home studios equipment staff/friends booking/rates for sale forum contact

16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not be

Moderators: kerble, Electrical-Staff

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby newberry on Mon Jun 16, 2014 10:01 pm

rustywire wrote:Notice how in the video, he maxes out the oscilloscope at 20khz, citing it's the ceiling of human hearing, which is false. Some people can hear up to 22khz, but that's not my point here. The point is in music, there are harmonics produced up to and beyond 70khz. Just because you may not be able to identifiably hear them, doesn't mean they cease to exist and influence how you hear everything from 20hz-20khz. Or 22khz. Or whatever you can hear and feel.


Anyone else care to weigh in on this? Is it false that 20kHz is the ceiling of human hearing? Rustywire, could you elaborate on how humans are influenced by audio beyond the 20Hz - 20kHz range? (eta: agreed, there are a few people that can hear up to 22, but 44.1 captures frequencies up to 22kHz)

I also take issue with an engineer saying
"Unfortunately, there is no point to distributing music in 24-bit/192kHz format. Its playback fidelity is slightly inferior to 16/44.1 or 16/48,"

A crude analogy I'm partial to, is like watching a movie shot in 24fps 16:9, then watching it using only half of the frames and 4:3. Sure there's enough information to convey the bigger picture, but it isn't an accurate reproduction of the whole. Increasing the bit depth (or aspect ratio) offers more information, and detail; but it's the sampling rate (fps), or number of snapshots and how they are captured & reconstructed that makes more of a dramatic difference in terms of completeness. However, the eye is much easier to fool than the ear...which in many cases is akin to convincing a human they've been fooled.


Yeah, I do think the ear is more easily fooled. I think anyone could plainly see the missing information if you showed them a 16:9 image, and then a 4:3 crop of the same image. However, if you take audio recorded at 24-bit (the article does say it's better to record at higher bit rates), and then convert it to 16/44.1, how many people would be able to tell the difference?
Last edited by newberry on Mon Jun 16, 2014 10:23 pm, edited 1 time in total.
User avatar
newberry
Lode Star of the Twenty-First Century
Lode Star of the Twenty-First Century
 
Posts: 2470
Joined: Thu Sep 28, 2006 1:30 pm
Location: Los Angeles

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby Anthony Flack on Mon Jun 16, 2014 10:11 pm

newberry wrote:Anyone else care to weigh in on this? Is it false that 20kHz is the ceiling of human hearing? Rustywire, could you elaborate on how humans are influenced by audio beyond the 20Hz - 20kHz range?


I think if that's a significant factor then we are going to have to rethink everything, including microphone and speaker design. Because nothing in the signal chain is designed with the accurate reproduction of ultrasonic frequencies in mind.
Anthony Flack
Greatest Man Who Ever Lived
Greatest Man Who Ever Lived
 
Posts: 8229
Joined: Fri Aug 03, 2012 12:27 am
Location: New Zealand

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby newberry on Mon Jun 16, 2014 10:28 pm

Here's a long discussion about human hearing range:

http://www.harmonycentral.com/forum/for ... 0/274313-/
User avatar
newberry
Lode Star of the Twenty-First Century
Lode Star of the Twenty-First Century
 
Posts: 2470
Joined: Thu Sep 28, 2006 1:30 pm
Location: Los Angeles

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby Rodabod on Mon Jun 16, 2014 11:00 pm



This is horse shit, and the guy is speculating. They would need to post a graph of the CD player and turntable playing silence as a comparison for a kick off (to show noise / dynamic range), but even then there are many more factors at play.
User avatar
Rodabod
Heaven-Sent Hero
Heaven-Sent Hero
 
Posts: 6127
Joined: Sat Apr 10, 2004 2:14 pm
Location: London

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby newberry on Mon Jun 16, 2014 11:41 pm

Rodabod wrote:


This is horse shit, and the guy is speculating. They would need to post a graph of the CD player and turntable playing silence as a comparison for a kick off (to show noise / dynamic range), but even then there are many more factors at play.


From that link:
Ok, so here are 2 plot spectrums from Audacity. The one that cuts off at around 22kHz is obviously the CD rip (top one). The other one is a vinyl rip. Same song, same album, same mix, and as far as I know, same master.


Ah, could be comparing different masters. The author of the article in the OP may not be right (though AFAIK no one has proven any of his facts wrong), but he certainly has a better grasp of the scientific method than "ploppy666."
User avatar
newberry
Lode Star of the Twenty-First Century
Lode Star of the Twenty-First Century
 
Posts: 2470
Joined: Thu Sep 28, 2006 1:30 pm
Location: Los Angeles

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby righthanded on Tue Jun 17, 2014 12:14 am

It seems totally worthless to compare a vinyl rip to a cd rip and not compare either to the master recording.
righthanded
suspicious flashlight
suspicious flashlight
 
Posts: 1387
Joined: Wed Aug 16, 2006 3:38 am
Location: elizabeth, nj

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby andyman on Tue Nov 21, 2017 7:42 pm

My coworker just claimed he bought some headphones that let you hear up as far as 44kHz, since that's what the frequencies go up to, apparently.
offal wrote:I just want to be able to play bitchin' single-stroke rolls before I die
User avatar
andyman
Humankind's Greatest Musical Genius
Humankind's Greatest Musical Genius
 
Posts: 4046
Joined: Sat Apr 15, 2006 9:31 am
Location: Cambridge

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby brephophagist on Tue Nov 21, 2017 8:50 pm

andyman wrote:My coworker just claimed he bought some headphones that let you hear up as far as 44kHz, since that's what the frequencies go up to, apparently.

Do they look like this?
Image
User avatar
brephophagist
suspicious flashlight
suspicious flashlight
 
Posts: 1366
Joined: Sun Oct 04, 2009 1:41 am
Location: of liver

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby Anthony Flack on Wed Nov 22, 2017 2:30 am

Were we having this discussion in multiple places? I remember covering a lot of the same ground with regard to Pono I think it was, and the discussion got pretty acrimonious which is a shame because speaking for myself I got to learn lots of interesting things about Nyquist/Shannon, noise, dithering, the fourier transform and all sorts of related subjects.

I've got consumer equipment, a consumer environment and consumer-grade ears, and would consider myself to be highly susceptible to confirmation bias. The best sound card I have has a dynamic range of 108 dB so already I would only be comparing 16 bit to 18 bit at best*. And I'm probably sitting in, I dunno, 20-30dBA of ambient noise maybe? So I'd have to crank a 16 bit audio file up to peak at maybe 130-140 dBA before the noise floor is apparent above that. Which is like jet engine volume.

So since I didn't trust my ears or my gear I had decided to investigate in the other direction and try to get a feel for what is going on by listening to ever-more downgraded signals and see how far you can push that. And that is quite interesting if anybody wants to try it.

The maths is pretty unequivocal - dithering turns 100% of quantization distortion into regular noise. And dropping bit depth raises the noise floor by 6dB per bit. And that's about it. 144dB of dynamic range is already 36dB more than my hardware can output and more than my ears can handle.

* And according to Wikipedia the best converters only manage 123dB (21 bit) dynamic range due to inherent limitations of integrated circuits.
Anthony Flack
Greatest Man Who Ever Lived
Greatest Man Who Ever Lived
 
Posts: 8229
Joined: Fri Aug 03, 2012 12:27 am
Location: New Zealand

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby bishopdante on Wed Nov 22, 2017 4:43 am

andyman wrote:My coworker just claimed he bought some headphones that let you hear up as far as 44kHz, since that's what the frequencies go up to, apparently.


The rated frequency response of sennheiser hd800 is 14Hz-44kHz, with a tolerance of -3dB.

That basically means that the problems are minimised working within the conventional audible range - not that anybody would actually be listening to ultrasonic frequencies.

When employing a speaker on the upper edge of its capabilities, that tends to be where the resonant modes and distortions occur. When designing a speaker system, just because the driver has an upper performance limit of 4kHz doesn't mean that it won't perform with less distortion crossed over at 1kHz.

By the same virtue, employing any system well within its performance margins usually results in the negative artefacts being minimised.
502 Bad Gateway · nginx/1.0.8
User avatar
bishopdante
Humankind's Greatest Musical Genius
Humankind's Greatest Musical Genius
 
Posts: 4213
Joined: Tue Apr 05, 2011 11:25 am
Location: London

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby MatthewK on Wed Nov 22, 2017 6:58 am

Well said FMBD. It's probably also worth pointing out that listening tests have shown that people are not able to pick between music with audio content above 25 kHz and the same music with those frequencies absent. So if the reason for reproducing those frequencies is supposedly musical, it's wasted effort, just like having a heater in the speaker to accurately reproduce the temperature of the instrument when it performed. Sure it's information from the performance, it's just not part of the musical experience.
User avatar
MatthewK
kareem abdul jabbar
kareem abdul jabbar
 
Posts: 501
Joined: Sun Apr 10, 2011 12:25 am
Location: Tasmania

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby andyman on Wed Nov 22, 2017 7:27 am

brephophagist wrote:
andyman wrote:My coworker just claimed he bought some headphones that let you hear up as far as 44kHz, since that's what the frequencies go up to, apparently.

Do they look like this?


I didn't have the heart to tell him...
offal wrote:I just want to be able to play bitchin' single-stroke rolls before I die
User avatar
andyman
Humankind's Greatest Musical Genius
Humankind's Greatest Musical Genius
 
Posts: 4046
Joined: Sat Apr 15, 2006 9:31 am
Location: Cambridge

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby bishopdante on Thu Nov 23, 2017 2:52 pm

MatthewK wrote:Well said FMBD. It's probably also worth pointing out that listening tests have shown that people are not able to pick between music with audio content above 25 kHz and the same music with those frequencies absent. So if the reason for reproducing those frequencies is supposedly musical, it's wasted effort, just like having a heater in the speaker to accurately reproduce the temperature of the instrument when it performed. Sure it's information from the performance, it's just not part of the musical experience.


Well... depends on the volume of playback. Loads of ultrasonic frequencies could easily make a person much more deaf (temporarily or permanently), and result in less perceived brightness. Certain sorts of metal-dome tweeters and compression drivers are notorious for creating ultrasonic frequencies due to distortion... it's not necessarily what you can perceive.

If you look at the typical distortion plot here for a dynamic speaker: https://www.klippel.de/know-how/measure ... rtion.html it should be noted that where THD is detected for a given frequency, what that means is that other frequencies are being generated octaves above it. Hence a speaker with a limited response can generate frequencies far above the upper limit of what frequencies you can drive it with, and a speaker with a wider frequency response and lower distortion might actually generate fewer high frequencies when driven with a given frequency.

Generally speaking the edges of a speaker's response tend to be subject to resonance and acoustics problems, so using a tweeter with a reasonable response up to 100kHz might not be a bad idea, even though you don't actually want to reproduce 100kHz, and a transducer playing down to 5Hz would equally be in its optimal range playing 60Hz.

Of course, that has absolutely no bearing on what sample rates make music sound like music, or whether your converters actually record or reproduce ultrasonic frequencies, Shannon/Nyquist etc.
502 Bad Gateway · nginx/1.0.8
User avatar
bishopdante
Humankind's Greatest Musical Genius
Humankind's Greatest Musical Genius
 
Posts: 4213
Joined: Tue Apr 05, 2011 11:25 am
Location: London

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby Facundo on Thu Nov 23, 2017 3:38 pm

There's no best sample rate because not every ADC and DAC have the same heart.

But if i had to sell an option in the telemarketing i would say... "Avoid Pi, Avoid Decimals. Infinite Decimals! Infinite errors... EMBRACE YOUR FUTURE WITH 88.2 (Hitler's dirtiest desk kept secret)."

The 88.2kHz are for that magical dithering mystery which is the exact double of the 44.1kHz and the logical source to downrating from, to 44.1 without smoke and sparks. I could add unheardable and untold misteries about, but i've chosen silence as the perfect explaination of this mystery.

https://www.soundonsound.com/sound-advi ... -recording
nuking to the bombs, shooting to the bullets
User avatar
Facundo
laney
laney
 
Posts: 157
Joined: Sat Nov 29, 2014 2:40 pm

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby Anthony Flack on Thu Nov 23, 2017 4:09 pm

The link you posted makes it quite clear that there is no benefit to downsampling by exactly half; it says that modern downsampling algorithms can convert any sample rate equally well.

Modern 'asynchronous' sample-rate conversion is far more sophisticated and works by analysing the source and destination sample rates and working out only the required sample values with huge precision. This achieves a technical performance that is significantly in excess of any real-world converter and very close to the 24-bit theoretical level — and that's achieved with any ratio of input-to-output sample rate. There is no measurable difference in performance between using simple integer ratios or complex ones.


Also, if for some reason your DAC chews up the signal when outputting at a sample rate it doesn't like, that's a problem with the DAC, not the data. It could be solved with a software driver that simply converts the data to the preferred format on the fly before outputting it. This is a point I made when talking about Pono - if it's the DAC that's the problem, then you don't need to remaster all your records at a higher sample rate; you could achieve the same result by simply converting your existing recordings to a higher sample rate, either beforehand, or better, as it's being played back.
Anthony Flack
Greatest Man Who Ever Lived
Greatest Man Who Ever Lived
 
Posts: 8229
Joined: Fri Aug 03, 2012 12:27 am
Location: New Zealand

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby bishopdante on Thu Nov 23, 2017 4:55 pm

Here is a really good reason for using high sample rates: http://www.musicofsound.co.nz/blog/why- ... mple-rates

If you aren't going to modify the audio significantly, 44.1kHz or 48kHz sample rate is absolutely fine - particularly if you are mixing through an analogue desk and doing EQ / dynamics in analogue.

24 bit does make a big difference.

Once you start pitch shifting and time stretching, and performing various sorts of extreme sound-warping, the difference between 192 and 44.1 is chalk and cheese. 44.1 will quite quickly sound like a modem.
502 Bad Gateway · nginx/1.0.8
User avatar
bishopdante
Humankind's Greatest Musical Genius
Humankind's Greatest Musical Genius
 
Posts: 4213
Joined: Tue Apr 05, 2011 11:25 am
Location: London

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby Facundo on Thu Nov 23, 2017 5:21 pm

I was talking of recordings i made or processes i'm envolved, most of all are 24bit or 32-floating point of bitdepth and 88.2 - 96kHz. For the case of classical music most external studios i worked with run DAWs like Pyramix or Magix Sequoia, most of them request 88.2 as samplerate and i thought (I suppose, but I did not ask that question that they are using these PCMs 24/88 to edit mix and master, finishing the product in DSD, something i considered non possible after seeing a pipe organ recording i made on a cathedral released out in red book CD format and in DSD format, thing which blew my mind.

Years ago i was searching for good SACDs of releases I wanted, like the Leo Kottle's armadillo album (really titled "6 and 12 string guitar"), and for the experiment I downloaded the image of the album made with a playstation and still available in rutracker.org. and started transcoding the data to PCM formats, to play the album in a adac with dsd support which i knew a friend of mine has in his studio, was a Mytek 8X192 ADDAs he's selling now. We tried all the transcodes I did to PCM from 192 to 44.1 and the best two were 88.2 and 96kHz in listening. We used Barefoot Micromain27s as stereo and also focal cms65. I have no technical info to support the choice of my taste, only ears and feels assuming that some imperfections in music from the technical signal path the music flows are flattering, although they deceive and almost always more pleasant to listen to than the perfection of the surgical steel of a tuning fork perfectly vibrating that note
nuking to the bombs, shooting to the bullets
User avatar
Facundo
laney
laney
 
Posts: 157
Joined: Sat Nov 29, 2014 2:40 pm

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby Anthony Flack on Thu Nov 23, 2017 6:35 pm

bishopdante wrote:Once you start pitch shifting and time stretching, and performing various sorts of extreme sound-warping, the difference between 192 and 44.1 is chalk and cheese. 44.1 will quite quickly sound like a modem.


If you're pitching down, then you're bringing ultrasonic frequencies back down into the audible range.

But if not, there's no difference in the information content between a lower sample rate with no ultrasonic frequencies, and a higher sample rate with the same high frequency roll-off. They can both be reconstructed as a continuous line at any resolution - which is what the article Facundo links to above is saying - and any shit caused by sound warping algorithms can be fixed by higher oversampling inside the algorithm. You're probably pushing the warping algorithm harder than it's designed to go and it's hitting the limit of its oversampling buffer or something.

I want to mark the distinction between the information content of the data itself and the behaviour of external processes. If the problem is with the data itself, then the only solution is to recapture it, which is what Neil Young was proposing to do with his Pono scheme. But if the problem is simply to do with how various external things deal with different sample rates, that can be fixed with a sample rate conversion of the existing data, providing it's done properly.

Lowered bit depth, with dithering, adds noise, no more and no less. Sounds like noise.
Anthony Flack
Greatest Man Who Ever Lived
Greatest Man Who Ever Lived
 
Posts: 8229
Joined: Fri Aug 03, 2012 12:27 am
Location: New Zealand

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby Boombats on Thu Nov 23, 2017 9:36 pm

Anthony.FLAC
User avatar
Boombats
Analingus Eggnog
 
Posts: 19429
Joined: Tue Oct 24, 2006 2:31 am

Re: 16 bit 44.1 kHz is just fine and 24 bit 192 kHz may not

Postby 24K on Fri Nov 24, 2017 12:47 am

Anthony Flack wrote:If you're pitching down, then you're bringing ultrasonic frequencies back down into the audible range.


No, not unless you are using specialist mic's & recording actual ultrasonic content. It's not going to magically appear in a recording with none.

Anthony Flack wrote:there's no difference in the information content between a lower sample rate with no ultrasonic frequencies, and a higher sample rate with the same high frequency roll-off. They can both be reconstructed as a continuous line at any resolution


There's loads of difference. Pitch down an octave & you've halved the transient response of the recorded material. You're going from 44100 discrete data points per second to 22050.

I think it may be the extra data points that make higher resolution audio, sharper & more focused not necessarily the extra frequency content.
The eye can be fooled with 24 frames a second, the ear - while far more primitive & basic in its construction is more sensitive to time based anomalies. Even a few milliseconds latency can be off putting if playing / monitoring through a DAW. Plus, if using plugins even a few samples can cause noticeable phasing issues if the delay compensation is incorrect.
User avatar
24K
mayor
mayor
 
Posts: 99
Joined: Wed Aug 22, 2012 1:39 pm

PreviousNext

Return to General Discussion

Who is online

Users browsing this forum: Bing [Bot], Google [Bot] and 12 guests