Human attention span and length of music

[VK RAMAN]

Most of the western music is 3 to 4 minutes and our sampradaya bhajan songs also in majority of cases 3-5 minutes, why is it; and why majority of our carnatic music is more than 10 minutes and if RTP 25 to 40 minutes.

[Nick H]

Most of the western music is 3 to 4 minutes

There are symphonies, concertos, even the more serious rock music. It is completely erroneous to say that "most" western music is 3-4 minutes. Somehow, though, as far as pop is concerned, this is what they could get on a record. You could then ask, why did they make the record that long? I Guess that is all they could do with early 78RPM records and it stuck, even when "singles" became 33.3RPM. Musical historians please correct me if I'm wrong!

Pop music singles do not typify "Western Music" anymore than Bollywood typifies Indian music, but both are made for their markets, and I'd agree that the Western pop market is a short-attention-span crowd.

Serious music requires some sort of variety, and or some sort of development. I think that it is pretty tough for the pop brain to cope with more than two things, so we have a call-and-response form called verse and chorus. The words call and response originate in religion, I think for believers and non-believers alike, religions to have been a part of our culture for long enough to have have had far-reaching effects. So far as mass celebration is concerned, doesn't religious (speaking generally) music have more in common with pop? It wants to get those short-attention span brains engaged. In fact, the last time I went to a church wedding, I commented, rather scornfully, to my companion, "You people have turned this into a series of teenage pop songs!" Whether one believes in it or not, the churches, through the ages, have given us some incredible, amazing, serious music of great beauty: now it is teenage pop.

Something that the pop brain can cope with is a story, so, although presented in a melodically simplistic parcel, the ballad can have quite a lot of different lines. More than most carnatic songs. I suppose.

Variety and development both require time. The carnatic song may repeat the same line many times, but there is both development and variety in its presentation. A symphony may take a simple melody and repeat it many times in different ways --- and then move on to something else in the next movement and the next. The more serious rock musicians needed more than 3 minutes a side: they could, like the classical composers, fill entire LPs with long-attention-span music.

I think that the pop brain (as I'm calling it, for want of a better word) likes repetition, but it likes regularity and it does not like to be surprised. It is certainly a short-attention-span brain!

I guess all I'm saying here is that serious music requires a long attention span, whereas trivial music does not.

Now my question, to add to the pile. Why, to those of us who love long-attention-span music, is that short-attention-span stuff so irritating?

[vasanthakokilam]

Nick: I am with you on the major parts of what you wrote.

VKRaman, here is a twist that you probably did not anticipate. CM songs, as written, are in general shorter than western classical music pieces, as written.

CM has a wide variety in terms of length as we see them performed in a concert. If you leave out the duration added by the performance aspect of CM ( alapana, niraval and kalpanaswaram ), then songs are not really that long. They can be grouped into various buckets of 3-4 minutes, 4-6 mintues, 6-8 minutes. I do not think it will take more than 8 minutes to render even a long song even with sangathis and numerous charanas.

Here is a quick back of the envelope calculation.

Consider the many smaller compositions of Thyagaraja which are typically performed in the medium tempo. The usual tempo is that it takes around 5.8 to 6 seconds to finish one avarthana of adi thala. Those small songs are typically like this: Pallavi: 2 lines ( 2 cylces ). Anupallavi : 2 lines : Charanam: 4 lines.
Let us say there are two sangathis for each line and each is repeated twice. So there is a factor of 4 increase in time. In between sections, the pallavi refrain is repeated along with one extra cycle gap for flourishes and theermanams. So we have ( 2 + 2 + 4 ) * 4 + ( 2 + 2 + 2 ) = 32 + 6 = 38 thala avarthanams = 228 seconds = 3.8 minutes. Let us round it to 4 minutes. A longer song in 2 Kalai will be around 8 minutes. ( take all these calculations as approximate to get to the ball park )

My brain prefers songs of different durations depending on context. When I am driving, I prefer not to listen to long alapanas. I need a lot of variety. So when I make drive time listening palylists or CDs, they are typically a little snippet of alapana ( like a minute ), the song plus niraval plus kalpanaswaras. And a few other songs that are just songs lasting 4 minutes etc. So I can cover quite a wide range in one hour of driving. Same way for my Gym playlists. There it tends to be on the shorter side and more upbeat peppy CM or ironically a long and interesting talk on CM ( like a lecture demonstration ) or other topics ( including university courses ). If the context is CM, I am of course comfortable and indeed expect the usual fare. ( though I probably will get bored with alapanas lasting more than 10 minutes but I am always game for RTP ).

If I am in the film song listening mode, it does not bother me that they are all in the 3-4 minute long since that goes along with my expectations. The trick there is to choose the right songs since a vast majority of them are unlistenable. So a lot of upfront work is required. Also, given the compressed time, in the good film songs ( old and new ) the music tends to be very tight and condensed which also adds to the richness in a different way ( sound tapestry, timbre, color, or whatever.. The film music directors are geniuses in what all they bring to the table to pack them all interestingly within that short time perido ).

[Nick H]

I guess I gave the impression that I am never in the mood for short-attention-span music. That would be quite wrong: I am, regularly, although in some of the situations VK would be listen to "SAS" (see I don't have the attention span to spell it out yet again!) I'd be listening to "LAS." Can, for instance, listen to long Hindustani alaps to soothe the rhythms of city driving, although they might be a bit soporific for the highway at night.

By about 85%, though, i do not really care for background music. Be it pop or classical, I'd rather give it my attention or turn it off.

The film music directors are geniuses in what all they bring to the table to pack them all interestingly within that short time period).

It is a skill and art in itself. There are some trivial Western songs that manage to tell a story in just a few words with a little melody. The way that they work, for me, raises them up from being just trivial pop.

[VK RAMAN]

My statement of 3-4 minutes per song is based my 25+ years in recording and replication of Cds and that may not represent the exposure Nick has. I just like to know is there a study done to make songs like that. 9 digit telephone numbers were brought into existence in US based on how humans can memorize 9 numbers quickly.

[vasanthakokilam]

VK Raman, I now know exactly what you are talking about. The User Interface design is all about that, and especially in the modern touch based gadgets, there are many such considerations. Companies do think about such things deeply.

On the song length, I do not really know if any one has made that conscious decision on that but as with many things with art, it is a process of trial and error over time. An equilibrium of sorts like this gets established. I was also making the related point that CM songs can also be sung in that 3-4 minute duration. So there is something to that 3-4 minute duration. Mahavishnu may have something to say on this, I suppose.

[Nick H]

VK Raman, I listen to very few 3-minute songs, but, as implies, that is because my taste in western music is of the less "popular" sort. There are hundreds of years of w. music that does not come in three-minute bites. Although, of course, some of it does.

I really think that the answer to your question is a combination of attention span, and historical/technological accidents in recording. Companies often do not think about these things. 640K memory enough for you? 70 minutes good enough for a CD? Many of which do not contain 23 songs. And yes, I've been listening to music of all sorts since before the CD existed. But, probably you have too.

Telephone numbers have to be big enough to meet demand. Doesn't India have ten-digit mobile numbers? This might have been based on one of those "studies" --- but I can tell you that whenever they did such a "study" on UK telephone numbers, changing them accordingly, they had to do it again because they ran out. Londoners were particularly inconvenienced by this.

Studies show... Words to be always treated with caution. But another thing that has had a big influence on commercial music is the demands and requirements of radio stations, and it is quite possible that some of those would have done studies.
(a quick look at Google suggest that I'm right: blame the limitations of 10-inch 78rpm records, the requirements of radio stations, and human nature)

[uday_shankar]

9 digit telephone numbers were brought into existence in US based on how humans can memorize 9 numbers quickly.

While this is factually wrong (i.e., the US uses a 10 digit system) I think as Nick pointed out the rationale too is based on demand rather that the memorizing capability of humans.

Music and "songs" are two entirely different things.

A symphony can last 40 minutes but it is not a song. Similarly, Dhrupad singers can sing an alaap followed by a bandish for hours together, somtimes an entire night (so they must require huge attention spans by your theory!). Out of the three classical systems of music I have listened to, only Carnatic music has a structured entity like a kriti (i,e., "song") the foundation for the underlying music. So it is fallacious to compare this with ligther song-based systems like Bhajans, film music or other popular forms. In fact, I would argue that among classical music listerners, the shortest attention span is among Carnatic rasikas. Hence the immense popularity of people like Ariyakudi who belted out song after song without pause.

[mahavishnu]

While this is factually wrong (i.e., the US uses a 10 digit system) I think as Nick pointed out the rationale too is based on demand rather that the memorizing capability of humans.

I agree that this is in error. But there there is more to the story though...

In 1956 George Miller, a noted cognitive psychologist at Princeton discovered a phenomenon called "chunking" during memorization of common data structures like digits. He found that humans had a processing capacity for keeping seven items/objects in memory, referred to as the 7 ± 2 law or Miller's Law. Researchers at Bell Labs in the late 1950s (and later) found that even for these seven digits in memory, it was easier to chunk in 3s and 4s like xxx-xxxx, which explains why we had phone numbers like 867-5309 (if you get this pop reference, excellent. now pl stop chuckling and read on).

And if you add a three digit area code to this, it becomes of the form 202-555-1212 and we can chunk this information into meaningful smaller 3/4 units. These items decay rather fast after initial recall and unless consolidated, their trace will disappear from memory or even get interfered with by new information. The time scale of recall of this phenomenon is restricted to a few seconds. Bell was especially interested in the human memory data due to the attribution of misdialed numbers and greater operator costs to the inefficiencies of encoding in human recall memory. However, this has little to do with attention and does not extend to cognitive processes happening in the time scales of 3-4 minutes.

Even if we restrict ourselves to western popular music, I do not think there is a cognitive theory of why an average pop song is 3-4 min long. The closest is something that a colleague of mine named Petr Janata at UC Davis has proposed. He believes that when we listen to music, the first few things that happen in the brain are related to the processing of pitch, key information and melody sequences. In CM, that would translate to identifying the pitch/tonal centres of the shadjam, ragam outline and possible violation of expectations if any. This is followed by higher level association processes, emotions associated with the mode/scale and then autobiographical memory processes (where have I heard this before, is it different from the previous experience etc). For there to be sufficient development of activity at both these timescales happening interactively, Janata proposes that a piece of music can be as short as 18-30 seconds or so. This does not explain why the average Beatles song is about 3 minutes long, but it tells us why the average sample in the iTunes store is about 20-30s long.

Talking about the the length of songs and the arbitrariness of all of this reminds me of a brilliant composition (tongue-in-cheek) by the great 20th century classical composer John Cage named 4:33, which is essentially 4 minutes and 33 seconds of silence. You can buy that in the iTunes store!

https://itunes.apple.com/us/album/433-s ... d412575119

[varsha]

Human attention span and length of music

Whose span ? Performers or the listeners .
One look at 50 % of audience in an average chennai concert is good enough for the artist to jettison the performing career and search for something more worthwhile .
Crossword puzzles , sudoku , popcorn munching , betel leaf chewing , craning around their own necks to catch sight of someone , or to hail another . Walking out to attend a mobile call . And a hundred another things .
There is an increasing mismatch between performer and the receiver . Sign of our times .
Good that CM is mostly a personal plea , so that the artist can sing for himself and get out .

I have sat through night long abhang sessions when not many songs were performed . And here is an artist who - in my books - is the ultimate in showcasing a raga within 90 seconds .
I have strung together 4 examples here = puriyakalyan , megh , brindavan sarang and hamsa kalyan.
http://www.mediafire.com/download.php?15g1j046m4av75i
Basavraj Rajaguru
At the other extreme , there are examples of a 56 minute yaman by amir khan , 72 minute hamsadwani by Kishori Amonkar without a dull moment forcing the listener to drop attention .
That is why 78 rpms will always remain a treat .
It will always depend on the capabilities of the performer to swim at both ends of the pool .
And on a listener who can walk the meditative track , filtering out the noisy thoughts that keep crowding in .

[Rsachi]

My two pennies:
Long back, phone calls were billed in slabs of 3 min. I also remember college debates had a 3/4 min limit per speaker. Since most radio songs were around 3 min., they said that one should brush teeth for the length of a song.

But I also think the time scale is entirely a function of the situation, mood, expectation, genre. Here is a 20 min plus great Suruti I am able to enjoy, lying in bed with 101F temp:
Shashank plays Suruti

[arunk]

I dont know if western popular music "naturally" (or driven by some inherent attention span limit of its fans) gravitated towards 3-4 minutes, here is my 2 cents (and a theory). Perhaps it has already been mentioned.

A lot of it has to do with radio-play and commercials. I remember an interview with Queen (British band) about their hit Bohemian Rhapsody and how the producers and many others thought it would never "make it" even if it was good owing to the length. The resistance was if it is too long, # of commercials come down => less revenue for radio stations => less royalty. So good music, but not "beneficial" economically to many of the stakeholders. Now this may have resulted in songs being limited to a certain duration which itself perhaps led to the fans of such music adjusting their expectations which in turn could have set the expectation over time.

Arun

[alpajnani]

While I don't agree western music or Indian music is truly 3~5 minute songs (as previously well illustrated by other posters ...... symphonies to RTP's span much longer), shouldn't we rather inquire what is the typical length of a poem/kriti/kirtana which (at least in the context of Carnatic Music) roughly rules how long one chooses to sing it. On the other hand, should the inquiry be why Ariyakudi and predecessors chose to sing shorter kriti's around the one or two main longer pieces (given I have been to Hindustani concerts where the whole concert of 4+ hours was just two raagas/"compositions").

That said, I do agree that the LP/cassette format didn't allow for longer recordings (and perhaps related to commercial/packaging limitations of multi-cassette/multi-LP sets). Don't we see better availability of longer RTP's and entire kutcheri's being available now with the advent of CD/DVD/mp3's as opposed to cassettes/LP records?.....and related to which I hope we see an end to the "track recording" of "songs" for Carnatic Music and stick to kutcheri formats instead

[VK RAMAN]

Uday Shanker - I stand corrected. I am in US for last 32 years and even then we end up making misstatement. Thank you for the correction.

[mohan]

Yes, the early standard 78-rpm phonograph records were 10 inches in diameter and recorded only 4 minutes of sound. Hence (for western popular music), the recording technology had an influence over duration of the composition.