Piwin -> RE: Picado Speed (Oct. 14 2022 7:51:45)
|
quote:
I would have thought there is rather a rate limit on processing chunks of spoken information/meaning - don't know what these are called, but they are not phonemes? Right. I guess it's a bit of a messy topic, but basically phonemes are considered the basic unit of meaning because of their contrastive value. For instance, how do you know whether I said bat or cat, desk or disk, fan or van, etc.? In each case there's only one sound that changes, and since it has the power to contrast different words that way, it's considered a phoneme. So it's in that sense that phonemes are considered to carry information. Note that you can have sounds that are different but don't have any contrastive value. Those are called allophones. For instance, in English [p] (as in the p in "lip") and [ph] (as in the p in "pit") are allophones of the phoneme /p/. They're different sounds, but they have no contrastive value. In the minds of English speakers they're both conceptualized as just a "p" and many may not even realize they're pronouncing them differently. Note that phonemes are language specific. While [p] and [ph] are not phonemes in English, they are in Icelandic, for example. If you're doing purely acoustic phonetics, without worrying about meaning, the basic unit is a "phone". You can think of phones as the assortment of all the different sounds that make up human language. Nothing language specific here, [p] and [ph] are phones regardless of which language they show up in and regardless of whether they have any contrastive value. It's really just two different approaches to the same subject matter, one focused solely on acoustics, and the other focused on meaning and functionality. Re: auditory processing dropping off around 10 sounds per second, I'll have to go check. I probably got ahead of myself by saying it sounds like a "continuous buzz". I remember Adam Neely's videos on rhythm vs pitch. The literature he was citing said that the perception of rhythm shifts into perception of pitch at around 50ms intervals, so 20 notes per second. And even there, personally even if I can't really perceive it as rhythm anymore, I do perceive that it's not just one uninterrupted sound. I'll have to go check, but I think the 10 value was more on whether you can "make sense" of what you're hearing enough to remember it. So, like, if I played 10 different sounds in a second, would you be able to tell afterwards what those sounds were, in what order they occurred, etc. I think the example Ricardo gives of three professional guitarists not being quite sure how many notes there were is an example of that. They're still hearing different sounds, but there's a drop in resolution. What the chunking model of language posits is that you have different levels of representation at varying levels of abstraction, operating in parallel. At the sensory level you have this flow of input that keeps on coming. That input is transient and your memory only allows you to handle so much of it at any given moment. So before you "lose" input to the new input that is coming in on that basic sensory level, you want to chunk it and pass it up to a higher level of representation. They posit that only information in the same level of representation can interfere with each other. So the input you just passed up to a higher level of representation isn't erased by the incoming input at the lower level of representation. And the further you get from the immediate sensory input, the more time you have to process chunks. An example from Mortensen: [[[Super][cali]] [[fragi][listic]] [[expi][ali]] [docious]] So, once you've chunked "super" and "cali" into "supercali", theoretically there's no risk that the incoming "fragi" could interfere with it. Only "fragilistic" and "expiali" could. You keep keep chunking up to higher levels, perhaps not simultaneously, but with a sort of cascading effect, e.g. while you're chunking "expi" and "ali" in one level, you might be chunking "supercali" and "fragilistic" in another level. And you keep on doing that game of chunking and passing it upwards until you have the full word. Then you can apply the same process to sentences and larger units of meaning. Anticipation would also play a role to facilitate the process. In this case, as soon as you hear "supercali", you probably already know what the rest of the word is going to be. Anyway, it's just a way of trying to explain how we process fast, transient linguistic input given the limitations of human memory. Whether we can extrapolate some of it to music, I don't know. Intuitively I feel like there are many similarities there, especially with regards to the role of anticipation. I'm not sure we could process sounds very well at those speeds if they didn't follow the patterns we expect them to follow due to our past experience of listening to music. But who knows.
|
|
|
|