• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

High Resolution Audio Controversy on Wikipedia

Joined
Nov 14, 2018
Messages
18
Likes
6
#1
I know this is a little random, but I was browsing Wikipedia's pages on High Resolution Audio, DVD-A, and Super Audio CDs recently and noticed something a little interesting: The SACD page leads with mentioning the Audio Engineering Society's double blind test that demonstrated that people could not discern a sound quality difference between CD and SACD quality. Now, if you check out the High Resolution Audio page there is mention of a 2016 Meta-Analysis that demonstrated that among listeners there is a small but statistically significant ability to discriminate between standard quality audio and high resolution audio, and when subjects are trained their ability to discriminate between formats improved dramatically. And If you head on over to the DVD-A page this Meta-Analysis is once again mentioned, alongside a new reference to a paper on the "hypersonic effect", which is essentially a phenomenon where our brain is actually affected by inaudible high-frequency sounds and contributes to improving the overall perception of music.

Now, with all this in mind, I thought to myself "well wait a second, if we know that High Resolution Audio is a highly controversial subject, and there is actually a decent amount of academic content out there nowadays that proves that High Resolution Audio can be reliably distinguished by people, why does the SACD page lead with mentioning that paper? Isn't that a little biased? And why are the papers mentioned on other Wikipedia pages about high-resolution audio not mentioned here?" so I headed on over to Wikipedia's SACD discussion page and was a little amused to see this exact argument going on.

What are your thoughts? Personally I used to be a little skeptical of High Resolution Audio, but now that I've taken some courses on Media Signal Processing (currently studying Digital Media in university :D), my views on the topic have changed. I'm happy to see that for the most part Wikipedia discusses High Resolution Audio in a positive light, but the introduction to that SACD page feels kind of dirty imo, and I would like to see it changed.
 
Last edited:

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
25,125
Likes
52,287
Location
Seattle Area
#2
I have edited a WIki page. It is a bit of a pain to figure out the system to do that. I think that barrier stops more people from contributing to them.
 
Joined
Nov 14, 2018
Messages
18
Likes
6
#3
I have edited a WIki page. It is a bit of a pain to figure out the system to do that. I think that barrier stops more people from contributing to them.
Yeah the system seems like a mixed blessing. I understand that it prevents blatant vandalism/misinformation on Wiki pages, but if someone doesn't agree with a change then it becomes a real headache to work out properly. It's funny though, I've heard about editorial bias on Wikipedia before and yet I never expected to see it on something like the SACD page.
 
Last edited:

Fluffy

Addicted to Fun and Learning
Joined
Sep 14, 2019
Messages
626
Likes
942
#4
Personally I used to be a little skeptical of High Resolution Audio, but now that I've taken some courses on Media Signal Processing (currently studying Digital Media in university :D), my views on the topic have changed.
Can you please say what made you change your mind? I haven't really seen convincing explanation why frequencies outside of the audible spectrum could be meaningful to music listening.
 

Hypnotoad

Active Member
Joined
Feb 17, 2019
Messages
230
Likes
226
Location
Melbourne, Australia
#5
I only ever download CD quality audio now, to my old ears it seems a waste of time to get all these so called HiRes audio files. YMMV
 
Joined
Nov 14, 2018
Messages
18
Likes
6
#6
Can you please say what made you change your mind? I haven't really seen convincing explanation why frequencies outside of the audible spectrum could be meaningful to music listening.
When we learned about how DACs work we were also taught about the effects of filtering and quantization. Suddenly I didn't just scoff at big numbers like 24bit and 96kHz because I now knew what they meant and the benefits they could provide. The biggest thing that pushed me over the edge though was learning that a greater bit depth means I can crank up my music to disgusting volumes without a song sounding 'rough' or grainy, and that was something I could easily try out for myself. Also, since I do listen to music fairly loud it was quite easy to notice the lower noise floor of my higher res content once I had quiet enough gear (Currently have a Topping D50 and Massdrop THX 789 for my Hifiman Ananda).

This being said, I don't think I've answered your question because you seem more concerned with high frequencies. Although I feel that there is a difference I am not exactly sure how to explain it. From my experience, if a specific sound naturally has a lot of high frequency content but is filtered, it will sound 'off' compared to a non-filtered counterpart. My best advice would be to listen to something hi-res with clean cymbals and try to figure out if you can tell a difference (ie. the start of "Anakin vs. Obi-Wan" from the 24bit192kHz remastered Star Wars soundtrack). But here's an actual academic paper on the matter that might help illustrate why preferring music with high frequency sounds isn't crazy talk: https://www.ncbi.nlm.nih.gov/pubmed/10848570. Essentially this paper demonstrates that while we cannot hear sounds above 20kHz, the existence of high frequencies affects our perception of the sound within our hearing range.
 
Last edited:

CDMC

Active Member
Joined
Sep 4, 2019
Messages
119
Likes
197
#7
When we learned about how DACs work we were also taught about the effects of filtering and quantization. Suddenly I didn't just scoff at big numbers like 24bit and 96kHz because I now knew what they meant and the benefits they could provide. The biggest thing that pushed me over the edge though was learning that a greater bit depth means I can crank up my music to disgusting volumes without a song sounding 'rough' or grainy, and that was something I could easily try out for myself. Also, since I do listen to music fairly loud it was quite easy to notice the lower noise floor of my higher res content once I had quiet enough gear (Currently have a Topping D50 and Massdrop THX 789 for my Hifiman Ananda).

This being said, I don't think I've answered your question because you seem more concerned with high frequencies. Although I feel that there is a difference I am not exactly sure how to explain it. From my experience, if a specific sound naturally has a lot of high frequency content but is filtered, it will sound 'off' compared to a non-filtered counterpart. My best advice would be to listen to something hi-res with clean cymbals and try to figure out if you can tell a difference (ie. the start of "Anakin vs. Obi-Wan" from the 24bit192kHz remastered Star Wars soundtrack). But here's an actual academic paper on the matter that might help illustrate why preferring music with high frequency sounds isn't crazy talk: https://www.ncbi.nlm.nih.gov/pubmed/10848570. Essentially this paper demonstrates that while we cannot hear sounds above 20kHz, the existence of high frequencies affects our perception of the sound within our hearing range.
I question that you are actually hearing a difference in hi res recordings. Let's take a look at some facts:

  1. Most listening rooms have a background noise level of 40-50 dbs. Studios, 35-40 db, the best 30db. Lets just assume you have a listening room that has a noise floor of 30db.
  2. The dynamic range of a 16 bit recording is 96 db before dithering (which makes it a bit higher). The most dynamic recording have a range of 30-40 db, most 10-20 db.
  3. So to effectively use the full resolution of a 16 bit CD, your peak listening levels would have to be 126 db, something that not even the loudest headphones or speakers can produce. The short is there is no listening environment that exceeds the dynamic range of a 16bit recording, and in reality a 14 bit recording.
If you are hearing grain and roughness when you crank your "music to disgusting levels" that has nothing to do with bit depth or sampling frequency, that is clipping your amplifier and/or overdriving your speakers/headphones.

As far as comparing hi res recordings to redbook, it is virtually impossible if you are just streaming them, as virtually all hi res recordings are mixed or remixed differently from their CD counterparts. In many cases they are nothing more than redbook recordings upsampled. The only way to to truly compare is to start with a true hi res recording, then down res it to redbook. Play the two recordings back through foobar and use the ABX add in and see how high you score in picking between the two. Every person who does this is unable to distinguish between the two.

Keep in mind there is a difference between what is needed on the listening end and the recording and mixing end. For listening, there is zero evidence that any differences can be heard beyond 16/44.1. For recording due to multiple layers and mixdowns, 96/24 is recommended to limit dither errors and loss of resolution going through multiple stages.
 
Joined
Nov 14, 2018
Messages
18
Likes
6
#8
For listening, there is zero evidence that any differences can be heard beyond 16/44.1.
Did... Did you even read the first post? Here is a link to the mentioned Meta-Analysis discussing how people can reliably tell the difference between CD and high resolution formats that among tested listeners there is a small but still statistically significant ability to discriminate between standard quality audio and high resolution audio : http://www.aes.org/e-lib/browse.cfm?elib=18296.

And here is a link to the study explaining the "hypersonic effect", which explains why people prefer music with content above 20kHz despite not being able to hear high frequency sounds: https://www.ncbi.nlm.nih.gov/pubmed/10848570.

And yes, I understand that in order to compare Hi-Res content to CD quality you would have to downsample a High Res track to 16bit, 44.1kHz.
 
Last edited:

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
25,125
Likes
52,287
Location
Seattle Area
#9
Joined
Nov 14, 2018
Messages
18
Likes
6
#10
That paper has been disputed quite a bit. I would not reference it as such.
That's fair. there are conflicting explanations on this subject. Should I instead reference subjects' responses in “The audibility of typical digital audio Filters in a high-fidelity playback system ” Convention Paper, Presented at the 137th AES Convention 2014, the paper you referenced in your post "High Resolution Audio: Does It Matter"? At least we can agree that for "some reason" there is an audible difference between high resolution and filtered content, although it is not entirely understood.
 

dc655321

Addicted to Fun and Learning
Joined
Mar 4, 2018
Messages
588
Likes
555
#11
Here is a link to the mentioned Meta-Analysis discussing how people can reliably tell the difference between CD and high resolution formats: http://www.aes.org/e-lib/browse.cfm?elib=18296.
Have you read that study carefully? Characterizing its findings as evidence of "reliably" differentiating seems like a big stretch.
Look at the uncertainties associated with the sub-study results (Figure 2 in the study).

If "hi-res" recordings were an improvement to, or even significantly different from 16/44.1, one would not have receive special training or squint with one's ears to differentiate - it would be obvious to all with healthy hearing.

And here is a link to the study explaining the "hypersonic effect", which explains why people prefer music with content above 20kHz despite not being able to hear high frequency sounds: https://www.ncbi.nlm.nih.gov/pubmed/10848570.
Amirm has beat me to the take-away message for the Oohashi paper. That one gets trotted out and beaten up so often, it should be a t-shirt or meme :)
 
Joined
Nov 14, 2018
Messages
18
Likes
6
#12
Characterizing its findings as evidence of "reliably" differentiating seems like a big stretch.
Look at the uncertainties associated with the sub-study results (Figure 2 in the study).
You are right, I should not have used the word 'reliably'. I should have said that among tested listeners there is a small but still statistically significant ability to discriminate between standard quality audio and high resolution audio. At least, according to that paper.
 

MZKM

Major Contributor
Forum Donor
Joined
Dec 1, 2018
Messages
1,183
Likes
2,077
Location
Land O’ Lakes, Florida
#14
That paper has been disputed quite a bit. I would not reference it as such.
In not a single one of the tests to prove these hypersonics are influential do they measure the FR of the speaker for both versions to check for tweeter overload. One test (I may be remembering wrong) used an MRI or something to check brain activity, they picked up brain activity differences between 44.1kHz and high-res, but picked up no activity when just playing the hypersonics, which should tell you right there that they are not influential and the differences are due to IMD in the treble.

In an online argument with someone, they referenced a study on the audibility of hypersonics, yet the study was using macaques and not humans. Some people just desperately want to believe that hypersonics (ultrasonics seems for fitting) are meaningful to us.

Not to mention that many speakers and some amps/DACs/preamps drop off ~20-30kHz.
 
Last edited:

Fluffy

Addicted to Fun and Learning
Joined
Sep 14, 2019
Messages
626
Likes
942
#15
When we learned about how DACs work we were also taught about the effects of filtering and quantization. Suddenly I didn't just scoff at big numbers like 24bit and 96kHz because I now knew what they meant and the benefits they could provide. The biggest thing that pushed me over the edge though was learning that a greater bit depth means I can crank up my music to disgusting volumes without a song sounding 'rough' or grainy, and that was something I could easily try out for myself. Also, since I do listen to music fairly loud it was quite easy to notice the lower noise floor of my higher res content once I had quiet enough gear (Currently have a Topping D50 and Massdrop THX 789 for my Hifiman Ananda).

This being said, I don't think I've answered your question because you seem more concerned with high frequencies. Although I feel that there is a difference I am not exactly sure how to explain it. From my experience, if a specific sound naturally has a lot of high frequency content but is filtered, it will sound 'off' compared to a non-filtered counterpart. My best advice would be to listen to something hi-res with clean cymbals and try to figure out if you can tell a difference (ie. the start of "Anakin vs. Obi-Wan" from the 24bit192kHz remastered Star Wars soundtrack). But here's an actual academic paper on the matter that might help illustrate why preferring music with high frequency sounds isn't crazy talk: https://www.ncbi.nlm.nih.gov/pubmed/10848570. Essentially this paper demonstrates that while we cannot hear sounds above 20kHz, the existence of high frequencies affects our perception of the sound within our hearing range.
As for bit depth, what @CDMC wrote is quite accurate. I also tested my own system to see if there is a benefit to higher bit depth than 16 – and there is, only when I'm using digital volume control. If I set my dac to 16 bit and reduce the volume by 60 db through windows, then raising up my amp to compensate, there is a little bit of added background noise, compared to if I do the same thing with the dac set to 24 bit. So for digital volume control, 24 bit is a safe bet. For audio files, there's no point.

And for the high frequencies, I also tested myself multiple times on some hi-res jazz recordings with cymbals that are guaranteed via spectral analysis to go way beyond 20khz. I did not hear any difference between the original sample rate and a down sampled to 44.1 version, neither sighted or blinded. Heck, I can't even hear straight 20khz tones (I go up to 17khz reliably).

So for me at least, hi-res is pointless. And even if a statistically significant effect can be demonstrated, it will probably still apply to a small percentage of the population, so it's shouldn't be a concern.
 

BDWoody

Major Contributor
Forum Donor
Joined
Jan 9, 2019
Messages
1,895
Likes
3,398
Location
Mid-Atlantic, USA. (Maryland)
#16
That's fair. there are conflicting explanations on this subject. Should I instead reference subjects' responses in “The audibility of typical digital audio Filters in a high-fidelity playback system ” Convention Paper, Presented at the 137th AES Convention 2014, the paper you referenced in your post "High Resolution Audio: Does It Matter"? At least we can agree that for "some reason" there is an audible difference between high resolution and filtered content, although it is not entirely understood.
In any of these studies you reference, does anyone suggest the mechanism of transfer of this high frequency information? The ear has physical limitations in terms of what it can process...
 

Eirikur

Senior Member
Joined
Jul 9, 2019
Messages
318
Likes
452
#17
If "hi-res" recordings were an improvement to, or even significantly different from 16/44.1, one would not have receive special training or squint with one's ears to differentiate - it would be obvious to all with healthy hearing.
I would choose the hires over a down-sampled version mainly because of the processing step: if you can get the "original" mixed & mastered version at 96/24 you can feed it straight to your DAC these days, i.e. cutting out the middle man.
 

leonroy

Member
Forum Donor
Joined
Feb 15, 2018
Messages
52
Likes
49
Location
London, UK
#18
So at tolerable volumes and with a young, healthy set of ears and on excellent equipment is there a simple TLDR; where we can say X bits and X kHz are sufficient for lossless distribution to end users?

From what I understand the 16bit/44.1kHz standard was partially selected due to technical limitations. In a perfect world if engineers at Sony/Philips had not had certain constraints on sampling rate or the storage size of CDs would they have picked 16/48 or 24/48 or something higher?
 

digicidal

Major Contributor
Joined
Jul 6, 2019
Messages
1,171
Likes
1,649
Location
Sin City, NV
#19
In addition to the issues with dynamic range, on the frequency side of things we also have the problem with our sensitivity - even in those whose ears are capable of the full 20kHz... the sounds will need to be at a painfully high volume in order to be detectable at all.

At the limits, audibility approaches the pain threshold so even if it is there, and is at a level you can perceive it... is it still worth listening to? Perhaps it is useful for a test, but unless the goal is hearing loss and serious fatigue after a few minutes... I don't see much value. As far as signal processing in the digital domain - that's another matter entirely.
 

Fluffy

Addicted to Fun and Learning
Joined
Sep 14, 2019
Messages
626
Likes
942
#20
From what I understand the 16bit/44.1kHz standard was partially selected due to technical limitations. In a perfect world if engineers at Sony/Philips had not had certain constraints on sampling rate or the storage size of CDs would they have picked 16/48 or 24/48 or something higher?
Talking about a world without constraints is kind of pointless. In that kind of world, we would all watch home video on our million-K TVs and listen to audio that's sampled in the terraherz. It's better to look at what is actually necessary, because once you've crossed the line of what is practically needed, you dive into the realm of diminishing returns and rapidly rising engineering complexity.

According to the Nyquist theorem, all you need for a 20khz band width is a 40khz sampling rate. To avoid the effects of imperfect filtering and aliasing, you can extend your safety range so that the sampling frequency is 44khz, and have an extra 2khz of bandwidth to perform your filtering in. The 44.1 was indeed chosen to fit broadcasting audio over tv signals (it divides nicely by both 30 fps and 25 fps), and the difference is negligible. 48khz fits better with the 24fps of films (44.1k can't be nicely divided by 24), so that's where that figure came from.
So at tolerable volumes and with a young, healthy set of ears and on excellent equipment is there a simple TLDR; where we can say X bits and X kHz are sufficient for lossless distribution to end users?
Yes. 44.1khz over 16 bit is more than enough.
 
Top Bottom