Ableton Forum

My thinking aloud:

I'd have thought if you can hear 20kHz, and tell it's pitch from others, then your ear will operate and discern 1/20kHz = 50us ear drum cycle. You brain obviously operates a lot faster though.

I'd have though the overall latency to respond to this mass of information, will depend on your concentration and I assume this is similar to reaction time (e.g. when hitting the brake on your car). For car "stopping distances" and "thinking distances" it's supposed to be maybe 3/4(?) of a second, but percieving music is a completely different ball game. So basically this depends on a "music chairs" vs "audio sounds odd" scenario.

Thats more what I was getting at, but in production terms I guess there are rules of thumb and effects that can be predicted in general which is good enough.

Angstrom wrote:
It's still safe to say that we perceive the world as it was, not as it is.

or maybe we perceive the world as we predict it to be based on previous sensory input.

pepezabala wrote:well, you can try it out with your DAW. I suppose that it should be somewhere at 10 milliseconds. That's where you notice latency more or less. Everything below 10 milliseconds gets perceived as simultaneous event.

However there should be a wide range of variables coming in. Pitch, loudness, environment etc.

but i understand now that there are people that state that live syncs perfectly while others see it as inferior

Let me grab a book.

Let's say we are crossing a busy street in the heart of Seoul miinding our own business (too much) without noticing that the pedestrian light has just turned red. This guy sitting in a huge truck gets all mad and hits on his 120dB car honk that just blows you away (without damaging your ears luckily!). Even though the honk lasted one second, you are not able to hear anything that occured for another 200ms or so. This is called post-masking. Pre-masking is a bit more interesting. [...] It so happens that some sounds that occur before the masker, if they happen close enough to the start time of the maskers acoustic event (honk) will also not be heard.

According to the graphic attached to this chapter, the pre-masking range is around 50ms. One of the algorithms that makes use of this is MP3.

pp. 12 - 13 | Introduction To Digital Signal Processing: Computer Musically Speaking (9789812790279): Tae Hong Park

mattadms wrote:Let me grab a book.

Let's say we are crossing a busy street in the heart of Seoul miinding our own business (too much) without noticing that the pedestrian light has just turned red. This guy sitting in a huge truck gets all mad and hits on his 120dB car honk that just blows you away (without damaging your ears luckily!). Even though the honk lasted one second, you are not able to hear anything that occured for another 200ms or so. This is called post-masking. Pre-masking is a bit more interesting. [...] It so happens that some sounds that occur before the masker, if they happen close enough to the start time of the maskers acoustic event (honk) will also not be heard.

According to the graphic attached to this chapter, the pre-masking range is around 50ms. One of the algorithms that makes use of this is MP3.

pp. 12 - 13 | Introduction To Digital Signal Processing: Computer Musically Speaking (9789812790279): Tae Hong Park

Hey, this is what I wanted to say!
Also, to increase the set of parameters here, the masking of course depends on the Volume difference of the 2 signals. So, if the second hit is louder than the first, the difference we can distinguish gets smaller!

Piplodocus wrote:I'd have thought if you can hear 20kHz, and tell it's pitch from others, then your ear will operate and discern 1/20kHz = 50us ear drum cycle. You brain obviously operates a lot faster though.

You have to be carefull here, as the different frequencies in the brain are not represented as a frequency of action potential, but of what cells fire them. 20kHz is way to fast for them often lasting some ms or less, but not as afst as 50µs. teach me wrong?!). But still you are right. Your brain operates very fast, but you have to think more of a billion slow CPUs than of one very fast. And this is also where our limits come from (reaction time etc..).


BTW.: I am glad that you all care about your most important device in music making

pre masking is a new concept for me but it reminds me of something else that I've noticed.

When you filter a continuous sound like a fan droning or a motor running you always suddenly become aware of it when it stops. It's as though in order to notice it stop you have to become aware of it just before it stops. More evidence that we only consciously experience events after they have been filtered, processed and interpreted by unconscious processes.

there's a difference between the highest frequency and the time it takes to notice that frequency though, eh?

Like, I can hear about 17khz, but I don't think my brain 'notices' that such a signal has started within 0.05 of a millisecond!

It needs time to say - "uh, I think , uh, a noise may have erm, started, or something, so lets have a listen to it." That's how my brain works.

There is some discussion of this in Curtis Roads' book 'Microsound'. Luckily it is in the sample chapter available online:
http://mitpress.mit.edu/books/chapters/ ... 7chap1.pdf

The discussion about time scales is about half way through the chapter.

Angstrom wrote:
It's still safe to say that we perceive the world as it was, not as it is.

Which technically, is even true with what we see, since the light has to travel from an object to our eyes, and then get processed by our brains. So in a way, we're always looking into the (very, very recent) past. Something to ponder...perhaps especially if you're high. /mindfuck

No man, because then, like, I am totally seeing the future?

Here's another answer: 1ms.

I was just reading this - Best Practices for Open Sound Control - and in the paper they say...

This limit of 1msec coincides with the threshold for just noticable difference in onset time between two auditory events.

...though this figure isn't referenced in the paper...

Not read all this, so sorry if some of this is allready posted.

Nervous impulse speed - about 100m/s or so (varies acording to specific nerve path) for anything moderately important - so say about 1/3 speed of sound over a distance of about 30mm (so about 0.3ms).

Actual processing of sound varies according to frequency - we deal with upper mids far faster than other frequencies - lower frequecies - could be 20ms or so - for high notes (rustling noises for example with alot of conent from 4K upwards) its alot faster. Changes in sound are more important than steady sounds. You also have to factor in that low frequencies naturally take longer to activate a fibre in the ear and it also depends upon whether the there was silence before or not in that band. Fibres need time to start mving at the same frequency and in phase with the presure changes they are sensitive to - which comes to bands:

Sound is handled in 24 or so bands, at an average about 1/3 octave for each band - if someone wanted to model human audio processing, then a 1/3 octave band filter to separate sounds into 24 bands for parallel processing and analysis is probably a good place to start. Each band kind of behaves like a single bin of an FFT - intensity and phase or time difference depending on the frequency, but theres alot more to it else we wouldnt be able to determine small (4th or less) intervals in music. Actually I think we do it via harmonics and phasing rather than separate processing of close notes - but thats just my guess - probably bollox

hi all,

a bit short on time today, just skimming through this interesting post.... i have a couple of other things which may be of interest, some of which has been partially covered. But if you want to explore this further i recommending the couple of books below, and no i dont work for amazon!

One aspect of this which i dont think has been mentioned so far in this topic is the processing of sound by the brain and subsequent "latency" involved in human reaction times ( this is more relevant to the musician than the listener)... sound travels through air, these waves hit our ears which transmit nerve / electromechanical pulses to the brain, the brain processes and then transmits to the arms or the feet. each one of these stages introduces a delay, with the biggest being the brain to the arms / feet of the musical performer (approx 1/5 of a second!!! sorry but i find that amazing...) -- excellent chapter on the subject by Max Matthews in this book: http://www.amazon.co.uk/Music-Cognition ... 294&sr=1-1

PHON curves (fletcher munson curves) - the loudness of a sound in a room (measured in dBSPL) has an effect on how clearly we hear certain frequencies. this topic is relevant to those who are interested in mixing techniques and theres a great chapter (chapter 2) on the subject in this book: http://www.amazon.co.uk/Mixing-Audio-Co ... 0240520688

happy listening / music making!