Cleared them up for me, isn't language great?
Anyway, the rest of the talk is hunky dory
Listen ScottJ, I appreciated your questions - you seemed genuinely interested in trying to understand this & the answers but if you intend to engage in your usual sniper tactics whenever someone comes along that seems to contradict what I say then please move on also.
ScottJ is plain wrong and anybody who listens to the video will see that for themselves - enough said.
Maybe I've made a mistake. Just to be sure, you don't sell any audio electronics and related accessories?
Are you saying I interpreted ScottJ's post incorrectly or Josh McDermott's video, incorrectly - it would help to be more specific?
BTW, I did neither but I'm interested in what you meant?
What he said "take any 10 second sound clip at 44.1KHz and you can decompose it into a sum of two streams of 441,000 numbers (essentially infinite choices)"
And he went on to say "you have no a priori knowledge of anything but the result"
And that only serves to confuse & is incorrect in this scenario where we are analyzing the signal mix to isolate one particular real world sound stream in that mix. There is NOT an infinite number of possible real world sound streams that will result in this same mix because sound generation is constrained by physics. McDermott himself says in the video "real world sounds are NOT random" & "real world sounds are a very, very, very, very small portion of the set of all possible sounds"
I don't know if ScottW listened to the video or understood it but what Scott posted directly contradicts what is said in the video & also misunderstands what auditory scene analysis is about?
Furthermore we do have a priori knowledge of this constrained behavior
It's not abstract mathematical modelling that we are talking about
Last edited:
"The problem that ASA has to solve can thus be rephrased as follows: given the flow of sensory information from the cochlea, which resembles a time–frequency analysis, the listener must determine the likely combination of physical sources in the world. Unfortunately, this is still an ill-posed problem. An exact solution being impossible, perceptual gambling must begin."
From this paper "Auditory Scene Analysis: The Sweet Music of Ambiguity"
From this paper "Auditory Scene Analysis: The Sweet Music of Ambiguity"
We can filter sound effectively because we learn from our previous experiences sounds which have regularity and are meaningful (and the rest which are useless, uncommon, not useful, unimportant, irregular). For example, we can filter the sound of background music (we might possibly know the title of the song), we can filter the sound of wind (in windy environment), etc. etc.. But we can do all this simply because we have intelligence.
Most autistics are low intelligence (so your experience might be with such individual?), so it is normal if they don't have the ability to narrow their perception. They have less experiences that they can understand in order to make effective filters.
Here's a great video illustrating what an amazing job your brain does processing the the sounds that reflect off the shape of your ears into your ear canal to give you 3D sound perception.
YouTube
I was quite blown away by this, watch it through to the final experiment.
As a Dyslexic I find Autism is interesting as it's on one end of the the same spectrum of problems.
I too find a lot of background noise a pain. If I want to concentrate on something I'm better of putting some headphones on to cut out the open plan office noise so that the noise I'm hearing is more predictable. The reduced ability to filter out external distractions means I was always the first person to spot a wasp enter a classroom full of kids. My Auditory memory for what's just been said is below par I struggle to remember phone numbers but my visual memory for pictures and Images was off the end of the chart used to measure it. Intelligence in those with Autism and Dyslexia is hard to compare to those without because the cognitive skills are different. Does my ability to assemble flat pack furniture with barely a glance at the instructions make me smarter than someone with a PHD that struggles with it - probably not, just wired a little differently.
YouTube
I was quite blown away by this, watch it through to the final experiment.
As a Dyslexic I find Autism is interesting as it's on one end of the the same spectrum of problems.
I too find a lot of background noise a pain. If I want to concentrate on something I'm better of putting some headphones on to cut out the open plan office noise so that the noise I'm hearing is more predictable. The reduced ability to filter out external distractions means I was always the first person to spot a wasp enter a classroom full of kids. My Auditory memory for what's just been said is below par I struggle to remember phone numbers but my visual memory for pictures and Images was off the end of the chart used to measure it. Intelligence in those with Autism and Dyslexia is hard to compare to those without because the cognitive skills are different. Does my ability to assemble flat pack furniture with barely a glance at the instructions make me smarter than someone with a PHD that struggles with it - probably not, just wired a little differently.
OK, thanks
As we age a lot of people find that they can't as easily follow the one conversation in the noisy room of other conversations as they were able to do when they were younger (I notice this myself). There may be a number of reasons for this one of which is some hearing loss at certain frequencies but another is some loss in the brains ability to do the analysis necessary to achieve the above "Cocktail party effect"
Yes, good video - hadn't seen it beforeHere's a great video illustrating what an amazing job your brain does processing the the sounds that reflect off the shape of your ears into your ear canal to give you 3D sound perception.
YouTube
I was quite blown away by this, watch it through to the final experiment.
As a Dyslexic I find Autism is interesting as it's on one end of the the same spectrum of problems.
I too find a lot of background noise a pain. If I want to concentrate on something I'm better of putting some headphones on to cut out the open plan office noise so that the noise I'm hearing is more predictable. The reduced ability to filter out external distractions means I was always the first person to spot a wasp enter a classroom full of kids. My Auditory memory for what's just been said is below par I struggle to remember phone numbers but my visual memory for pictures and Images was off the end of the chart used to measure it. Intelligence in those with Autism and Dyslexia is hard to compare to those without because the cognitive skills are different. Does my ability to assemble flat pack furniture with barely a glance at the instructions make me smarter than someone with a PHD that struggles with it - probably not, just wired a little differently.
Here's an article along the same lines - what all these demonstrate is that we have learned, over time, the particular spectral changes our own pinnae make to the sound hitting our eardrums & we have built an internal model of sounds which accommodates to this personal spectral profile. Changing the pinna shape & we change the spectral profile thereby losing some aspects of our hearing ability.
How the Shape of Your Ears Affects What You Hear - The New York Times
BUT live with that changed pinna shape for a while & we will accommodate to the new spectral profile & these abilities return.
They did the same with vision having volunteers wear prism glasses for a week. The glasses initially flip the what we see upside down but within a week we have accommodated to the new orientation & everything looks normal.
Here's the thing - take the glasses off after a week & we almost instantly revert to our normal vision put them back on & we very quickly find this is normal too (even after a couple of weeks elapsed time) - so our brain has learned & created two internal models which deal with the way the image falls on the retina (in actual fact the eye's normal image is inverted on the retina & the prism corrects this inversion to an upright orientation of the image on the retina)
Last edited:
BTW, just as a bit of interesting fun - here's a sound tunnel which only allows sound travel in one direction - I'm sure it could be DIYed?
One-way sound tunnel offers novel way to control acoustic waves
One-way sound tunnel offers novel way to control acoustic waves
Yes, I call that a summary of part of what goes on in digital audio. The deal is that clock deviation presents as Frequency Modulation and ultimately at the loudspeaker as Doppler distortion (additional Doppler distortion !). The deviations at 1/f band if great enough amplitude cause perceptible change in timing in lower frequency notes. Higher repetition rate deviations are less so directly audible but subjectively still present as maskings and distortions. Getting low frequency deviations really low amplitude fixes a whole bunch of problems and this is described subjectively as solid stable etc sounding bass and low bass....this is what Jocko is on a about, hi Jocko. Clock jitter is 'voicing' in digital systems.......low jitter is good, low jitter with low 1/f component is generally better. Clocks satisfying the above can still benefit by tweaking, I am having interesting fun with this with my 'goop', crystal oscillator packages are readily prone to local field effects lol.
Dan.
Thanks for chipping in, Max
My take on low 1/f noise is a bit different to what you laid out above.
The result of close-in phase noise of oscillators IMO is that it results in a slight blurring of the reproduction of each frequency - instead of a pure 1KHz tone being reproduced, it is deviating in frequency randomly either side of this 1KHz - so we no longer have a purity of tone reproduced.
Why we notice this more prominently in the bass & in soundstage has to do with psychoacoustics. Our auditory processing localizes sounds using mainly ITD (Interaural Time Difference) below a couple of hundred Hz & mainly ILD above. So when low frequencies are reproduced in a spectrally purer form, we perceive the start of these LF sounds more accurately in time - hence these sound objects are perceived as more solid & better placed in time & as a result the sound stage becomes more real
Yes, it was only when I could get some NDK oscillators that Jocko had measured & sorted from a large batch, I was sure they had low close-in phase noise that I could be sure what the audible effects of this low 1/f was. Up to this without spending a large amount of money you cannot guarantee you are getting low 1/f oscs.
IME, the biggest uplift in sound quality related to clocks is attendance to the PS supplying them
My apologies MM, I didn't intend being contentious. I was simply assuming the choice of words came from the standard math lexicon.
We're on the same page...jitter is vibrato....anybody with a synth will tell you that modulating/filtering/dynamics of the vibrato control voltage will cause all manner of audible effects/voicings etc.
Of course digital system jitter/vibrato magnitude is much lower level than the synth example but it has the same effects none the less.
Jitter/vibrato also causes tremolo, just to add a little confusion into the mix.
.What is the Difference Between Vibrato and Tremolo?
Sure that's the first thing to attend to, but after that the remaining clock uncertainties can be set to an order or orders that are subjectively satisfying or subjectively innocuous or subjectively absent.
Similar in a sense to your experience of Jocko's selected crystals, I can at will dial up any of the above three subjective findings.....dead clean can end up sounding boring long term, a bit of the right 'reclocking' (controlled/known added clock noise) can do wonders for long term enjoyment and satisfaction (and more).
How and why this can be digs deeper into physics, physiology and neural processing/perception.
Dan.
- Status
- Not open for further replies.