision based on height-- even when it was obvious that the tone was clearly a semitone different from the one I'd just correctly dropped-- I placed the egg aside and waited until I could make a decision by its quality.

This has led to my creating groups of non-dropped eggs for each of my weaker categories (yellow through orange). This is effective for learning-- both because I have to decide whether differently-sounding eggs are in fact in the same category as each other, which forces me to evaluate their quality rather than their height, and because I am able to compare one group to another and confirm the qualities that define each group. After having played quite a few Large games where the categories of yellow through orange started sounding alarmingly indistinguishable from each other, it was quite a relief when comparing among these groups started to again bring out the distinct qualities of each... the wooden yellow, the whiny grellow, the smooth green, the twisted orange.

I'll probably make another main-page update when I finish Level 12 because, of course, at some point I will need to be able to evaluate whether or not this process is actually working to instill some kind of absolute pitch. As I have said (repeatedly), the fact of the matter is that every learning attempt since 1899 has succeeded in teaching listeners to know a few tones, only identifiable independently of any musical context, with the knowledge typically fading without constant practice. So far, I can reliably identify only a few tones, and not in music. (I'm constantly practicing so I can't say much about persistence of memory.) The level-up procedure provides interesting data about categorical perception, but what about the definitive note-naming skills? I'm aware of three tests, and I may make a stand-alone tool that will allow me to evaluate myself at any given time and thereby track my progress as a wannabe absolute listener.

aruffo
Site Admin
Posts: 1694
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

Postby aruffo » Mon Jun 16, 2014 10:49 pm

A strange thing that I do. When I am playing carefully, trying to be certain of each choice, and I nonetheless drop an egg on the wrong vat... I will in disbelief grab the same egg and drop it on the same vat again. Sometimes I do this two or three times (knocking the multiplier down to 1x) just because I was so sure!

aruffo
Site Admin
Posts: 1694
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

Postby aruffo » Wed Jun 18, 2014 4:40 pm

Done with the Large games..! The identification chart didn't change appreciably, but the discrimination charts look closer to the ideal than any other discrimination results so far.

Image
Image

I can't know whether this result will persist after the Jumbo games, but it's a nice thing to see here. I can see that the Yellow category is still weak, of course, and I wonder if my performance for Red is somehow related to its being an overall tonic.

aruffo
Site Admin
Posts: 1694
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

Postby aruffo » Tue Jun 24, 2014 10:52 pm

There's the end of the Level 10 games. The discrimination this time is less than ideal. I think the difference may be that after the Large games, I indicated as "different" only those tones for which I actually heard a difference, whereas after the Jumbo games I indicated "different" for any tones where I suspected I might be hearing a difference.

Image

Identification didn't turn out too badly, I suppose:

Image

What this shows is what I experienced: I know the lower categories by their individual quality-- notice the near-100% in each category with the steep drops at each border. The upper categories, not so much. I'm quite good at not mistaking red for rellow, which is kind of groovy... but I am not at all confident about any categories over red. When I encounter each one, I do recognize their quality, but I am not confident about it. So I place each egg over the color that I believe it belongs to, and when I finally hit one that I do actually know, then I get the others as a cascade. Even so, I still allowed myself to guess often enough that the categories' gradual drops at each side reveal my uncertainty.

Now I have to stop playing for a while and program the higher levels. I had thought that I would be able to complete Level 11 before going back to the programming interface-- but, unfortunately, I'm discovering that the assumption I initially made (that Level 11 would be the terminal level) may mean that Level 11 will behave unpredictably, and I certainly don't want to lose my progress thus far. So, programming interface, here I come..!

Axeman
Posts: 151
Joined: Wed Jul 20, 2005 4:20 pm
Location: New Zealand

real world effect?

Postby Axeman » Sun Jun 29, 2014 5:11 am

Are you noticing any real world differences in how you are hearing yet Chris?

aruffo
Site Admin
Posts: 1694
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

Postby aruffo » Sun Jun 29, 2014 11:34 am

No, not yet. I was thinking about that recently, and wondering if I should be discouraged about that. For the moment I've decided not to be discouraged, because

- I only just reached Level 11, so I only just reached 12 pitches
- I only just started playing APA regularly again, so I'm awfully rusty in pinpointing chroma
- Level 11 only has one octave and one timbre, so without pinpointing chroma I am unlikely to recognize tones that aren't that

And, perhaps most importantly, I think of my conclusions about literacy and phonetic awareness (e.g., why I included Chordhopper). As I progress in the game, and re-train myself using APA, I should become able to identify isolated individual tones in the real world. But until and unless I teach myself to read and recognize "words", any absolute skill I might gain will not be useful in music.

That said, as I mentioned, now that I'm on Level 11 and have all 12 pitches in play, I'm going to program a tool to assess my actual identification skills, using all three tests that I've found in the literature: straight identification, X-random-walk-X, and half-octave sorting. I don't expect to do very well right off the bat, but it would be encouraging to see myself improving at each of these tasks over time.

fyaseo
Posts: 1
Joined: Tue May 19, 2015 6:19 am

Great

Postby fyaseo » Tue May 19, 2015 6:26 am

Great infograph by @aruffo :) :D


Return to “Absolute Pitch Painter”

Who is online

Users browsing this forum: No registered users and 2 guests