I work with automatic environmental sound recognition. As a limitation of my work, I use only audio without other contextual information to do the recognition.

In nature, many sounds are very alike, for example a bouncing ball or knocking on the door [1]. This similarity between sounds, often confuses us. Thankfully, context can help us to disambiguate.

Many people have worked with automatic sound recognition. Once, I read that in a given experiment, when they asked people to identify sounds, the subjects stated that by listening to only 3 seconds of the sound, it was sometimes was not enough.

In another paper I read, they were stating that in some cases, humans need to listen for approximately 2 seconds to give meaning to the incoming sound. Sadly, I don't recall the reference.

Does anyone know if this is true, and can back it up with a reference?

[1] Disambiguating sound through context, MA. Niessen.

[2] Acoustic Environmental recognition. L. Ma

  • Tommyrot. Never start out with "I once read." There are 596.4 wrong sources for every correct source (numerical value made up). I, and most of my musically-inclined friends, can identify not only the instrument but also the specific recording in under 1/2 second. Identifying a musical pitch can be done with less than 0.1 second of input. Oct 17 '14 at 11:20
  • this has already been answered on our sister site, Sound. I'll migrate this there and close as dupe.
    – Rory Alsop
    Oct 17 '14 at 11:46
  • Rory Alsop. my question is different from the other suggested, but I will edit it to ask better.
    – jessica
    Oct 17 '14 at 18:13

It's not clear in which context you would have heard this, but taken at face value it is demonstrably false. Time yourself saying a single-syllable word, such as "cat". Did it take less than 2 seconds? (yes) Would another English-speaking human understand the meaning? (yes) Therefore this is obviously not true.

From a musical perspective, in just about any music you listen to, there's a good chance you will hear a large number of notes (probably the majority of them) that are less than 2 seconds. At a common tempo of 120 bpm, 2 seconds is the length it takes to play four quarter notes, and you can have lots of musical meaning in a shorter time than that.

  • The context is for environmental sound recognition. For example in (Acoustic environment claaification, Ling Ma). They report humans have difficulty to understand sounds with lenght of 3 seconds
    – jessica
    Oct 16 '14 at 18:26
  • There are some songs I can recognise; unprompted & with no warning, just caught the intro on the radio etc, by the 2nd beat.
    – Tetsujin
    Oct 16 '14 at 18:33
  • i can recognise a hi-hat in ½ second, same for a kick drum... less sure if I heard either a road-drill [jack hammer] or helicopter in the background of a movie I was watching. I think context plays a great part.
    – Tetsujin
    Oct 16 '14 at 18:43
  • You are right tetsujin. But without context?
    – jessica
    Oct 16 '14 at 18:48
  • Jessica - you keep saying this, but what do you mean?
    – Rory Alsop
    Oct 19 '14 at 20:39

There are a lot of sounds like lightning striking nearby that are way shorter than two seconds. Speech consists of sounds changing at a much more rapid rate than that.

And hopefully you need less than two seconds for both recognizing a car horn and initiating reaction to it.

I think you need to figure out some missing context for the statement you are thinking of.

  • Hope this answer does not get deleted.
    – Tim
    Oct 17 '14 at 10:39

I'm sure that I read in one of Denis Smalley papers that 6ms is most minimal time distinction the ear. After that point the audio information becomes a single stream of sound, like the difference a drummer playing a flam or a roll.

I can't remember exactly which paper it was in, but I remember it from when I was studying spectromorphology. If want to reference the papers then there are links at the bottom of his page: http://www.city.ac.uk/arts-social-sciences/academic-staff-profiles/professor-denis-smalley

However, what I think you're referring is the psychological effect of reduced listening. With nothing to associate with the certain listeners may claim they need to hear more to recognise the sound. My feeling is that wouldn't help them as it may have more to with the lack of a reference point.

It may be worth looking into the work of Trevor Wishart. He talks about speech recognition quite a lot, and some of his writings might be useful to you.

Hope that helps.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.