Archimago has posted results of a detailed, well-crafted study and survey about high-resolution music. He provided 140 listeners with three pairs of files (read the procedure used for the test), and asked them to determine which one was the high-resolution, 24-bit file.
The results were as expected. For two of the three files, the results were 50/50, exactly what you would get if you guessed randomly. For one of the files – the one with the greatest dynamic range, which should have benefited most from the additional resolution, 47% of people picked the correct file as being the 24-bit version, and 53% got it wrong. Again, nearly the same as random.
What’s most interesting about this test is the fact that it seems to have been very well constructed. The tester asked a number of questions to be able to analyze the listeners performing the test, including their musical experience, approximate value of their stereo equipment, and their confidence in their answers.
For example, those who claimed to be very confident or certain of their answers were still either 50/50 or more wrong than right:
Musicians who were more confident than average got even worse results:
And those with some experience of recording, mixing and editing, with high confidence, were only slightly better than random:
Most interesting were the results of people who claim to be audio hardware reviewers. They were more often wrong than right:
The article concludes:
In a naturalistic survey of 140 respondents using high quality musical samples sourced from high-resolution 24/96 digital audio collected over 2 months, there was no evidence that 24-bit audio could be appreciably differentiated from the same music dithered down to 16-bits using a basic algorithm (Adobe Audition 3, flat triangular dither, 0.5 bits).
And Archimago went on to say:
The results of this survey appear to support the notion that high bit-depth music (24-bits) does not provide audible benefits despite the fact that objectively measurable DACs capable of > 16-bit resolution are readily available at very reasonable cost these days.
This is a solid, well-crafted test that shows, yet again, that high-resolution music is a marketing ploy.
Poor audio engineers…
Poor audio engineers…
These studies miss the point that fewer people own CD Players and CD-ROMs these days. The important question is does Hi-Res sound better than 256k iTunes, and the answer is yes.
Prove it.
These studies miss the point that fewer people own CD Players and CD-ROMs these days. The important question is does Hi-Res sound better than 256k iTunes, and the answer is yes.
Prove it.
Ha. Well, I can only prove it to my own satisfaction. Maybe you won’t be impressed:
1) I am certain that on my equipment a lossless ripped CD sounds better than 256k iTunes.
2) I assume that Hi Res doesn’t sound any worse than a lossless ripped CD (I’ve never actually listened to any Hi-Res since I own both a CD Player and a CD Rom Drive).
3) Therefore, Hi-Res sounds better than 256k iTunes.
This doesn’t seem to farfetched to me. My point is that until someone starts offering CD quality downloads then the quality of CD sound is irrelevant to many people.
Lots of blind tests have shown that people can’t distinguish between a good compressed file (AAC 256 kbps or MP3 320) and CDs. You say they sound better to you; have you done a blind test?
Lots of people offer CD quality downloads in FLAC or Apple Lossless format. Just not the big vendors, yet.
Ha. Well, I can only prove it to my own satisfaction. Maybe you won’t be impressed:
1) I am certain that on my equipment a lossless ripped CD sounds better than 256k iTunes.
2) I assume that Hi Res doesn’t sound any worse than a lossless ripped CD (I’ve never actually listened to any Hi-Res since I own both a CD Player and a CD Rom Drive).
3) Therefore, Hi-Res sounds better than 256k iTunes.
This doesn’t seem to farfetched to me. My point is that until someone starts offering CD quality downloads then the quality of CD sound is irrelevant to many people.
Lots of blind tests have shown that people can’t distinguish between a good compressed file (AAC 256 kbps or MP3 320) and CDs. You say they sound better to you; have you done a blind test?
Lots of people offer CD quality downloads in FLAC or Apple Lossless format. Just not the big vendors, yet.
Are you saying that you can’t tell the difference between 256k iTunes and a CD?
I haven’t done a blind test because I’ve got a life.
If there is a one stop shop for cd quality downloads then they need to spend more on advertising.
Thanks.
No, I’ve done blind test, and I can’t reliably distinguish the two. It’s not about having a life; if you’re going to make a statement such as the one you made, and don’t care about proving it, then your statement is just an expression of faith.
Presumably the equipment one listens with may influence whether one could distinguish 256k mp3 from a flac file?
I’ve certainly tried to hear a difference between 16 bit and 24 bit files and I have to say I cannot (Cambridge Audio DACmagic > Naim Nait3 > Mission 752), though not a blind test.
Robert
Read the original article; the test asked people how much their hi-fi equipment was worth, and it did tend to skew towards a higher-priced set of hardware.
I read Archimago’s postings on the Squeezebox forums when he conducted the test. Of course Archimago’s test wasn’t mp3 vs flac but 16 bit vs 24 bit. Isn’t Charles asserting he can distinguish between 256k iTunes files and 16 or 24 bit?
Setting aside what I can or cannot hear, I still think it would be more useful to compare Hi Res w/ 256k since those are the competing downloadable formats.
I’m not sure HiRes (i.e. 24 bit) is the main competition for 256k. Most of the places I buy music from offer 16 bit Flac, or mp3 (sometimes Wav). 24 bit still seems quite uncommon to me. And I think most people are happy with mp3.
Maybe I need to look around more. I’m just a clueless CD buyer.
Maybe HiRes files will take off beyond audiophile circles if Neil Young’s Pono is a success…
Ever heard of confirmation bias?
Yep. Your point is…?
You have a life so you can’t do a blind test, but you somehow have the time to come I here and post comments. Funny how that works.
Are you saying that you can’t tell the difference between 256k iTunes and a CD?
I haven’t done a blind test because I’ve got a life.
If there is a one stop shop for cd quality downloads then they need to spend more on advertising.
Thanks.
No, I’ve done blind test, and I can’t reliably distinguish the two. It’s not about having a life; if you’re going to make a statement such as the one you made, and don’t care about proving it, then your statement is just an expression of faith.
Presumably the equipment one listens with may influence whether one could distinguish 256k mp3 from a flac file?
I’ve certainly tried to hear a difference between 16 bit and 24 bit files and I have to say I cannot (Cambridge Audio DACmagic > Naim Nait3 > Mission 752), though not a blind test.
Robert
Read the original article; the test asked people how much their hi-fi equipment was worth, and it did tend to skew towards a higher-priced set of hardware.
I read Archimago’s postings on the Squeezebox forums when he conducted the test. Of course Archimago’s test wasn’t mp3 vs flac but 16 bit vs 24 bit. Isn’t Charles asserting he can distinguish between 256k iTunes files and 16 or 24 bit?
Setting aside what I can or cannot hear, I still think it would be more useful to compare Hi Res w/ 256k since those are the competing downloadable formats.
I’m not sure HiRes (i.e. 24 bit) is the main competition for 256k. Most of the places I buy music from offer 16 bit Flac, or mp3 (sometimes Wav). 24 bit still seems quite uncommon to me. And I think most people are happy with mp3.
Maybe I need to look around more. I’m just a clueless CD buyer.
Maybe HiRes files will take off beyond audiophile circles if Neil Young’s Pono is a success…
Ever heard of confirmation bias?
Yep. Your point is…?
You have a life so you can’t do a blind test, but you somehow have the time to come I here and post comments. Funny how that works.
My brain can tell the difference between CD audio and 256 Kbps AAC, but my ears can’t.
I can tell the difference between CD audio and SACD, however.
My brain can tell the difference between CD audio and 256 Kbps AAC, but my ears can’t.
I can tell the difference between CD audio and SACD, however.
This is going to be a popular post. Those Pono-ponies really don’t want to hear this.
It is interesting and perhaps instructive to see that the one person here who disputes the results never thought to do a blind test (and apparently refuses to)
Chip, I assume you are talking about me. I just reread the article and I’m certain that I have not disputed a thing in it. Apparently I started a controversy by stating that I could tell the difference between 256k and lossless files. If it makes everyone feel better I will change that to “I believe I can tell the difference between 256k and lossless.” It’s the same to me either way.
I suppose that’s the point – ‘belief’ versus proof.
This is going to be a popular post. Those Pono-ponies really don’t want to hear this.
It is interesting and perhaps instructive to see that the one person here who disputes the results never thought to do a blind test (and apparently refuses to)
Chip, I assume you are talking about me. I just reread the article and I’m certain that I have not disputed a thing in it. Apparently I started a controversy by stating that I could tell the difference between 256k and lossless files. If it makes everyone feel better I will change that to “I believe I can tell the difference between 256k and lossless.” It’s the same to me either way.
I suppose that’s the point – ‘belief’ versus proof.
I have been doing quite a bit of testing recently with the following observations:
90% – distinguish the difference between 24/96 and 256Kbps AAC
70% – distinguish the difference between 16/44 and 256Kbps AAC
60% – distinguish the difference between 24/96 and 16/44
I started with known good 24/96 source material (Brahms, Shostakovich and Mahler) in AIFF. I then down sampled with XLD to 16/44 AIFF and 256Kbps AAC. I have observed the following over 10 runs with ABXTester:
It appears I can reliably tell the difference between 24bit and 16 bit AIFFs when compared to compressed AAC files. But when it comes to the diference between 24/96 and CD quality there’s a noticeable drop in my ability to detect a difference.
My main audio source is a SSD based Mac Mini dedicated to music playback. Where possible services like Spotlight and Time Machine are turned off. Other than a USB DAC and Ethernet, all other ports are left unused. I also use a better quality mains lead than the one supplied by Apple. I’m using a Naim DAC and Naim amplification on a separate spur/power supply.
I also observed the following:
– I scored lower using the NAD D 3020; which also has a 24 bit asynchronous DAC. However Naim equipment can be very detailed and revealing, whereas the NAD is definitely warmer. I used headphones for the comparison to eliminate the effect of speakers and rooms.
– I prepared 3 well recorded and mastered tracks. Quite a lot of 24 bit content for sale these days is nothing more that up sampled 16 bit material. There’s very little genuinely good 24 bit material out there.
– I couldn’t test this objectively; ABX doen’t support it, but I generally find I get more detail and enjoyment when using iZotope to do the sample rate conversion during playback compared to Apple’s CoreAudio. I’ve also tried BitPerfect but find I get more enjoyment from Audirvana.
– I don’t own one, but I have spent a few hours with a Chord’s Hugo DAC. It can handle DSD (SACD) files. Subjectively at least I feel there is a bigger difference when comparing CD quality PCM and 24 bit DSD files. The problem is that there is a very little DSD content.
With all of that being said, I feel that at least right now I would agree with the survey, that ‘Listeners Cannot Distinguish Between CD-Quality and High-Resolution Music Files’. The problem in my view is not so much the way we buy, encode and store our music but the quality of the source. It’s very hard to recover what wasn’t there in the first place, and there is an abundance of badly recorded and mastered music as to render the discussion about 24 bit and CD quality moot.
Ultimately we don’t listen to music because we enjoy science experiments, we listen to be entertained and moved. Some of my favourite recordings are more than 50 years old and are technically pretty average, but they don’t stop me enjoying the music. Far from it.
I have been doing quite a bit of testing recently with the following observations:
90% – distinguish the difference between 24/96 and 256Kbps AAC
70% – distinguish the difference between 16/44 and 256Kbps AAC
60% – distinguish the difference between 24/96 and 16/44
I started with known good 24/96 source material (Brahms, Shostakovich and Mahler) in AIFF. I then down sampled with XLD to 16/44 AIFF and 256Kbps AAC. I have observed the following over 10 runs with ABXTester:
It appears I can reliably tell the difference between 24bit and 16 bit AIFFs when compared to compressed AAC files. But when it comes to the diference between 24/96 and CD quality there’s a noticeable drop in my ability to detect a difference.
My main audio source is a SSD based Mac Mini dedicated to music playback. Where possible services like Spotlight and Time Machine are turned off. Other than a USB DAC and Ethernet, all other ports are left unused. I also use a better quality mains lead than the one supplied by Apple. I’m using a Naim DAC and Naim amplification on a separate spur/power supply.
I also observed the following:
– I scored lower using the NAD D 3020; which also has a 24 bit asynchronous DAC. However Naim equipment can be very detailed and revealing, whereas the NAD is definitely warmer. I used headphones for the comparison to eliminate the effect of speakers and rooms.
– I prepared 3 well recorded and mastered tracks. Quite a lot of 24 bit content for sale these days is nothing more that up sampled 16 bit material. There’s very little genuinely good 24 bit material out there.
– I couldn’t test this objectively; ABX doen’t support it, but I generally find I get more detail and enjoyment when using iZotope to do the sample rate conversion during playback compared to Apple’s CoreAudio. I’ve also tried BitPerfect but find I get more enjoyment from Audirvana.
– I don’t own one, but I have spent a few hours with a Chord’s Hugo DAC. It can handle DSD (SACD) files. Subjectively at least I feel there is a bigger difference when comparing CD quality PCM and 24 bit DSD files. The problem is that there is a very little DSD content.
With all of that being said, I feel that at least right now I would agree with the survey, that ‘Listeners Cannot Distinguish Between CD-Quality and High-Resolution Music Files’. The problem in my view is not so much the way we buy, encode and store our music but the quality of the source. It’s very hard to recover what wasn’t there in the first place, and there is an abundance of badly recorded and mastered music as to render the discussion about 24 bit and CD quality moot.
Ultimately we don’t listen to music because we enjoy science experiments, we listen to be entertained and moved. Some of my favourite recordings are more than 50 years old and are technically pretty average, but they don’t stop me enjoying the music. Far from it.
Kirk, you’ve been hitting these skeptical, audiophile-dissing posts pretty hard lately, pursuing withering dismissiveness toward vinyl, the Pono project, and high-res audio playback; bestowing categorical authority to DBX testing (no matter how informal and amateur it is); and elevating good mastering as the ultimate virtue in sound quality. My problems with this center mostly on tone and emphasis. I find the endless rounds of online debate between subjective and objective opponents surrounding lossy vs. high-rex and digital vs. analog to be generally quite arid and reductive and often extraordinarily nasty (on both sides). On that score I was struck by how friendly and non-absolutist the discussion of this test was over at Archimago, where the results are presented with a winning air of tentative modesty and open acknowledgement of the test’s limitations and possible flaws. You, on the other hand, are making sweeping and definitive claims for the test, and being sharply dismissive (“Prove it…”) when someone challenges any part of this crushing defeat for the subjectivists. I guess I’m saying I’m sorry to see you embracing this sharp-stick-in-the-eye approach. I’d love to see more positive advice and reporting on what makes for superb hi-fi sound reproduction and less ideological warfare in a Manichean, Us-vs.-Them mode. Less talk of snake oil and marketing scams and more road maps to musical beauty.
If this weren’t an industry built around scams, I would react differently. If you want musical beauty, listen to the music; don’t chase after some ideal of perfect sound that is unattainable.
So that’s it. Anything higher than 256k is a scam. Got it. Thanks.
Not at all. If that’s all you get out of this, you might want to take a bit more time reading.
High-resolution music is based on voodoo; such as frequencies that dogs can’t even hear, as though they change the way we hear music. CD quality is excellent; but hi-res is essentially about marketing.
Well, that seems somewhat non-responsive, and also seems to attribute to me a futile ideal-chasing that I don’t actually embrace. I DO want musical beauty; that’s why I care about the technology of gear and software that conveys it to me, and enjoy talking about it and debating it companionably rather than riding down the opposition with the Ride of the Valkyries turned up to 11. In any case, some great mastering engineers whose work I respect believe some of the things you find contemptibly dishonest or stupid. I’m basically agnostic and don’t like people who talk mean on either side.
Kirk, you’ve been hitting these skeptical, audiophile-dissing posts pretty hard lately, pursuing withering dismissiveness toward vinyl, the Pono project, and high-res audio playback; bestowing categorical authority to DBX testing (no matter how informal and amateur it is); and elevating good mastering as the ultimate virtue in sound quality. My problems with this center mostly on tone and emphasis. I find the endless rounds of online debate between subjective and objective opponents surrounding lossy vs. high-rex and digital vs. analog to be generally quite arid and reductive and often extraordinarily nasty (on both sides). On that score I was struck by how friendly and non-absolutist the discussion of this test was over at Archimago, where the results are presented with a winning air of tentative modesty and open acknowledgement of the test’s limitations and possible flaws. You, on the other hand, are making sweeping and definitive claims for the test, and being sharply dismissive (“Prove it…”) when someone challenges any part of this crushing defeat for the subjectivists. I guess I’m saying I’m sorry to see you embracing this sharp-stick-in-the-eye approach. I’d love to see more positive advice and reporting on what makes for superb hi-fi sound reproduction and less ideological warfare in a Manichean, Us-vs.-Them mode. Less talk of snake oil and marketing scams and more road maps to musical beauty.
If this weren’t an industry built around scams, I would react differently. If you want musical beauty, listen to the music; don’t chase after some ideal of perfect sound that is unattainable.
So that’s it. Anything higher than 256k is a scam. Got it. Thanks.
Not at all. If that’s all you get out of this, you might want to take a bit more time reading.
High-resolution music is based on voodoo; such as frequencies that dogs can’t even hear, as though they change the way we hear music. CD quality is excellent; but hi-res is essentially about marketing.
Well, that seems somewhat non-responsive, and also seems to attribute to me a futile ideal-chasing that I don’t actually embrace. I DO want musical beauty; that’s why I care about the technology of gear and software that conveys it to me, and enjoy talking about it and debating it companionably rather than riding down the opposition with the Ride of the Valkyries turned up to 11. In any case, some great mastering engineers whose work I respect believe some of the things you find contemptibly dishonest or stupid. I’m basically agnostic and don’t like people who talk mean on either side.
I have always found the sound quality of SACD to be obviously superior to CD (I’m talking about classical music). However, it has been suggested on this site, and others, that this perceived difference stems from the superior recording master used for the SACD. I recently purchased a “Mastered for iTunes” recording of Brahms German Requiem from the iTunes store, where it was easy to compare this to the “normal” of version the same recording sold as part of an Otto Klemperer collection. The difference was instantly noticeable — the MFiT version had more bass, wider dynamic range and greater “presence”.
Apple states that MFiT releases employ high-resolution 26/96 masters. My take on this is that High Resolution technology has great benefit mainly in the recording/mastering process, but the value of delivering hi-res files directly to the consumer is uncertain (hence the controversy).
Looked at this way, both camps can be right: Apples hi-res MFiT process results in superior sounding masters that are beyond CD quality (i.e. CDs using conventional mastering techniques), and this results in 256k files that sound (at least) as good as CDs.
Among those who don’t believe that thousand-dollar speaker cables will change their lives, the main suggestion for improving music quality is the mastering process. There are plenty of examples of good (and bad) mastering that show how important that part of the process is. High-res files of poorly mastered albums will still sound bad, and many people who do buy high-res files are starting to realize this.
But audiophiles are easily swayed. If they’re used to paying exorbitant prices for gadgets, they don’t flinch at paying more for music, because it’s high-resolution. (Even though many cases have been found where so-called 24-bit files have just been up sampled from 16-bit.)
Yes, MfiT is based on 24/96 masters, and the files are directly compressed to AAC 256 kbps from those, skipping one step (that of 24/96 > CD). So there is a possibility that they may sound better. But in discussions with some people in the classical music industry, it’s not that simple; some people don’t think it makes much of a difference.
As you say, hi-res files are essential for the recording/editing/mastering process, but only because changes made to these files results in – to simplify – fewer rounding errors than with lower resolution files. People don’t record in 24/96 because of the dog-only frequencies; they record in those specs because when they alter the files, there are fewer artifacts.
I have always found the sound quality of SACD to be obviously superior to CD (I’m talking about classical music). However, it has been suggested on this site, and others, that this perceived difference stems from the superior recording master used for the SACD. I recently purchased a “Mastered for iTunes” recording of Brahms German Requiem from the iTunes store, where it was easy to compare this to the “normal” of version the same recording sold as part of an Otto Klemperer collection. The difference was instantly noticeable — the MFiT version had more bass, wider dynamic range and greater “presence”.
Apple states that MFiT releases employ high-resolution 26/96 masters. My take on this is that High Resolution technology has great benefit mainly in the recording/mastering process, but the value of delivering hi-res files directly to the consumer is uncertain (hence the controversy).
Looked at this way, both camps can be right: Apples hi-res MFiT process results in superior sounding masters that are beyond CD quality (i.e. CDs using conventional mastering techniques), and this results in 256k files that sound (at least) as good as CDs.
Among those who don’t believe that thousand-dollar speaker cables will change their lives, the main suggestion for improving music quality is the mastering process. There are plenty of examples of good (and bad) mastering that show how important that part of the process is. High-res files of poorly mastered albums will still sound bad, and many people who do buy high-res files are starting to realize this.
But audiophiles are easily swayed. If they’re used to paying exorbitant prices for gadgets, they don’t flinch at paying more for music, because it’s high-resolution. (Even though many cases have been found where so-called 24-bit files have just been up sampled from 16-bit.)
Yes, MfiT is based on 24/96 masters, and the files are directly compressed to AAC 256 kbps from those, skipping one step (that of 24/96 > CD). So there is a possibility that they may sound better. But in discussions with some people in the classical music industry, it’s not that simple; some people don’t think it makes much of a difference.
As you say, hi-res files are essential for the recording/editing/mastering process, but only because changes made to these files results in – to simplify – fewer rounding errors than with lower resolution files. People don’t record in 24/96 because of the dog-only frequencies; they record in those specs because when they alter the files, there are fewer artifacts.
The Guardian offers their take on the subject here:
http://www.theguardian.com/technology/2014/aug/21/mp3-cd-24-bit-audio-music-hi-res
The article is dated August 21st.
The Guardian offers their take on the subject here:
http://www.theguardian.com/technology/2014/aug/21/mp3-cd-24-bit-audio-music-hi-res
The article is dated August 21st.
Interesting. And the music used was classical, which should reveal more problems, at least with dynamic range and high-end stridency, than pop. Strings in particular suffer under poor MP3 conversion.
Another area worth testing, though, is low-end. Dance music with serious low-end can be compromised in MP3 conversion. MP3s sometimes have that wobbly weak low-end compared the lossless versions. It would be most noticeable though on a good sound system (like a good clean club system) that can fully reproduce those frequencies. 320k without any bass-filtering might still be indistinguishable from lossless, though…
I have to do some double-blind tests on myself in the studio and see whether I can tell the difference between highest quality MP3 and lossless.
Interesting. And the music used was classical, which should reveal more problems, at least with dynamic range and high-end stridency, than pop. Strings in particular suffer under poor MP3 conversion.
Another area worth testing, though, is low-end. Dance music with serious low-end can be compromised in MP3 conversion. MP3s sometimes have that wobbly weak low-end compared the lossless versions. It would be most noticeable though on a good sound system (like a good clean club system) that can fully reproduce those frequencies. 320k without any bass-filtering might still be indistinguishable from lossless, though…
I have to do some double-blind tests on myself in the studio and see whether I can tell the difference between highest quality MP3 and lossless.
You can’t just SAY ‘well crafted study’ when it’s completely uncontrolled and weights single guesses of random internet clowns the same as trained listeners conducting statistically significant ABX sessions.
It’s one thing to handwave away the fact that extended non-chance results over large sessions illustrate the statistical presence of a FUGITIVE perception that’s sometimes significant and sometimes not. Sensory input is fuzzy, it’s always a gray area. We can determine how often you will notice what’s ‘behind the veil’ and that determines how much the failings of lossy/low-bit audio will bug you, and it’s not the same as ‘infallibly always noticing the problem’.
That’s a really lame, low bar to set, this idea of ‘if you don’t notice the unpleasant artifact every single second, go and live with it forever because you deserve no better’.
Or it was, until we invented ‘ask random people on the internet to self-report and call that a controlled study that means anything’. Did you even have a mechanism for screening out fake answers?
Bits are cheap. It is extremely dumb to crusade on behalf of lossy compressing and artificially limited word lengths in a world where the sky’s the limit. And it’s extremely suspect to try and emotionally shame people away from the terrible risk that they might listen for themselves and have an emotional reaction in favor of higher quality.
Nobody would ever argue that these deeper word lengths and higher sample rates are not measurably higher quality by any measurement standard. This is always about the proposition that humans should be ASHAMED to attempt to consume more resolution than they can process. For goodness sake, why?
Might as well go bathe in a large Ziploc bag, then go sailing in a kiddie pool. Expansiveness is free. Or, technically, now expansiveness is about $400 and a set of headphones, and my, aren’t people outraged.
This study is one of many. Audiophiles are constantly saying that the studies are no good, because they don’t agree with what they want them to show.
Bits are not cheap. I’m all for lossless formats, but high-resolution music generally costs twice that of CD-quality or compressed downloads. It’s all about deception, and trying to convince people that it makes a difference, when it simply doesn’t.
Would you pay, say, twice as much to have a film in 8K resolution, just because you can? The day will come when the TV/movie industry tries to pass that off on us (they’re struggling to get people to accept 4K), and, unless we have extremely large displays, it simply doesn’t make a difference because the eye can’t see that resolution. It’s similar with music.
You can’t just SAY ‘well crafted study’ when it’s completely uncontrolled and weights single guesses of random internet clowns the same as trained listeners conducting statistically significant ABX sessions.
It’s one thing to handwave away the fact that extended non-chance results over large sessions illustrate the statistical presence of a FUGITIVE perception that’s sometimes significant and sometimes not. Sensory input is fuzzy, it’s always a gray area. We can determine how often you will notice what’s ‘behind the veil’ and that determines how much the failings of lossy/low-bit audio will bug you, and it’s not the same as ‘infallibly always noticing the problem’.
That’s a really lame, low bar to set, this idea of ‘if you don’t notice the unpleasant artifact every single second, go and live with it forever because you deserve no better’.
Or it was, until we invented ‘ask random people on the internet to self-report and call that a controlled study that means anything’. Did you even have a mechanism for screening out fake answers?
Bits are cheap. It is extremely dumb to crusade on behalf of lossy compressing and artificially limited word lengths in a world where the sky’s the limit. And it’s extremely suspect to try and emotionally shame people away from the terrible risk that they might listen for themselves and have an emotional reaction in favor of higher quality.
Nobody would ever argue that these deeper word lengths and higher sample rates are not measurably higher quality by any measurement standard. This is always about the proposition that humans should be ASHAMED to attempt to consume more resolution than they can process. For goodness sake, why?
Might as well go bathe in a large Ziploc bag, then go sailing in a kiddie pool. Expansiveness is free. Or, technically, now expansiveness is about $400 and a set of headphones, and my, aren’t people outraged.
This study is one of many. Audiophiles are constantly saying that the studies are no good, because they don’t agree with what they want them to show.
Bits are not cheap. I’m all for lossless formats, but high-resolution music generally costs twice that of CD-quality or compressed downloads. It’s all about deception, and trying to convince people that it makes a difference, when it simply doesn’t.
Would you pay, say, twice as much to have a film in 8K resolution, just because you can? The day will come when the TV/movie industry tries to pass that off on us (they’re struggling to get people to accept 4K), and, unless we have extremely large displays, it simply doesn’t make a difference because the eye can’t see that resolution. It’s similar with music.
There is always a question of what is good enough or what we can audibly discern. There are certain things that we know to be true without a doubt.
One is that without equipment accurate enough to maintain a pure signal path, the arguments is diminished.
Two, without speakers that are able to accurately reproduce the signal from the equipment, the argument is diminished.
Three, compression formats are designed to “trick” the listener. In fact compression will often times find subtleties in music (a background bongo strike, a faint high piano note) and it will process those frequencies at those levels to be more pronounced or it will add some processing in order to ensure those subtleties are maintained. This often adds noise or reduces the overall soundstage of the score. At first our ears like the increased definition but over time they will fatigue because the brain will slowly process them as sounding less natural. In fact the same things happen in Video or in the Television market. Earlier HD TV’s would not be able to show all of the color gradations accurately, so they would “process” the video to showcase the contrast. Meaning one pixel color would be clearly defined from the color of the adjacent pixel. Our eyes at first like this because our brains enjoy being able to clearly discern objects from one another – but in reality a man’s suit isn’t all jet black like on some televisions. It is mix of millions of small gradations of grey, and dark grey and medium grey and light blacks etc. If you watch TV long enough on a cheap set your eyes will grow weary of the high contrast and over time you will long for the more natural looking display of colors that your eyes normally see in the real world.
Four, don’t confuse hi-resolution with accurate reproduction. Fats Domino probably had access to a two track analog recording setup and when those tracks were full they bumped them all to one track and recorded another instrument or vocal on to the free track. Analog sounds good but it degrades quickly when copied over and over like back in the old days – so some music no matter how much remastering is done will never be hi-resolution. Its not about high resolution audio – its about listening to as much of the original recording as possible.
Five, different file types can be analyzed on a scope and looked at for noise, accuracy etc. There is empirical evidence as to which formats are technically better or worse. There is argument about what our ears can discern, but if you only want to listen to single drum bang the same note over and over then the quality of the music is probably irrelevant. I would argue that some musical genres lend themselves to compression more than others. Rap has a lot of offering that are limited in terms of musical variation. If I wanted to enjoy a genre such as Rap that relied heavily on low frequency back beats and vocals, I would worry more about my equipment and speakers than I would the file format (within reason). For instance, I would have a crossover to the sub at about 100-hz and I would bi-amp my speakers so that I could get as much snap out of the base as possible. Maybe center channel speaker for vocals, implementing a 3 channel setup vs stereo. I would think that a good compression format sampled at 256-khz or 320-khz would sound fine – and the world would be none the wiser.
Six, familiarity is everything. So as in point number two, the fact that 50 percent of the people took a test and failed to accurately identify the higher resolution format only means that compression schemes are smart and are accenting certain things our ears like. But on exceptional equipment with exceptional speakers you will over time start to discern what is really musical and what is not. You will start to notice subtleties in soundstage and in the responsiveness of the woofers. Certain songs will bring this out more than others. You will become familiar with the room and what you acoustics you are used to. Great equipment and great speakers will also expose crappy recordings, they will also expose crappy compression schemes. You will start to notice noise and unnatural attempts to make something sound a certain way. When people hear high quality tweeters made of diamond or beryllium they often comment that the highs are too grating or aggressive – this is often a result of listening to music that is compressed and has a good deal of noise in the higher frequencies (a well known side effect of lossy compression formats)
Seven, and most important. Its all about what makes you happy. Some folks like gear and some folks like music. Don’t let your love for one diminish the other. If you enjoy music and you want find the best possible reproductions of that music and you are wiling to invest in the gear to resolve those subtleties, then by all means go for it. If you just want the amp with the most warmth or the most headroom or the lowest signal to noise ration then by all means do that, but either way your room and your equipment will have its own personality and you will ultimately have to follow where that leads.
In closing – in my humble opinion – hard drive space is cheap these days. If you can get the majority of your music saved to AIFF format at CD Quality of 44-khz then you will have the best chance of enjoying your music as it was intended. The cost to store a reasonably sized catalog is really peanuts. You can store about a 1000 albums on a half Terabyte disk or 500mb. That is really cheap. Once you have that you never need worry about mathematical errors of uncompressing lossless files such as FLAC or ALAC on the fly. You can still save tagging and metadata in the AIFF file. IF and I say IF you want to get into the world if true high resolution audio. I wouldn’t worry so much about going backwards, because chances are the original recording wasn’t sampled at a high enough bit rate to care. I would look to the future, do research into the recording and find music that was mastered originally with the intent of being distributed in high resolution. Then and only then will your ears have a chance of finding some Shangri-La. But even then, the value in it is only what you perceive it to be. The scientists say that many of these frequencies are out of our range of perception, the enthusiasts and believers say that when it is all put together in a package that the sum is greater than its parts and there is an unmistakable fullness and perfectness that emerges… Its for you to decide.
There is always a question of what is good enough or what we can audibly discern. There are certain things that we know to be true without a doubt.
One is that without equipment accurate enough to maintain a pure signal path, the arguments is diminished.
Two, without speakers that are able to accurately reproduce the signal from the equipment, the argument is diminished.
Three, compression formats are designed to “trick” the listener. In fact compression will often times find subtleties in music (a background bongo strike, a faint high piano note) and it will process those frequencies at those levels to be more pronounced or it will add some processing in order to ensure those subtleties are maintained. This often adds noise or reduces the overall soundstage of the score. At first our ears like the increased definition but over time they will fatigue because the brain will slowly process them as sounding less natural. In fact the same things happen in Video or in the Television market. Earlier HD TV’s would not be able to show all of the color gradations accurately, so they would “process” the video to showcase the contrast. Meaning one pixel color would be clearly defined from the color of the adjacent pixel. Our eyes at first like this because our brains enjoy being able to clearly discern objects from one another – but in reality a man’s suit isn’t all jet black like on some televisions. It is mix of millions of small gradations of grey, and dark grey and medium grey and light blacks etc. If you watch TV long enough on a cheap set your eyes will grow weary of the high contrast and over time you will long for the more natural looking display of colors that your eyes normally see in the real world.
Four, don’t confuse hi-resolution with accurate reproduction. Fats Domino probably had access to a two track analog recording setup and when those tracks were full they bumped them all to one track and recorded another instrument or vocal on to the free track. Analog sounds good but it degrades quickly when copied over and over like back in the old days – so some music no matter how much remastering is done will never be hi-resolution. Its not about high resolution audio – its about listening to as much of the original recording as possible.
Five, different file types can be analyzed on a scope and looked at for noise, accuracy etc. There is empirical evidence as to which formats are technically better or worse. There is argument about what our ears can discern, but if you only want to listen to single drum bang the same note over and over then the quality of the music is probably irrelevant. I would argue that some musical genres lend themselves to compression more than others. Rap has a lot of offering that are limited in terms of musical variation. If I wanted to enjoy a genre such as Rap that relied heavily on low frequency back beats and vocals, I would worry more about my equipment and speakers than I would the file format (within reason). For instance, I would have a crossover to the sub at about 100-hz and I would bi-amp my speakers so that I could get as much snap out of the base as possible. Maybe center channel speaker for vocals, implementing a 3 channel setup vs stereo. I would think that a good compression format sampled at 256-khz or 320-khz would sound fine – and the world would be none the wiser.
Six, familiarity is everything. So as in point number two, the fact that 50 percent of the people took a test and failed to accurately identify the higher resolution format only means that compression schemes are smart and are accenting certain things our ears like. But on exceptional equipment with exceptional speakers you will over time start to discern what is really musical and what is not. You will start to notice subtleties in soundstage and in the responsiveness of the woofers. Certain songs will bring this out more than others. You will become familiar with the room and what you acoustics you are used to. Great equipment and great speakers will also expose crappy recordings, they will also expose crappy compression schemes. You will start to notice noise and unnatural attempts to make something sound a certain way. When people hear high quality tweeters made of diamond or beryllium they often comment that the highs are too grating or aggressive – this is often a result of listening to music that is compressed and has a good deal of noise in the higher frequencies (a well known side effect of lossy compression formats)
Seven, and most important. Its all about what makes you happy. Some folks like gear and some folks like music. Don’t let your love for one diminish the other. If you enjoy music and you want find the best possible reproductions of that music and you are wiling to invest in the gear to resolve those subtleties, then by all means go for it. If you just want the amp with the most warmth or the most headroom or the lowest signal to noise ration then by all means do that, but either way your room and your equipment will have its own personality and you will ultimately have to follow where that leads.
In closing – in my humble opinion – hard drive space is cheap these days. If you can get the majority of your music saved to AIFF format at CD Quality of 44-khz then you will have the best chance of enjoying your music as it was intended. The cost to store a reasonably sized catalog is really peanuts. You can store about a 1000 albums on a half Terabyte disk or 500mb. That is really cheap. Once you have that you never need worry about mathematical errors of uncompressing lossless files such as FLAC or ALAC on the fly. You can still save tagging and metadata in the AIFF file. IF and I say IF you want to get into the world if true high resolution audio. I wouldn’t worry so much about going backwards, because chances are the original recording wasn’t sampled at a high enough bit rate to care. I would look to the future, do research into the recording and find music that was mastered originally with the intent of being distributed in high resolution. Then and only then will your ears have a chance of finding some Shangri-La. But even then, the value in it is only what you perceive it to be. The scientists say that many of these frequencies are out of our range of perception, the enthusiasts and believers say that when it is all put together in a package that the sum is greater than its parts and there is an unmistakable fullness and perfectness that emerges… Its for you to decide.
I use 24 bit in audio production all the time. It’s ‘easy’ to tell the difference between 24 bit and 16 bit files — but generally ONLY if you turn the volume way up on a very quiet passage, a fade-out, a reverb-tail, or the like.
And if you were to leave the playback volume at the level required to hear that difference during the loud section of properly mastered program material, they would be peeling you off the back wall of your listening room.
I use 24 bit in audio production all the time. It’s ‘easy’ to tell the difference between 24 bit and 16 bit files — but generally ONLY if you turn the volume way up on a very quiet passage, a fade-out, a reverb-tail, or the like.
And if you were to leave the playback volume at the level required to hear that difference during the loud section of properly mastered program material, they would be peeling you off the back wall of your listening room.
Hey folks. Sorry to go off on a tangent. But the post has been most useful. I was on the verge of investing in Hi Res – but for now have decided to go back to my large vinyl collection and wait till theres a leap enough in res to clearly make a difference. So its CDs for chilling and LPs for Music. And I can tell the difference between the two! Cheers from Mumbai.
Hey folks. Sorry to go off on a tangent. But the post has been most useful. I was on the verge of investing in Hi Res – but for now have decided to go back to my large vinyl collection and wait till theres a leap enough in res to clearly make a difference. So its CDs for chilling and LPs for Music. And I can tell the difference between the two! Cheers from Mumbai.