Head-Fi.org › Forums › Equipment Forums › Sound Science › What makes piano sound so hard to reproduce?
New Posts  All Forums:Forum Nav:

What makes piano sound so hard to reproduce? - Page 8

post #106 of 191
Quote:
Originally Posted by Chris J View Post

Just gotta say, a very thought provoking post!

Too bad that the mics linked to roll off at 20 kHz, some quite steeply.

 

@bigshot: Same.

 

I'm still waiting for something other than anecdotes.


Edited by xnor - 12/30/12 at 12:50pm
post #107 of 191
Quote:
Originally Posted by xnor View Post

Too bad that the mics linked to roll off at 20 kHz, some quite steeply.

 

@bigshot: Same.

 

I thought this was the Science Forum, not the sarcastic rude comments forum.

post #108 of 191
Im sorry if I offended any microphones.
post #109 of 191
One thing that is interesting about folks who have problems in sound science... They always seem to speak about specific brands of equipment and quote sales literature. Their opinions revolve around products. You don't see that as much among the SS regulars.
post #110 of 191
Quote:
Originally Posted by Chris J View Post

 

I thought this was the Science Forum, not the sarcastic rude comments forum.


that doesn't mean acting like a yes-man. Criticizing is always welcomed especially when people affirm things without measurements or scientifical tests.

post #111 of 191
I'll post a quick comment about LPs... When cutting an LP master, they usually rolled off the higher frequencies, because above 15kHz or so, the groove modulations would be so delicate, a few plays with an elliptical stylus would wear them down to a mush of distortion, particularly at the inner grooves where there is much less surface area and reduced resolution. My brother had a test record he got with his Macintosh system that had test tones up to 20kHz. The highest frequency tones were shot in about ten plays.

There are high frenquencies on records however. It's called surface noise and distortion.
post #112 of 191
Quote:
Originally Posted by bigshot View Post

There are high frenquencies on records however. It's called surface noise and distortion.

 

Foul, ref. !!! Low blow there.biggrin.gif

 

w

post #113 of 191
I apologise if I have offended any LPs.
post #114 of 191
Quote:
Originally Posted by Puranti View Post


that doesn't mean acting like a yes-man. Criticizing is always welcomed especially when people affirm things without measurements or scientifical tests.

Thanks for reminding me how much I hate the science fiction forum, I can't take the oversized egos and pontificating opinions.
post #115 of 191
Quote:
Originally Posted by Chris J View Post


Thanks for reminding me how much I hate the science fiction forum, I can't take the oversized egos and pontificating opinions.

There are pros and cons to everything. For me, everything except the sound science forum is filled with opinions based on naive fantasies. 

post #116 of 191
Quote:
Originally Posted by Chris J View Post


Thanks for reminding me how much I hate the science fiction forum, I can't take the oversized egos and pontificating opinions.


Well I you take those critics as oversized ego or pontificating opinions that's too bad for you. I'm myself a complete ignorant in sound science but I know how to recognize a valid argument from another, and if I'm wrong I would be glad to be corrected.

Science is not a matter of ego.

post #117 of 191
Quote:
Originally Posted by bigshot View Post

I'll post a quick comment about LPs... When cutting an LP master, they usually rolled off the higher frequencies, because above 15kHz or so, the groove modulations would be so delicate, a few plays with an elliptical stylus would wear them down to a mush of distortion, particularly at the inner grooves where there is much less surface area and reduced resolution. My brother had a test record he got with his Macintosh system that had test tones up to 20kHz. The highest frequency tones were shot in about ten plays.
There are high frenquencies on records however. It's called surface noise and distortion.

Interesting comment about rolling off above 15KHz for LP masters.  We never did that in any of the LP mastering sessions I was involved with, nor was excessive HF wear even considered.  Our master tapes were flat to 20KHz, so level set was important, but no HF limiting or rolloff was used.

 

Don't forget CD-4 (the only accepted discrete 4 channel LP), introduced in 1972, which had an FM carrier recorded at 30KHz with bandwidth to 48KHz. You could destroy the carrier by playback with the wrong stylus type (if I recall, a conical stylus did really bad things), but with the right one it hung in there pretty well.  The CD-4 technology didn't last, but it did serve to improve conventional stereo LPs, but in cutting and stylus shapes. The Shibata stylus, for example, could play up to 50KHz and not wear the record up to a few hundred plays.

 

HF wear was a function of stylus shape and condition along with tracking force. The conical stylus had a small contact surface which resulted in high surface contact pressure and high record wear. The elliptical stylus was the first improvement, and provided a larger contact surface with much lower contact pressure for the same tracking force. 

post #118 of 191

I do agree that any claim SHOULD be supported by evidence. 

 

We live in a real world, resources like time, money, etc are limited. Hope this statement does not sound like anecdotal evidence to some.

 

I would never post something I have no experience with and can not be confirmed  by listening. Another matter is how to produce scientifically valid evidence. Not that can not be done, in theory at least, trouble is, as usual , it looks expensive.

 

It looks so expensive no real time or "slow time" analyzer processing data acquisitioned real time for whatever usable amount of time that could at least approach the complex matrix of human hearing perception has been not even attempted at - at least to my knowledge. Dynamic range of human hearing is 120 dB or so; hope we agree on this one. Meaningful frequency response that does affect our perception is from DC to approx 100 kHz. Here, some of you will probably go ballistic. Hold the horses for a moment - I will try to explain.

 

Whenever you are looking at some scientifically verifiable result regarding sound, it will be, in vast majority of cases, static. Like constant frequency at constant level etc. The most information regarding real life performance of any device that is still generally menageable to  (re)produce you would get from the sqaure wave test. The shape of square wave is extremewly sensitive for either amplitude or phase anomalies - the moment anything is off by the slightest amount, square wave will no longer appear square. Just look at the measurements of square wave of headphones - these look so bad that about highest frequencvy one can talk about that still resembles anything approaching square wave is 300 Hz or so - try 1 kHz and it will turn in most cases into something you would generally not thought of as it has any relation to square wave at all. 

 

We were at CD vs vynil before; why do you think is the test frequency of square wave for CDs 400 Hz vs analog's 1 kHz - NO prizes for guessing that.

In short, a decent, not TOTL of TOTL... of TOTL phono cartridge will always have better square wave response than any RBCD no matter the cost. TOTL of TOTL of...TOTL phono cartridges even challenge DSD at 5,6 MHz in this regard  - no PCM, including 192/24, comes even close!

 

Yet - even square wave is in a way static - it does not change its amplitude or frequencies, like any real music does all the time ( with the exceptions of some really long notes, but these are more exceptions confirming the rule ). You get the picture where I am hinting at - building an analyzer that could lock onto any shape/frequency/amplitude signal and analize it for THD, noise, frequency response, etc, etc - would cost dearly.

 

What we have now is a pretty good set of static measurements that are generally limited to 20 Hz - 20 kHz bandwidth. Yet the mechanisms that will seriously affect analog playback lie way lower and way higher than that. Below 20 Hz, there WILL be resonance due to compliance of the stylus suspension interacting with the mass of the cartridge plus the effective mass of the pick up arm. Above 20 kHz, sometimes WAY WAY above 20 kHz,

will be resonances in stylus cantilever, resonance due to stylus effective mass yielding against the vynil groove which is not ideally rigid but does have its compliance - etc. My experience is that if you take care of response below 20 Hz, all the way to DC, or in case of turntable spinning at 33 1/3 RPM down to 0,55 HZ  - and response above 20 k, you generally get really decent 20-20K - so to speak for free.

 

I have enough of test records to perform any measurements save the pulse response - that JVC disc cut not at half speed mastering but one tenth of speed mastering  to my knowledge was never commercially available, JVC used it to develop its pick up cartridges for CD-4 quadrophonic system. I do not have any more equipment but an 20 MHz analog oscilloscope - allowing me to see the results, but I can not produce hard copy etc. I would have been delighted to have the measuring setup as described here : 

 

http://www.theanalogdept.com/images/spp6_pics/TT_Design/MechanicalResonances.pdf

 

I have never seen this setup or anything approaching it in real life - manufacturers tend to limit themselves to the bare essentials like frequency response and channel separation 20Hz - 20 kHz. And even these charts are included mostly with quite expensive cartridges. 

 

Today, using PC and some software, it would probably be possible to cut the cost of the setup appreciably - but in any case, we are still talking five figures of $/Euro. The same/similar repeats itself for measuring headphones - that HEAD Acoustics artificial head Tyll Harstens uses costs alone in excess of 20K (depending on the model) , add Audio Precision analyzer, etc, etc, etc - sorry , do not have the kind of money required to do that, let alone justify it as bussines/commercial proposition. Big commercial audio/video magazines do not do it - too expensive. Producing anecdotal evidence in form of subjective impessions on sound quality instead.

 

Yes, I have heard of digital storage oscilloscopes and likely of any gizmo imaginable that could provide me with the capability to produce scientifically verifiable results. But I tend to stash every available cent in better equipment either for vynil playback or music recording than to invest unnecessarily into measuring equipment that would in effect be used mostly to prove my sonic impressions. I hear as I hear, I have accumulated enough listening to either live or reproduced music to know what does sound good and approaching the real thing and what does not. And know at least some corelations between measurements I can see on the display of an analogue oscilloscope and sound played back either through headphones ( mostly  used ) or speakers. To some, these results will forever be considered anecdotal evidence - as I said above, I lack the funds to measure my findings to the level it would allow them to become scientifically verifiable, repeatable etc,etc as required by the measuring crowd - not the willingness to do it, or willingness to get additional education I openly admit to lack to some degree or another - but claims I made should definitely be possible to prove beyond any shadow of the doubt, given the conditions mentioned above can somehow be met. Large corporations, like Sony, have lost a teth or two - in proving SACD was better than CD, eventually abandoning the support for SACD/DSD due to commercial reasons - how on earth you can ask of an individual to produce proof or evidence if Sony with its resources could not do it ? They have the money, they just went the wrong way about proving its superiority - and failed. I perhaps know which points they have overlooked - no money to prove it - moot point. 

 

What really lies behind it all is the fact that in order to support the extended bandwidth, consumers would have to replace most, if not all, of their equipment for reproduction. Sony did introduce an audio system capable of playing back DSD/SACD back in the day - best part were the tweeters good to aboout 100 k or so, rare as hen's teeth, even rarer seen for sale, expensive as hell - but GOOD. And then the mastering engineerr of your favourite musician happens to be a die-hard 20-20k guy and any gains your machine has can not be put to use - that is what in effect happened in the real world. Sales low > cancellation of the whole programme. Only after that did Sony allow Korg and others to introduce DSD recorders to the general public.

 

I decided to deviate from my anedcdotal evidence ways in case of burn-in for headphones, because flaming I received for posting my method on head-fi was too much for me to let it go  - by splurging some cash I would heve certainly preffered to put to better use by arranging for the measurements of almost ( less than 5 hour play time ) new headphones before and after the 48 hour burn-in process.

 

Bruel & Kjaer - with everything documented, down to the serial numbers of the very equipment used for the test. 

 

Will be posted ASAP,  pending on the delivery of the results.

post #119 of 191

Square waves, really?

 

The problem I see is that you keep on posting walls of text with claims, anecdotes and a lot of irrelevant stuff. When somebody dissects your posts, all you do is moving on to a bunch of new claims etc. posting another wall of text, instead of actually addressing the criticism.

Maybe you just like to see your own writing, or it is cognitive dissonance or you just don't care.

 

Anyway, looking forward to your burn-in measurements. Repeatable, right?

post #120 of 191

Square waves.  Really. 

 

I've tested things with square waves for decades.  Everything from tape recorders to broadcast transmitters to telephone lines, digital recording systems, to speakers, mics, amps and preamps.  The one constant to all of that is, very few devices can pass a mid-band frequency square wave without significantly altering its appearance. But there are two things to know about that.  First, to pass a mid-band square wave unaltered requires extremely wide frequency response with extremely flat phase response.  But, second, and this is key, the relationship between how a reproduced square wave looks and how it sounds is extremely fuzzy at best, and in limited bandwidth systems the appearance of a square wave not an indication of how that device will sound. A perfect square wave requires response DC to well into the MHz, because the square shape requires a theoretically infinite number of odd-order harmonics summed together without change to amplitude phase.  Square wave testing is therefore hypersensitive to bandwidth and phase.  It's already been demonstrated that human hearing is relatively insensitive to non-flat phase response.  In practice, the actual appearance of a reproduced square wave is less important than the analysis of its harmonic content in terms of amplitude and phase.  Today there are other measurement methods that measure amplitude and phase response directly.

 

No bandwidth limited recording system, analog or digital, will reproduce a clinical square wave, though high sample rate digital systems come much closer than any analog system. 

 

There are no acoustic square waves in nature.  

 

Last comment, not square wave specific... having response down to DC on a turntable is just a bad idea.  The proof goes back to research done in the 1970s that relates extreme subsonic response with higher levels of IMD, and even causes doppler shift of mid band signals because of woofer cone movement in response to rumble.  No need to go into the sources of rumble, except to say you can't avoid it.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › What makes piano sound so hard to reproduce?