or Connect
Head-Fi.org › Forums › Misc.-Category Forums › Music › vinyl rip vs cd
New Posts  All Forums:Forum Nav:

vinyl rip vs cd - Page 12

post #166 of 334

@Brunk, (and anyone who might misinterpret those graphs)

 

The graph on the left is fundamentals only.

For example, cymbals, if fundamentals go up to 16Khz, there will be harmonics at 32kHz, at 64kHz, at 128kHz etc, almost to infinity. That's how sound works in real life, that's what we hear when someone bashes a cymbal. Analog.

 

Hence the importance of recording in analog, mixing in analog, mastering in analog, and if it has to be converted to digital at the end of the chain, it's important to keep as much high end (natural harmonics) in there, to keep the 'real sounding sound'. Sampling at 96kHz will allow frequencies up to 48kHz to remain relatively intact. That will sound "pretty realistic". Sampling at 192kHz will allow frequencies up to 96kHz to remain relatively intact. So those cymbals will sound realistic.

 

That's how it works.

 

If you sample at 44.1kHz, every harmonic over 22.05kHz will be missing, so cymbals, violin, trumpet etc, will be missing harmonics on into infinity.

It also affects how we perceive the lower harmonics, as Bob Katz explains in his book (see earlier in this thread).


Edited by ploppy666 - 11/10/13 at 1:00pm
post #167 of 334
Quote:
Originally Posted by ploppy666 View Post
 

@Brunk, (and anyone who might misinterpret those graphs)

 

The graph on the left is fundamentals only.

 

You aren't reading it correctly the red part of the bars are the fundamentals. The yellow part is the audible harmonics. Above a certain number of harmonics, they become inaudible because the higher harmonics exist at a much reduced volume, or because of the acoustic principle called auditory masking. Only a few audible harmonics exist about 12kHz, and even those trail off before the 20kHz threshold of human hearing. (which is exactly what that chart indicates.)

 

Above 5kHz, humans can't perceive pitch any more, so the sound we are talking about isn't musical. It's more a matter of ringing in the ears. Human perception of harmonics does not extend to infinity. Sound engineers usually work towards preserving 3 levels of harmonics. Beyond that, audibility falls off sharply.

 

Super high frequencies can cause problems. Frequencies near the edge of perception can cause headaches and listening fatigue if they are too high. Super high frequencies beyond that can cause harmonic distortion extending backwards into the audible range in most consumer audio equipment.


Edited by bigshot - 11/10/13 at 12:58pm
post #168 of 334

My mistake about cymbals.

 

But many instruments have harmonics above 20kHz, including cymbals.

post #169 of 334
Quote:
Originally Posted by bigshot View Post

 

Above 5kHz, humans can't perceive pitch any more, so the sound we are talking about isn't musical. It's more a matter of ringing in the ears.

 

Why say that? People who don't know anything about audio might think we can chop off everything above 5kHz. We can't, it would sound horribly dull.

post #170 of 334
 
Super high frequencies beyond that can cause harmonic distortion extending backwards into the audible range in most consumer audio equipment.

 

This is called aliasing, correct?

post #171 of 334

double post

post #172 of 334
Quote:
Originally Posted by ploppy666 View Post
 

My mistake about cymbals.

 

But many instruments have harmonics above 20kHz, including cymbals.

 

Not audible ones. I had the luxury of having access to a full professional grade recording workstation. I played around with SACDs with super audible frequency content. I tried ducking various frequencies to see how it affected the music. Although I could hear certain high frequencies if I isolated them all by themselves, when I ducked them using the equalizer in music, I found that it didn't make any difference. I did a little research and found that this was due to a psychoacoustic principle called "auditory masking". In a nutshell, the concept of masking is that boosted level in one frequency will cover up frequencies and octave above. In cymbals, this is important, because there is loud harmonics all over the place above the fundamental. After one or two levels of harmonics, masking kicks in, and even though there is information at higher harmonics, human ears can't hear them.

 

Just because you can measure a frequency, it doesn't mean that you can hear it.


Edited by bigshot - 11/10/13 at 1:06pm
post #173 of 334
Quote:
Originally Posted by ploppy666 View Post
 

 

Why say that? People who don't know anything about audio might think we can chop off everything above 5kHz. We can't, it would sound horribly dull.

 

That isn't what I said. I said we can't perceive pitch any more. That just means it isn't a part of the music. Above that, it's the nuances of the sound of the instruments. It's the difference between a car radio or boom box where you can hear the music being performed, but with a full range system you can hear the bowing on the strings or subtle harmonics on percussion.

 

It really isn't horribly dull, because we listen to limited range audio all the time in our daily life on telephones, table top radios, TV sets, etc.

post #174 of 334
Quote:
Originally Posted by Tus-Chan View Post

 

This is called aliasing, correct?

 

It has to do with distortion in the amplifier circuitry... I'm not that technical a person. Here is an article on it... http://productionadvice.co.uk/high-sample-rates-make-your-music-sound-worse/

post #175 of 334
Quote:
Originally Posted by bigshot View Post
 

 

Not audible ones. I had the luxury of having access to a full professional grade recording workstation. I played around with SACDs with super audible frequency content. I tried ducking various frequencies to see how it affected the music. Although I could hear certain high frequencies if I isolated them all by themselves, when I ducked them using the equalizer in music, I found that it didn't make any difference. I did a little research and found that this was due to a psychoacoustic principle called "auditory masking". In a nutshell, the concept of masking is that boosted level in one frequency will cover up frequencies and octave above. In cymbals, this is important, because there is loud harmonics all over the place above the fundamental. After one or two levels of harmonics, masking kicks in, and even though there is information at higher harmonics, human ears can't hear them.

 

Just because you can measure a frequency, it doesn't mean that you can hear it.


This is massively misleading to people outside of the music business. If we can't hear frequencies above 20kHz, why bother sampling at 44.1kHz? Why not sample at 20kHz?

Answer: Because it's NOT THAT SIMPLE. Sony and Phillips decided to make CDs at 44.1kHz because only then would 22.05kHz make the cut. Just.

 

Why do top studios record audio at 192kHz today, if we can only hear up to 20kHz? Again, because it's not that simple.

 

Post-production digital processing such as filtering, compression and equalisation creates distortion in the audible band (up to 20kHz) because of digital errors. If you double the sample rates, those errors are spread over twice the bandwidth, so half of the errors will be in the inaudible bandwidth (above 20kHz). Therefore, the higher the sample rate, the less errors will be audible. Every time you double the sample rate you halve the audible errors.

Theoretically even 192kHz is not enough, i.e it's not perfect. Digital probably never will be.

 

One thing we should be able to agree on (and I don't understand why you supposed audiophiles won't agree on this) is that 44.1kHz is way too low.


Edited by ploppy666 - 11/10/13 at 1:23pm
post #176 of 334
Quote:
Originally Posted by bigshot View Post

 

It really isn't horribly dull, because we listen to limited range audio all the time in our daily life on telephones, table top radios, TV sets, etc.

Again, you're being very misleading.

 

I challenge you to do a low pass at 5kHz on any piece of music and claim "it's not horribly dull".

 

Are you really that deaf?

post #177 of 334
Quote:
Originally Posted by ploppy666 View Post
 


This is massively misleading to people outside of the music business. If we can't hear frequencies above 20kHz, why bother sampling at 44.1kHz? Why not sample at 20kHz?

Answer: Because it's NOT THAT SIMPLE. Sony and Phillips decided to make CDs at 44.1kHz because only then would 22.05kHz make the cut. Just.

 

Why do top studios record audio at 192kHz today, if we can only hear up to 20kHz? Again, because it's not that simple.

 

The reason that the redbook standard extends up above 20kHz was because back then, filtering out the noise generated above the audible range required a rolloff that wasn't a precise "brick wall". They allowed a little extra so the rolloff wouldn't extend into the audible range. Today, oversampling DACs have improved upon this, and redbook sound is boosted to a rate where a more accurate brick wall filter can be applied, then output as normal redbook. They didn't include frequencies above 20kHz in the redbook standard because they wanted people to actually hear them. It was a technical compromise.

 

Studios require a little bit more latitude for mixing. But most don't record at 192 because it causes all kinds of problems with the analogue elements of their systems. Here is a more detailed explanation if you are interested.

 

http://people.xiph.org/~xiphmont/demo/neil-young.html

post #178 of 334
Quote:
Originally Posted by ploppy666 View Post
 

Again, you're being very misleading.

 

I challenge you to do a low pass at 5kHz on any piece of music and claim "it's not horribly dull".

 

Are you really that deaf?

 

Turn on your TV set. Does that sound horribly dull to you? How about your car stereo?

 

I provided you with some articles to read and digest. I'd suggest you do that before you start trying to criticize what I'm saying.

post #179 of 334
Quote:
Originally Posted by ploppy666 View Post
 


This is massively misleading to people outside of the music business. If we can't hear frequencies above 20kHz, why bother sampling at 44.1kHz? Why not sample at 20kHz?

Answer: Because it's NOT THAT SIMPLE. Sony and Phillips decided to make CDs at 44.1kHz because only then would 22.05kHz make the cut. Just.

 


44.1 kHz is the sampling rate, which determines the maximum frequency that can be encoded via the Nyquist-Shannon theorem, but there's more than that to it. It means that there are 44.1 thousand digital "snapshots" of the analog signal per second. The effect of sampling rate is not limited to the maximum sonic frequency that can be encoded -- it also has to do with how accurately the analog signal is represented digitally on the whole. It's a matter of resolution in general. Think of Reimann sums -- you get a more accurate approximation of the integral as the number of bars per unit increases. 


Edited by manbear - 11/10/13 at 1:50pm
post #180 of 334
Quote:
Originally Posted by analogsurviver View Post
 

 

98 dB divided by 14 microsec = 7 dB/microsec 

 

An average high quality phono cartridge, either decent MM or MC, has rise time of 10 microseconds or less

 

78 dB divided by 10 microsec = 7.8 dB/microsec

78 dB divided by   6 microsec =  13 dB/microsec

 

The best cartridge I had the pleasure to measure and audition, had a rise time of 3 ( in a word : three ) microseconds. Sound - out of this world...

 

 

If this makes such a difference how come hardcore vinyl pundits like Uncle Ivor are incapable of detecting the presence of a simple A/D/A loop which nobbles this performance back down to red book limits and yet apparently captures all the analog goodness clicks and pops included ?

 

Unless you are playing square waves this improved rise time is simply irrelevant and is irrelevant anyway as we hear any square wave above  10K square wave as a 10k sine wav as all the harmonics (which start at 30k) are inaudible.

 

A while back I seriously posted a question on RAO about the better rise time of LP, I could not find anyone who could empirically show that this was in any way audible i.e the rate of signal change that both CD and LP are capable of are superior to our ability to detect them. So while LP can render these changes better we cannot hear to difference otherwise the LP/LP and A/D/A comparisons would be a slam dunk for LP.

 

In fact we know that our ears are far less sensitive to short-duration transients

 

In fact the shortest sound we can perceive at all is about 22 microseconds, what do we hear this as, not music certainly we hear it as a pop or click, but put two together and you will only hear one, i.e you cannot resolve two adjacent 22 microsecond pulses, this is due to the our hearing which integrates sound events to about 50 milliseconds so you can hear a 5hz square wave as a set of separate impulses (clicks) but even a 50hz square wave is perceived as continuous, so however impressive your cart may be is entirely moot.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Music
Head-Fi.org › Forums › Misc.-Category Forums › Music › vinyl rip vs cd