ASAR Pattern Recognition -small steps for man –one giant leap for maturizing in audio - diyAudio
Go Back   Home > Forums > General Interest > Everything Else

Everything Else Anything related to audio / video / electronics etc) BUT remember- we have many new forums where your thread may now fit! .... Parts, Equipment & Tools, Construction Tips, Software Tools......

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 12th April 2011, 05:48 PM   #1
mige0 is offline mige0  Austria
diyAudio Member
 
mige0's Avatar
 
Join Date: Jan 2005
Location: Austria, at a beautiful place right in the heart of the Alps.
Default ASAR Pattern Recognition -small steps for man –one giant leap for maturizing in audio

LOL, in humoresk memory to great fakes and great jokes in recent history :
Neil Armstrong performing Gary Peach:
„One small step for man, one giant leap for mankind“
One small mistake... The proof that Neil Armstrong got his famous 'One small step for man' Moon-landing line WRONG | Mail Online
British scientist claims he coined Neil Armstrong's moon landing quote - Times Online


************ **************** ************* ************* ***************




This isn't exactly about what audio guys are concerned with in usual speaker design process – but may be worth to put together nevertheless, as there seems to be a loooot of confusion among audio guys on this topic.
Beyond that, it may shade a light to strong (but wrong) believes among audio folks that – IMO - currently (and for quite some time now) act as massive mind blockers.

Though it may not sound like from the beginning - it will culminate in the question if we „hear in math terms“ or if we have the ability to recognize and distinguish specific sonic patterns in the time domain.


What will be addressed along the course:
the myth around „ill diffraction“ - most prominently popularized by Earl Geddes
the myth around „ill stored energy“ - most prominently popularized by Siegfried Linkwitz

- above two points of view finaly being highly interwoven with the (pretty limited) current paradigm of „frequency domain thinking“ - most prominently popularized by Floyd Toole.


To make my point I'd like to introduce a „new“ analysis method that I'd like to call PBW analysis = Packed Burst Wavelet analysis.

Being looking for quite some time for a visualization tool that most pin point displays time domain behaviour to easily distinguish the specific properties of CMP systems, I finally arrived at PBW analysis method that provides exceptional high resolution in time *and* frequency domain.
Its a form of visual representation to what most basically could be addressed as „Acoustic Source & Acoustic Room“ (ASAR) pattern.
As we will see, this pattern differs significantly from „simple filter“ patterns.

Anybody knows, I'm neither a math magician nor any well skilled in Matlab / Octave coding, so I'd like give credit to Elias Pekonen who has pioneered wavelet analysis for the DIY community and also to Jean-Michel LeCleach, both of which have been providing a looot of good work regarding spectral analysis, so that I finally was able to adapt wavelet analysis to my needs.



Michael
  Reply With Quote
Old 12th April 2011, 05:49 PM   #2
mige0 is offline mige0  Austria
diyAudio Member
 
mige0's Avatar
 
Join Date: Jan 2005
Location: Austria, at a beautiful place right in the heart of the Alps.
PBW (Packed Burst Wavelet) analysis is just another tool in the box capable to provide – what some might call – another *bunch of pretty pictures*.

The kind of visualization I was aiming for, is to most clearly show the specific pattern of CMP behaviour / ASAR patterns.

In the „Horn Honk Wanted“ thread there already were analysis methods introduced to bring out patterns related to delay effects (= echo and looped echo) .
3-D PBW (Packed Burst Wavelet) analysis is just another – slightly refined - method to display time domain impacts.

In analogy to what I've shown with „Thermal Distortion“ we possibly should look at CMP and ASAR in categories.
Technically seen, its all based on one and the same mechanism but of course is *perceived* differently :

1.) very short time delay = chassis related (cone / dome / foil brake up)
2.) short time delay = related to speaker construction or due to multi-source overlap (second sources created by diffraction, honk of horns, TL behaviour, OB, back loaded horns, etc. or difference in time of flight from multiway speakers for example)
3.) long time delays = rooms (boundary reflections in rooms and room modes for example).

So ASAR patterns basically embraced each and every effect *not* related to „simple filters“ but to *time delay effects* .

More philosophically speaking – those time delay effects actually are an interwoven property to acoustic rooms (hence ASAR pattern) making room perception indistinguishable from CPM perception.
Of course all three above mentioned „delay time categories“ have equivalents in any original sound event (not any related to sound reproduction !), we are trained and already „familiar“ with recognizing such patterns for as long as mankind exists.

As ASAR patterns most basically could be characterized as „repetitive“, there is no loss of information happening (in theory) with adding a few more such patterns during reproduction.
Even so, we do not aim after that - from a „fidelity“ perspective – most people do not have any problems to „see through“ ASAR patterns that do not stem from the original sound event.
Thats even true for actually pretty annoying ASAR patterns like horn honk for example.

It means, that our ear brain system can detect, value and suppress or highlight (to some degree) specific ASAR patterns according to „getting used to a listening environment“.

Otherwise it would not have been possible to enjoy music in such a wide variety of listening rooms and with that many different speaker (and headphone) constructions as we do for centuries now.
Actually, technically seen, this isn't any different from sitting first row or balcony when enjoying the same concert – the impression / presentation may be different and picking up details will take more or less effort – but its actually „all there“.

At a „second thought“ it becomes obvious that the stability of ASAR patterns in time domain is more so a quality criterion.
This immediately explains the impact of jitter – both „true“ electronic generated one and from second order thermal and acoustic effects that act „as if“ – and whose absence / diminishing always is recognized as a clear gain in reproduction performance.



Michael
  Reply With Quote
Old 12th April 2011, 05:50 PM   #3
mige0 is offline mige0  Austria
diyAudio Member
 
mige0's Avatar
 
Join Date: Jan 2005
Location: Austria, at a beautiful place right in the heart of the Alps.
Lets start out with a variant of what also has been shown back in the „Stored Energy Bogus“ thread :
Debunking SL's „Stored Energy“


A simple sine burst – overlayed by a specific impulse response


First off - here the impulse response of a „simple filter“ 1kHz / Q10 / +6dB EQing :

Click the image to open in full size.


Below is whats seen if we look at a bunch of sine bursts (look out for the second from left):

Click the image to open in full size.


below is whats seen if we look at a single sine burst (rectified):

Click the image to open in full size.


and finally here is whats seen if we take a bunch of wavlet analysed sine bursts – between 500Hz and 5kHZ - and display it in a 3D plot :

Click the image to open in full size.


and the same plot animated :

http://www.kinotechnik.edis.at/pages..._6dB_boost.gif (6MB)




Michael

Last edited by Pano; 13th April 2011 at 02:46 AM. Reason: image overload
  Reply With Quote
Old 12th April 2011, 05:51 PM   #4
mige0 is offline mige0  Austria
diyAudio Member
 
mige0's Avatar
 
Join Date: Jan 2005
Location: Austria, at a beautiful place right in the heart of the Alps.
To further warm up - same below for a „simple filter“ 1kHz / Q10 / -6dB EQing :

the impulse response:

Click the image to open in full size.


a single sine burst (rectified):

Click the image to open in full size.


and 3D plot plus animation :

Click the image to open in full size.


http://www.kinotechnik.edis.at/pages...10_6dB_cut.gif (7MB)


OK – so far we *should* be familiar with – no ?


Michael
  Reply With Quote
Old 12th April 2011, 05:52 PM   #5
mige0 is offline mige0  Austria
diyAudio Member
 
mige0's Avatar
 
Join Date: Jan 2005
Location: Austria, at a beautiful place right in the heart of the Alps.
Now proceeding towards more interesting ASAR patterns / CMP behaviour :


the impulse response of a single 100% echo happening after 3ms :

Click the image to open in full size.


a single sine burst (rectified):

Click the image to open in full size.


and 3D plot plus animation :

Click the image to open in full size.


http://www.kinotechnik.edis.at/pages...r_echo_3ms.gif (7MB)





Michael
  Reply With Quote
Old 12th April 2011, 05:53 PM   #6
mige0 is offline mige0  Austria
diyAudio Member
 
mige0's Avatar
 
Join Date: Jan 2005
Location: Austria, at a beautiful place right in the heart of the Alps.
And another ASAR patterns / CMP behaviour created by looped echos (three times looped by 6dB attenuation) to complete the story :


the impulse response :

Click the image to open in full size.


a single sine burst (rectified):

Click the image to open in full size.


and 3D plot plus animation :

Click the image to open in full size.


http://www.kinotechnik.edis.at/pages..._echo3_1ms.gif (7MB)





Michael

Last edited by Pano; 13th April 2011 at 02:49 AM. Reason: image overload
  Reply With Quote
Old 12th April 2011, 05:54 PM   #7
mige0 is offline mige0  Austria
diyAudio Member
 
mige0's Avatar
 
Join Date: Jan 2005
Location: Austria, at a beautiful place right in the heart of the Alps.
So what we see in that plots – though they have not come out as visually perfect as I would have liked and may be replaced by better ones eventually :

1.) In the presence of CMP behaviour / ASAR patterns - any sound event gets „framed“
This means that at the beginning and the end there is always a „plateau“ for as long as the delay time lasts
For the time slot of delay time (the „opening frame part“) there is no spectral disturbance – same is true for the „CMP tail“ (the „closing frame part“) after source shut down.

2.) This „CMP framing“ is basically independent of frequency and as such can not be treated by „simple filter“ measures (just the whole signal can be attenuated / „knocked down“).

3.) The „combing“ of comb filter effects is not happen immediately – its happening delayed.

4.) There is no e-function rise nor tail with CMP behaviour / ASAR patterns – but - this can be sort of mimicked to some extent by looped echoes.

5.) As the „framing“ happening with CMP behaviour / ASAR patterns is „sharp edged“ at the points of spectral „discontinuity“ along the time line, we basically face the sonic specialty of „signal start“ / „signal end“ more than once.
This means that the spectral distribution of „signal start“ is overlaid to the original signal at delay time, adding high frequency content where its not expected to happen – if we can say so for such short events.
Same – to some degree – for „signal end“.
Also - for frequency bands of destructive interference – „signal start“ happens at time slot „signal end“ too.
This is one further specific property with any CMP behaviour / ASAR pattern in addition to the framing „plateaus“ and the combing time slot – though, quite often overlooked.


Taking above conclusions one step further, we easily see that CMP behaviour / ASAR patterns are creating kind of time / frequency grid upon any sound event.

As our ear is sensing sounds basically just in this manner (clues at specific frequency bands at specific time slots) we are of course relaying on the exact relationship of that „time / frequency grid“ in ASAR pattern recognition.
If those „matrix“ get degraded – by jitterering the time axis for example – our perception of ASAR patterns is compromised too.

So it may well be that – while enjoying Bethoven – we may not be busy with solving filter equations but rather comparing the beauty of matixes to each other.

From this point of view, valuing the *quality* of presentation / reproduction may rather be judging integrity and precision of those time / frequency matrices than anything else - but thats a different story.


Michael

Last edited by mige0; 12th April 2011 at 05:56 PM.
  Reply With Quote
Old 12th April 2011, 05:54 PM   #8
mige0 is offline mige0  Austria
diyAudio Member
 
mige0's Avatar
 
Join Date: Jan 2005
Location: Austria, at a beautiful place right in the heart of the Alps.
Anything presented above isn't exactly „new“ of course (- what actually is ?)

BUT

above visualisation of CMP framing / ASAR pattern hopefully may trigger re-thinking of time domain impacts.
This specific sonic pattern is most basically *related to room effects in audio perception* at many, many scales.

Following a lot of IMO fruitless discussions about binaural perception with respect to room effects – audio guys seem to have forgotten that most basic encoding of room effects into an audio stream is based on CMP behaviour – which by itself – isn't related to binaural reproduction nor to stereo mic'ing at all.

To get those room related effects „right“ – meaning - to have a system where things „fall into place“ is a quality of its own, that is (and was ever since) way higher in my list than to get tonal balance right – though – until now, I would not have been in a position to pinpoint what „quality“ I exactly was after.

Besides better understanding of perception of 3-D effects , this specific „time domain quality“ of CMP framing / ASAR patterns may also shade some light as to why some systems may sound gritty – especially when a notch in FR is EQ to flat that actually isnt a notch from „simple filter behaviour“, but is part of a CMP pattern.
Simply put, its applying the wrong tools if one tries to fill up – so called - „deep nulls“ (actually originating from CMP behavour) with a graphic or parametric equalizer.

Also – its a mis-belief that such „deep nulls“ cant be corrected at all – they actually can be corrected 100% to flat by „appropriate measures“ (as shown way earlier) - BUT – in reality, one will run out of steam pretty fast. Here also, CMP behaviour shows that its quite a different animal compared to „normal“ filter behavour.

Also – applying low pass filters do not provide *any* cure to (unwanted) CMP behaviour - as is the mis-belief of someone I usually highly respect – or possibly to be more precise - this kind of cure is the kind of „cure“ only people with „peacemaker personality“ usually find appropriate to „solve their problems“.

Also - it can be seen quite clearly:
If a frequency response plot is showing „deep nulls“ one would assume intuitively that this frequencies are „dead“.
Actually that's not the case at all !
At certain time slots those frequency bands are as alive and intact as any others - its just that at *some time slots* those frequency bands are mute.
Hence – when it comes to CMP behaviour / ASAR patterns, the concept of FR is simply void and misleading.



Michael
  Reply With Quote
Old 12th April 2011, 05:55 PM   #9
mige0 is offline mige0  Austria
diyAudio Member
 
mige0's Avatar
 
Join Date: Jan 2005
Location: Austria, at a beautiful place right in the heart of the Alps.
Having presented „another bunch of pretty pictures“ - what actually to learn from that ?

Basically - as said in the beginning – all above culminates in the question: „how do we hear ?“

If „we hear“ in math terms – well – then CMP behaviour and its basic component „delay“ is just the same filter behaviour than any other filter behaviour in the world – there would be no need to make any fuss around CMP.

BUT

if „we hear“ in terms of pattern recognition – well – then CMP behaviour can tell us:

1. about reflections caused by room boundary
2. about reflections and looped reflections of specific speaker makes
3. about source origin related to single versus multi placed sources

Point one
tells us in most basic terms about „room perception“ and possible identification thereof
(the one that was coded into the music we listen to, as well as the one we listen in)

Point two
tells us in most basic terms about „speaker perception“ and possible identification thereof
(horn honk, pronounced diffraction effects, cone brake up etc)

Point three
tells us in most basic terms about „perception of source origin“ and possible identification thereof
(with respect to what was coded into the music – with possibly some „special effects“ from mixdown – as well as with respect to the specific speaker make we may listen to – like multiway, line source, single or coax source)

All in all we may have learned that we actually „hear“ a pretty specific systemic pattern that has one and the same root cause turning out in different „perceptional scales“.

But of course – as an audio guy with pretty limited and biased perception of audio history once pointed out – „CMP does not cure common cold“


And I may add: neither does ASAR pattern recognition...





Keep swingin'
Michael
  Reply With Quote
Old 13th April 2011, 02:50 AM   #10
Pano is offline Pano  United States
diyAudio Moderator
 
Pano's Avatar
 
Join Date: Oct 2004
Location: Milliways
Blog Entries: 4
Do not load up the page with heavy animation files. Provide links instead.
I have removed the embedded images and provided links.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
giant BIB audio nirvana super 8 nirvana Full Range 26 19th July 2010 10:46 AM
Audio Alchemy Dac Man Alex Marx Digital Source 16 6th March 2010 10:05 AM
POOR MAN Teres Audio Versus Motor guglielmope Analogue Source 0 20th May 2009 09:25 PM
Made two steps forward and three steps back pjpoes Solid State 5 3rd June 2007 02:22 PM
Audio Zone Room Attacked By Giant Squid! Film at 11! Steve Eddy Everything Else 10 18th January 2004 05:30 PM


New To Site? Need Help?

All times are GMT. The time now is 06:40 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2