[ab initio]

It was an issue in the analogue age. Keeping the noise floor down low required Dolby filtering for cassette tapes, and LPs had a noise floor much higher than anything we're discussing here.

They still do!

 
Cheers

 

[bigshot]

The irony is that the audiophools who argue that 96dB of dynamic range in digital isn't enough are the same ones who happily sit down and listen to LPs with dynamic ranges that are a fraction of that.

Numbers in the abstract aren't helpful. They need to be put in context.

 

[Thad-E-Ginathom]

Yes, those who still sit down and listen to vinyl (rather than those who are new to it and blown away by the romance) are very used to the noise floor.
 
In the digital world, I am think it a bit odd when I hear claims like, "I changed this blah or that blahblah and noticed in immediate lowering of the noise floor." If I could hear the noise floor at all (at least at non-injurious levels) I'd think my equipment was faulty.
 
The numbers game is something I have never been very good at, and it was only a few days back that I actually realised how close (in "notes") those bigger Hz number are: every octave is doubling of frequency. 24khz to 48khz is just one octave; 48khz to 96khz is one more, and 96khz to 192khz just one more. Or would be, if we could hear them. 

 

[cjl]

Let me give you a little wake up call regarding relative proportions...

Look at the waveform. Do you see the difference between the size of the first half and the second? That gives you an idea of how much sound exists up above 12kHz. It isn't empty. You can crank the heck out of the volume and blow your ears out on those frequencies, but at normal listening volume, you ain't going to hear that stuff, particularly buried under the rest of the spectrum.

My illustration was intended to point out how unimportant the frequencies above 12kHz are to listening to music at normal listening levels, not to deny that those frequencies exist altogether. Some people think that the octave above 10kHz contributes to "treble". Not even close. Having some kind of understanding about how different frequencies contribute to music makes it a lot easier to judge quality improvements. Some things mattter a lot. Some things don't mean jack diddly.

One other quick note. Everyone talks about 12kHz and 14kHz and 16kHz as if they are miles apart. That range is less than three notes difference. "Do Re Me". 10kHz to 20kHz is about one octave. It accounts for about 10% of the full range of human hearing, and the least important part of what we hear by far.

 
1) The waveform in that image is on a linear scale, but your hearing is on a log scale. You can't just point at it and say "look, the wave is small so it's clearly inaudible"
 
2) Above 10kHz does contribute to treble. It's the extreme high end of the range, admittedly, but it is quite audible if there is something wrong in that frequency range
 
3) 12k, 14k, and 16k are very different from a standpoint of audibility - even though they are only 3 notes (or less) apart, the steepness of the rolloff of human perception means that a 12kHz cut will be very audible in some music, while a 16kHz cut would be nearly inaudible in nearly all music.
 
Here's an image of my file's waveform (with the correct 12k high pass filtering in the second half) on a log scale - you can clearly see the amount of high frequency content present in the second half. I'd also like to point out that you chose a sample (intentionally or not) that has relatively little high frequency content - I could relatively easily find several pieces of music that have substantially more content above 10kHz (and where it would be much more audible to cut all frequencies above 12k).
 

 

[cjl]

For example, here's another file processed the same way, with the second half a copy of the first with a sharp highpass at 12k (and this is even in a linear scale, not a log scale, so there's a LOT of high frequency content in this file...). I also took this same file and did the opposite (hard low pass filter at 12k) and listened to it, and it is VERY obvious with this example. I can do a full ABX and post logs if anyone wants, but it's a pretty obvious difference. I didn't try moving the cutoff higher to see where it becomes inaudible, but I suspect given my hearing, that level would be somewhere in the 16kHz range or so (maybe even a bit higher - this file has a lot of high frequency content in it).
 

 

[bigshot]

You keep applying absolutes to what I am saying. I'm not saying frequencies above 12kHz are not present in music or not audible at all. What I am saying is that they are unimportant to the enjoyment of music.

It's above the fundamentals and major harmonics of most musical instruments. It's far beyond the point where humans can't discern pitch any more. It's generally at a very low level compared to other frequencies, making it susceptible to masking. It's at the bleeding edge of human hearing where just about everyone's hearing starts to roll off. Without a direct A/B comparison, it's not likely anyone would even notice it missing.

Even the sub bass (20 to 40Hz) which is poorly served in many home stereos is more important than the octave above the treble range (10-20kHz).

People put WAY too much emphasis on numbers and not enough on sound. 12-14kHz sounds really nice and big and important on paper, but iin practice, it is barely significant at all. The only way to think this stuff is important is to keep looking at numbers on a page and not translate that into real world sound.

The most important frequencies to the reproduction of music are the ones covered by Fletcher Munson. Next in importance are the mids below that. Then the bass and treble at the ends. Then the sub bass. Dead last is the top octave. It accounts for 10% of the range and far less than that of music.

 

[cjl]

You keep applying absolutes to what I am saying. I'm not saying frequencies above 12kHz are not present in music or not audible at all. What I am saying is that they are unimportant to the enjoyment of music.

It's above the fundamentals and major harmonics of most musical instruments. It's far beyond the point where humans can't discern pitch any more. It's generally at a very low level compared to other frequencies, making it susceptible to masking. It's at the bleeding edge of human hearing where just about everyone's hearing starts to roll off. Without a direct A/B comparison, it's not likely anyone would even notice it missing.

But that's precisely what I'm saying - in some music, it is GLARINGLY obvious if you apply a cut above 12kHz. It isn't subtle. In the song you posted, it would be pretty subtle, but that's more a function of the choice of music than anything else. I would not consider a system high-fidelity unless it could accurately reproduce up to at least 16kHz or so (preferably 20), because given certain musical samples, the difference is extremely obvious, and the >12kHz content is NOT at a low level even relative to other frequencies (and important to the enjoyment of the music, in my opinion). I'll post a sample file illustrating what I mean in just a minute...

 

[ab initio]

I'll post a sample file illustrating what I mean in just a minute...

 

 

[cjl]

OK, here are two samples of the beginning minute or so of a song. One has been brickwalled at 12kHz, one is the original flac. I bet more than 90% of people could easily abx these two files (the 10% being those with significant high frequency hearing loss).
 
https://dl.dropboxusercontent.com/u/40020825/Sample1_frtest.flac
https://dl.dropboxusercontent.com/u/40020825/Sample2_frtest.flac
 
EDIT: Here's two more samples (treated the same way - one is untouched, one lowpassed at 12kHz) from the same song, just to give a couple more examples...
 
https://dl.dropboxusercontent.com/u/40020825/Sample3_frtest.flac
https://dl.dropboxusercontent.com/u/40020825/Sample4_frtest.flac

 

[cjl]

Oh, and for what it's worth...
 
foo_abx 1.3.4 report
foobar2000 v1.3.2
2014/05/06 15:28:58
File A: C:\Music\Sample1_frtest.flac
File B: C:\Music\Sample2_frtest.flac
15:28:58 : Test started.
15:29:29 : 01/01  50.0%
15:29:52 : 02/02  25.0%
15:30:01 : 03/03  12.5%
15:30:19 : 04/04  6.3%
15:30:46 : 05/05  3.1%
15:31:00 : 06/06  1.6%
15:31:15 : 07/07  0.8%
15:31:25 : 08/08  0.4%
15:31:31 : 09/09  0.2%
15:31:42 : 10/10  0.1%
15:31:53 : 11/11  0.0%
15:32:16 : 12/12  0.0%
15:32:21 : 13/13  0.0%
15:32:24 : 14/14  0.0%
15:32:28 : 15/15  0.0%
15:32:31 : 16/16  0.0%
15:32:34 : 17/17  0.0%
15:32:37 : 18/18  0.0%
15:32:40 : 19/19  0.0%
15:32:46 : 20/20  0.0%
15:32:57 : 21/21  0.0%
15:33:05 : 22/22  0.0%
15:33:09 : 23/23  0.0%
15:33:13 : 24/24  0.0%
15:33:16 : 25/25  0.0%
15:33:18 : 26/26  0.0%
15:33:20 : 27/27  0.0%
15:33:23 : 28/28  0.0%
15:33:27 : 29/29  0.0%
15:33:30 : 30/30  0.0%
15:33:45 : Test finished.
 ----------
Total: 30/30 (0.0%)
 
(Note the timestamps - the difference isn't subtle, so I could determine which was which within a matter of a couple of seconds for the most part, usually without even playing the reference files)

 

[bigshot]

I'm afraid I can't play Flac files.
 
But again, I'm talking about relative importance. I don't seem to be able to get that across to you. A double blind test completely misses the point of what I'm saying.

 

[cjl]

I know you're talking about relative importance, but you keep asserting quite enthusiastically that the upper octave doesn't contribute much to the music, and I strongly disagree. You keep acting as though your own aesthetic preference and thoughts on which factors matter is the only correct way to set up a system, and because you don't believe that upper octave is important, it clearly must not be important to anyone, with any genre or type of music. All I'm trying to show is that for certain music types and sound samples, this is clearly incorrect.
 
As for the double blind test, it shows that the difference between the two files is clearly audible (especially when you account for how quickly I was able to discern the difference). Once audibility has been shown, it largely comes down to preference (including preference in music style). Honestly, I'd even dispute your assertion that the uppermost octave (10-20kHz) is less important than the lowermost octave (20-40Hz) - I'd rather have a system with rolled off bass below 40Hz, but good treble response up to 20kHz than a system with rolled off treble at 10k but perfect bass. Of course, ideally, you'd have a system with good response 20-20k.
 
(Oh, and if you want a way to play flac files, VLC works quite nicely, and is compatible with Apple, Windows, and most versions of Linux... http://www.videolan.org/vlc/index.html)

 

[ComradeDylie]

   
No it was the whole idea that inaudible frequencies can make music sound better. How are they going to sound at all when they're inaudible?!
 
By the way, there have been studies that have shown that super-audible frequencies have absolutely no impact on sound quality. In fact, music doesn't really contain any, because the only musical instruments capable of producing them are cymbals and triangles, and in those, auditory masking from lower harmonics totally obliterates the ones in the upper range of hearing. Inaudible frequencies would be doubly inaudible!
 
Here is a "for instance" for you... This is an Apple Lossless file of a digital recording of Bach. The first half is the recording as it appears on the CD. The second half is just the frequencies above 12kHz.
 
http://www.vintageip.com/test/freqresponsetest.m4a
 
Silence... Yes, there isn't even anything above 12kHz in a lot of music, much less above 20kHz...

Oh..thats not circular at all.  I didnt say better.  I said more natural.  Perhaps it would be better to say "feel" instead of sound.  Or that we would 'perceive' them to be more natural.  Whatever.  You clearly have old man hearing loss if you cant hear those high frequencies. Me and that other dude can.  Sorry bro, for your sound blindness =(.  lol kidding.  
 
Can we get a piccolo, chimes, and cymbal only track up in here?  Try cutting out the high freqs out of that and tell me if it sounds right...oh wait...sound blind...
 
And you know what else we dont need besides frequencies over 12khz?  Letters past W.  I mean who even uses them?  We only need the first 23.
 
I like how we are talking about bit depth and sampling rates and if it makes a difference.  And then now we are on whether or not for sure, 100% AUDIBLE frequencies matter or not...oh no...
 
On the matter of relative importance:
 
Relative importance of posting on head-fi compared to anything that actually matters:  nil.  But here we are.  So....exercise?  Grow a garden?  Develop new clean fuels? Learn something new?  Play with your kids?  Cook dinner instead of eating McDonalds...again!? Brush your teeth?  Vacuum?  Dust?  Help cure diseases?  Idk...its all relatively important stuff eh?  But here we are.

 

[Xenophon]

  OK, here are two samples of the beginning minute or so of a song. One has been brickwalled at 12kHz, one is the original flac. I bet more than 90% of people could easily abx these two files (the 10% being those with significant high frequency hearing loss).
 
https://dl.dropboxusercontent.com/u/40020825/Sample1_frtest.flac
https://dl.dropboxusercontent.com/u/40020825/Sample2_frtest.flac
 
EDIT: Here's two more samples (treated the same way - one is untouched, one lowpassed at 12kHz) from the same song, just to give a couple more examples...
 
https://dl.dropboxusercontent.com/u/40020825/Sample3_frtest.flac
https://dl.dropboxusercontent.com/u/40020825/Sample4_frtest.flac

Indeed easy, not a subtle difference.  At least in this sample it's clear that removing the >12kHz frequencies makes the music somehow soulless.

 

[ab initio]

Bigshot,
 
How much do you think that one's personal preferences in music affect what aspects of sound reproduction they value the most? If I understand correctly, you are particularly fond of classical music, which, I think, tends to have much less content in the highest octave of human hearing, than, say, modern rock or metal, which feature much more percussion (notably the cymbal crashes).
 

The most important frequencies to the reproduction of music are the ones covered by Fletcher Munson. Next in importance are the mids below that. Then the bass and treble at the ends. Then the sub bass. Dead last is the top octave. It accounts for 10% of the range and far less than that of music.

That said, I definitely agree with this ranking of relative importance. For reference, I listen to music which varies from late baroque through modern classical to 60s and 70s classic rock and progressive rock through modern progressive rock, progressive metal, and hard rock. I don't listen to really any pop, hip-hop, country, or electronic anything. I don't know how those genre's might affect my priorities for sound reproduction. Maybe folks who favor pop, hip hop, and/or electronic dance type genre's might place bass quantity (and quality? maybe not?) higher up in their priorities? I'm not sure if there are electronic genre's that would use oodles of super high frequency synthesized sounds? I would find that painful, but the dog might rock out to that.
 
It's sort of funny (at least, I get a huge kick out of the fact) that the least important frequencies in music are the ones that make music files so stinking big. It's the highest octaves that take up all the disk space, just for the sake of satisfying Nyquist theorem for those frequencies. Every extra octave makes the file double in size. 
 
Because this thread is lacking in sample calculations (at least to my satisfaction), let's do one:
 
If we assume that we can hear from 20Hz to 20kHz, then we can hear a 10 octave range. The octaves are (in Hz):

 Octave: Start: End: Nyquist:            1          20          40          80            2          40          80         160            3          80         160         320            4         160         320         640            5         320         640        1280            6         640        1280        2560            7        1280        2560        5120            8        2560        5120       10240            9        5120       10240       20480           10       10240       20480       40960

 
Every time another octave needs to be captured in a recording, the resulting file size doubles. For the less technically inclined (do they every visit this forum?) we call a doubling in size with a unit increase "exponential growth", which is the the mathy way of saying something gets awfully expensive in a hurry. The theoretical minimum size of an uncompressed PCM file (per second) is the [Nyquist frequency] times [the bitdepth of the recording] times [the number of channels e.g., 2 for stereo].
 
If we ignored the top octave, and only recorded music up to 11025 Hz (i.e., sampled at 22.05 kHz), we would cut down the file size by 50%. If the music is band limited below 11kHz, then the recording  with sampling rate of 22050 Hz would still have exactly the same fidelity as a recording at 44100 Hz (or 88.2kHz, or 192 kHz, or 32 million THz, or whatever 30x super DSD is, etc.). This is throwing out 1 octave out of 10, i.e., the 10% of the range that bigshot mentioned above. Adding additional octaves beyond 20kHz extends the "musical range" by only a small fraction ("musical range" is in quotes, because nobody can even hear those frequencies) at the expense of doubling the file size for each extra inaudible octave.
 
176.4 kHz audio increases the octave range of the recording by 20% over the 44.1kHz version, all of it inaudible to humans, at the expense of increasing the file size to 400%. Your pet bats will thank you, though!
 
This is not a "glass is half full" vs "half empty" kind of thing. (well, one might say that 96kHz recordings are half full). For HiRes formats, this is a "glass is twice as big as they need to be (or four times, etc..)" kind of thing.
 
I recommend that we change the term "High Resolution audio" to "Highly Inefficient audio". I bet it will sell even better! You could shorten it to "High-In" audio. I wonder how this will go over in the other forums.
 
Cheers