Slashdot Mirror
Why Distributing Music As 24-bit/192kHz Downloads Is Pointless
An anonymous reader writes "A recent post at Xiph.org provides a long and incredibly detailed explanation of why 24-bit/192kHz music downloads — touted as being of 'uncompromised studio quality' — don't make any sense. The post walks us through some of the basics of ear anatomy, sampling rates, and listening tests, finally concluding that lossless formats and a decent pair of headphones will do a lot more for your audio enjoyment than 24/192 recordings. 'Why push back against 24/192? Because it's a solution to a problem that doesn't exist, a business model based on willful ignorance and scamming people. The more that pseudoscience goes unchecked in the world at large, the harder it is for truth to overcome truthiness... even if this is a small and relatively insignificant example.'"
I know, Stephen Colbert is Reddit's hero and they're starting to infiltrate this site as well, but seriously. Call them lies. That's what they are, that's what they -deserve- to be called. Are people really that passive-aggressive and afraid of expressing themselves that they won't call someone who lies a liar any more?
lossless formats and a decent pair of headphones and a set of really expensive MONSTER CABLES will do a lot more for your audio enjoyment than 24/192 recordings.
There, ftfy.
I find your well-reasoned and respectfully written response to be full of helpful counterpoints and useful references. I wish to subscribe to your newsletter.
>There is a huge problem with file sizes
Not any more, pumpkin.
We hit the terabyte size in drives a couple of years ago. There's no reason to be buying this format vs "archive quality" cd-audio or other lossless.
Buy/rip lossless. Transcode to lossy as needed. Anything else and you're being ripped off.
I listen to real music with real instruments. The "swish" you get in high-frequency percussion with lossy algorithms is annoying as fuck.
--
BMO
You are missing the point of the article. 192KHz is not 192kbps.
Double blind test results or I will continue to believe that you are suffering from Illusory superiority.
-1 overrated isn't the same thing as "I disagree".
I have a PhD in Digital Music Conservation from the University of Florida. I have to stress that the phenomenon known as "digital dust" is the real problem regarding conservation of music, and any other type of digital file. Digital files are stored in digital filing cabinets called "directories" which are prone to "digital dust" - slight bit alterations that happen now or then. Now, admittedly, in its ideal, pristine condition, a piece of musical work encoded in FLAC format contains more information than the same piece encoded in MP3, however, as the FLAC file is bigger, it accumulates, in fact, MORE digital dust than the MP3 file. Now you might say that the density of dust is the same. That would be a naive view. Since MP3 files are smaller, they can be much more easily stacked together and held in "drawers" called archive files (Zip, Rar, Lha, etc.) ; in such a configuration, their surface-to-volume ratio is minimized. Thus, they accumulate LESS digital dust and thus decay at a much slower rate than FLACs. All this is well-known in academia, alas the ignorant hordes just think that because it's bigger, it must be better.
So over the past months there's been some discussion about the merits of lossy compression and the rotational velocidensity issue. I'm an audiophile myself and posses a vast collection of uncompressed audio files, but I do want to assure the casual low-bitrate users that their music library is quite safe.
Being an audio engineer for over 21 years, I'm going to let you in on a little secret. While rotational velocidensity is indeed responsible for some deterioration of an unanchored file, there's a simple way of preventing this. Better still, there have been some reported cases of damaged files repairing themselves, although marginally so (about 1.7 percent for the .ogg format).
The procedure is, although effective, rather unorthodox. Rotational velocidensity, as known only affects compressed files, i.e. files who's anchoring has been damaged during compression procedures. Simply mounting your hard disk upside down enables centripetal forces to cancel out the rotational ruptures in the disk. As I said, unorthodox, and mainstream manufactures will not approve as it hurts sales (less rotational velocidensity damage means a slighter chance of disk failure.)
I'd still go with uncompressed .wav myself, but there's nothing wrong with compressed formats like flac or mp3 when you treat your hardware right
--
BMO
As a former audio engineer with some ranking success, I can tell you that it's true -- delivering high-sample rate audio as an end format is really pointless. It hardly makes sense in a studio, and definitely is illogical for the distribution of a final mix.
However, there is an increase in quality using 24 bit. Most people just assume increasing the bit depth is the same as increasing the sample rate, but this is incorrect and short-sided. With higher bit depths, you can get your analog components operating a little further away from the noise floor. This also makes dithering much less noticeable (the noise you hear when you crank the volume up as a song fades out). Why? There are more "levels" for each sample to be recorded into. It's like going from 16 to 24 bit color. You would notice this.
For the 192 KHz fans out there, there is direct and proven mathematical reasoning for why 44 KHz audio is plenty. That, and your equipment probably can't produce it. Your converters probably suck at this frequency, and your ears definitely can't vibrate that quickly. More samples doesn't "smooth out" the waveform.
When you can tell the difference between 44.1/16 and 192/24 in a double blind trial, come back and we'll talk.
Subjective opinions about audio quality, particularly those accompanied by words like "deaf" or "idiot", are worse than useless. Subjective listening is deeply suggestible and unreliable. Claimed differences among any acceptably well designed audio electronics virtually always disappears under rigorous and controlled testing.
To give just one example, listeners reliably prefer the louder source in subjective testing, even if the difference is not consciously perceptible. If a 192/24 D/A is just 0.1db louder than a 44.1/16 source, listeners will tend to describe it in all sorts of subjective terms... "edgier," "richer," "more forward," "cleaner impact," "deeper soundstage" etc when in fact it is simply a little louder.
"Because it's a solution to a problem that doesn't exist, a business model based on willful ignorance and scamming people."
which happens to be a business model that works, unfortunately
intellectual property law is philosophically incoherent. it is your moral duty to ignore it or sabotage it
Your cat is not "listening", it is simply tolerating that annoying racket that you call "music" in exchange for food, body heat, clean kitty litter, etc.
Did you listen to it double blinded? No? Then I don't care what your confirmation bias tells you that you heard. The difference is beyond your ability to hear, but not beyond your ability to deceive yourself into believing what you want to believe.
-1 overrated isn't the same thing as "I disagree".
A group of sixty audio professionals and audiophiles did a series of controlled double blind trials published in the Journal of the Audio Engineering Society. They found no perceptible degradation caused by a 16-bit/44.1kHz A/D/A.
http://www.aes.org/e-lib/browse.cfm?elib=14195
No loss from the original sampling, i.e. they didn't loose any information in the compression. Most music is sampled at (correct me if I'm wrong someone?) 44kHz, I forget how many bits, I think 16. The thing being touted is sampling it at 192kHz with 24bit resolution, which is much higher on both counts, and therefore, in theory, should produce better quality reproduction of the sound based on oversampling and reduction of the signal to quantization noise rate. The point the TFA makes is that human ears can't hear the difference, although I think that some audiophiles may beg to differ.
FWIW, I have quite bad ears, a recording needs to be quite bad before I notice it. I'm an electronic engineer though, so I know all the theory...
One thing I know, and that is that I am ignorant...
You mean like, honkies, spics, niggers, dune coons, prairie niggers, kykes, faggots, chinks, canucks, wops, guineas, krauts, and polocks? I think that's everybody anyway, my apologies if I left out any group, I try to be an equal opportunity offender, challenging people to be adults and get over their group identitied. Criticism welcome. Cowardly disapproval spurned.
I know, Stephen Colbert is Reddit's hero and they're starting to infiltrate this site as well, but seriously. Call them lies. That's what they are, that's what they -deserve- to be called. Are people really that passive-aggressive and afraid of expressing themselves that they won't call someone who lies a liar any more?
Okay, everybody, listen up: Anonymous Coward is having a rough day so let's all be extra nice to him!
"I like to lick butts!" by MobileTatsu-NJG (#32700246) (Score:5, Informative)
For the rest of us on /. haven't we had all of our music in FLAC for a decade now? I don't even listen to music much and mine is.
My music is mostly stored in whatever the default is for YouTube videos that I've saved locally. I'm apparently even less of a music fan than you are.
Fun fact: I'm also an audio technician. Yes, I can hear the occasional damaged sound, but I'm not enough of an asshole to care.
You do not have a moral or legal right to do absolutely anything you want.
Truthiness refers to a specific kind of lie-- a lie that sounds true, and that a large segment of people really want to be true. The kind of thing that's close enough to true for AM radio talk show hosts.
And now... I'll get off your damned lawn. Don't forget to take your teeth out before falling asleep.
That doesn't make sense. 48k and 96K are sampling rates, so the problem wouldn't be in encoding and decoding. If there was a quality problem, it would be analog to digital converters those transferring to digital formats are using and the digital to analog converers a sound system has. You seem to be conflating sampling rate and bitrate. There have been dramatic improvements for the same bitrates in the last 20 years.
This is my signature. There are many like it, but this one is mine.
Not if you don't know any better. ;-)
Seriously, its been so long since I've seen a live band I don't know what a drum is supposed to sound like.
At my age my ears are not so hot.
Sig Battery depleted. Reverting to safe mode.
Indeed. One of the overlooked but highly important issues with sampling rates is that although you can represent up to Nyquist in a periodically sampled signal, that is the limit for infinite length recordings. For finite-length recordings, it isn't all or nothing, represented perfectly or not at all -- instead the uncertainty (read: representation error) increases as you approach Nyquist.
Too bad Shannon and Nyquist are dead. It seems they've completely misunderstood the math. How embarrassing they passed on before you could correct their mistake. Now they'll never know.
I used to think like you. Spent thousands on audio equipment.
Now that I'm deaf in one ear I listen to MP3s through $24 headphones.
Being deaf saves a lot of money.
This space available.
Double blind test or gtfo. The peer reviewed research says you can't hear it. Talk is cheap, show us some data.
-1 overrated isn't the same thing as "I disagree".
There is a huge problem with file sizes (so both hard drive space and download bandwidth) with lossless files, so no, it's not entirely without problems.
I own (legally, even) somewhere on the order of 2500 CDs.
I have ripped all of them to FLAC (lossless).
Total size, under 600GB. I could easily fit my entire collection on a single HDD five years ago. Today, they don't even count as the biggest single directory on my home file server (hell, not even third place - Though in fairness, I do collect historically-significant Linux distro ISOs).
FWIW, even ripped raw rather than compressed as FLAC, they would still fit on a single 2TB drive. Audio really doesn't present all that much of a problem these days.
That's a profound misrepresentation of how hearing works.
Here's an oversimplified and inaccurate explanation. The ear's mechanism relies on different frequencies providing the highest level of excitation at different places. Your trained nervous system recognizes each different place as a different tone.
For most people, there is no place where sounds above 20 kHz will irritate a nerve ending enough to send an impulse to your brain. Thus, no sound higher than 20 kHz is audible, and 20 kHz corresponds to a 40 kHz sampling rate. (One sample at the low point on the wave, the next sample at the next high point, etc.
Contribute to civilization: ari.aynrand.org/donate
I recently remixed a classic recording for sony records. The files where rolled off of tape at 24bit/96k. 48k I can understand but 96k is pointless. WAAAAAAY beyond the range of human hearing. In the old days, things like cymbals and brass could really stick out because the encoders and decoders where just not where they are today.
Anyone that tells you they can hear the difference between 48k and 96k is dreaming. Its the quality of the recording that counts more than anything these days.
The difference is that the antialiasing filters are much simpler and have a gentler roll-off when sampling at 96kHz. The high-order filters necessary to ensure adequate attenuation at Nyquist and above when sampling at the lower rates have this tendency to ring.
They won't believe you. They believe their ears must be superior to those pseudo-audiophiles. Your post should have ended all discussion, but *sigh* it won't.
Fair point. The people who go on about 24/192 probably don't really listen to the kind of music which is affected by the loudness war. Most audiophiles I know are heavily into jazz or classical music, the recordings of those usually try to be quite faithful to the original.
One thing I know, and that is that I am ignorant...
1. Find post asking for results of a properly conducted double blind test.
2. Ramble on about your various stereo equipment for a couple paragraphs, show a complete ignorance of confirmation bias.
3. Completely fail to provide the requested evidence, wasting every ones time.
4. ???
5. Profit!
-1 overrated isn't the same thing as "I disagree".
I'm not deaf, but I've never spent more than $10 on headphones.
You'll be in for one heck of a shock the day you hear what music actually sounds like.
no it isn't. verisimilitude is, roughly, the quality of being believably realistic. truthiness is like "verisimilitudinous lying," i.e. the apparent realism is misleading, often toward the exact opposite of the truth.
"They were pure niggers." – Noam Chomsky
There may be no theoretical benefit, but since there's no such thing as an ideal sampler or filter or quantiser, it has many practical benefits.
Here is a quick example. You sample at 44 kHz. The first Nyquist zone is from 0 to 22 kHz, the second one is from 22 to 44 kHz (with flipped spectrum.)
Now, say that some [mechanical] harmonic from some instrument has frequency of 33 kHz. We don't hear those with our ears (parts of the ear are too massive to vibrate fast enough) so no harm done. The orchestra is playing as usual.
But now record this orchestra with an imperfect antialiasing filter (there are reasons why a perfect one wouldn't do you much good anyway.) The 33 kHz harmonic falls into the 2nd Nyquist zone. It will be played back as if it was (22 kHz - 11 kHz = 11 kHz.) Can you hear 11 kHz? Most people hear it just fine. Think about it for a moment. There was no 11 kHz signal in the original spectrum; there was 33 kHz, an inaudible one. The artifact showed up because a [lossy] mathematical operation was performed on the data that describes the signal. The resulting distortion produced an audible tone where none was present originally.
However if you encode at, say, 128 kHz sampling rate, things change. First, the antialiasing filter - even if it is of the same architecture - will have its cutoff way below the Fs/2. This means that signals of the second Nyquist zone will be attenuated by many tens of dB - essentially they can be completely eliminated because nobody cares what you do to ripple and phase above 30 or 40 kHz. Second, for the alias to show up it has to be in LF radio band now, starting at 128 kHz. Microphones aren't even mechanically capable of picking up those frequencies. And finally, if that 33 kHz harmonic passes through the filter (with the same mediocre attenuation as in the first example) ... it will be played back as 33 kHz, and it won't go anywhere. The amplifier will filter it, and the speakers will attenuate it greatly. In other words, a serious distortion that was present when you are sampling at 44 kHz disappears when you are sampling at a much higher rate.
My last hearing test has shown that I can hear up to 21khz. I play Tin Whistle, Great Highland Bagpipe, Ceilidh Pipe, and Guitar. I have heard the rattle of a live sax. I have heard a delicate triangle ringing out over a live orchestra. I have heard live trumpet. I've spent quite a bit of time training my ears to hear those sounds.
.wav .wav .wav and FLAC, encoded with the FLAC reference encoder
I have consistently failed to find a difference between the following in ABX tests I have run:
192/24 and 44/16
96/24 and 44/16
44/16
My reference tracks have been Pink Floyd's "Time", Sirenia's "Meridian", Bach's "Herz und Mund und Tat und Leben" part 7 conducted by Nikolaus Harnoncourt.
The reference system was a PC with an Asus Xonar Essence sound card, a Rogue audio Perseus pre-amp, a pair of Rogue M-180 monoblock power amps, and Vandersteen Signature 2ce speakers. (My father's sound system and my PC).
Of course, msobkow will claim that since I like Highland Bagpipes my hearing is inferior, and I can't hear the differences because he's better than me.
That said, I do like having music in 192/24. Why? Because I can play with it. I can edit it, there's more headroom. If I feel that "Another Brick in the Wall" just needs a tin whistle part, well, I'll have an easier time editing it in without distortion. But for listening? Nope.
Not a sentence!
training doesn't make one's senses better. it trains the observer's brain to relay the appropriate signals, rather than ignoring them.
i can spot a boom mic in shot almost subliminally. i can spot jitter of all kinds, motion-compensation artifacts, compression artefacts, spots on film (white and black), and can even tell if a cameraman was running out of film, and when the roll was likely to end by looking at the subtle increase in spottiness. other people can't spot these things.
that said, my eyes are pretty poor. my ears are pretty poor, but i can spot when a (perceptibly) lossy source has been used in a master well before i whip out the spectral view. other people can't.
that said, decent mp3 (lame preset standard, or even medium) flies by undetected. ditto the equivalent transparent settings in all audio encoders. ditto a decent h.264 compared to the film scans it came off, when viewed with the same chroma sampling (otherwise it'd be cheating to compare 4:4:4 with 4:2:0).
my wife can tell you every ingredient that goes into a tiny sample of food. i need twice as large a sample to correctly identify only half as many ingredients. my senses are trained (though not as well), but not as sensitive. good thing considering i work in media production, not food.
my point - you're fooling yourself if you think you have better senses than an average joe - you've just trained you brain to pick different things. they probably enjoy the movie more than you...
On warm summer nights I enjoy sitting on my front porch, with a dry gin made from hand-picked juniper berries, some artisan cheese and bread made out of flour that has been milled before sunrise. And if I am in the mood for it, I also enjoy 192kHz music with my bat friends. For us discerning people this is just a standard of living.
Not wanting to go deaf, I use high quality devices with low THD percentages so I can listen at lower volume with maximum impact. Most people don't realize that high volumes are much less necessary as noise is removed and SNR goes up. With a very low noise level, you can play music at relatively low volumes that sounds incredibly good, whereas the high THD injection from a pair of crappy headphones or terrible stereo will cause you to turn up the volume repeatedly to counteract the noise.
- Michael T. Babcock (Yes, I blog)
I could maybe save you an additional 50%. I have a friend who is also deaf in one ear. You could go halfsies and spend only $12 on a headphone. Which one of your ears works?
When I listen to music, its not for the data -- its for the feeling. You should try listening to music for the feeling too ;-)
My opinion.
- Michael T. Babcock (Yes, I blog)
Many people think a "factoid" is a small fact. Actually a factoid is something that sounds true, but is actually false.
You were mistaken. Which is odd, since memory shouldn't be a problem for you
My favourite audiophile rebuttal quote:
"If your hifi costs more than your music collection you have missed the point." - Unknown Source
The higher the sampling rate smoother the signal.
Well... no. There's enough information in a low sampled curve. As TFA explains it, the output isn't "jagged" when played back in analog.
Human perception wise a audio signal recorded at 96KHz sampling rate might well be indistinguishable from one sampled at 192Khz
as explained in the article:
- Yup the human ear won't hear anything aboe 20kHz sounds, because it doesn't have any receptors for that.
But there are some real-world problems that come into the mix. No audio installation is perfect. You always get distortions.
- Thus, a 192kHz sampled file could contain frequencies up to 96kHz. These are sound which can't be heard in theory. In practice if you throw 96kHz frequencies to a sub-optimal speaker, the speaker can barf a lot of distortions, including distortion below the the 20kHz. So not only are you trying to output a sound that can be heard, but you force the speaker to produce bad noise *which* is audible.
But my thinking is that future technologies might let you do interesting things with the extra bit of data which is useless to us right now.
Hard to do anything with those bits at all. We simply lack the anatomic feature to do anything with them. Unless you do something like transpose everything at lower frequencie (slow down everything 2x = move everything 1 octave lower). At which point you aren't really outputing the original sound anymore. You're simply using the data to produce new sounds that weren't here to begin with.
The only practical use-case for this would be zoologist studying animals whose sound are beyond the human hear range. In that case "moving everything a couple of octave down" would help the scientist have an approximation with which he can work (to find rythms or other variation that are inaudible in the original frequency range). But that has nothing to do with hearing music made by human, for humans, with instruments designed for human hearing ranges.
Kind of like with digital pictures which are too noisy or blurred, but which might be cleaned up with future algorithms to give us a slightly more useful picture.
The situation with pictures is slightly different. What you're speaking about is spacial frequency. I.e.: resolution.
And human eyes can percieve way much more than some blurry low-res pictures. And in addition to that, there's this thing called zooming which makes perfectly sense to record picture at higher resolution. Because looking at details is simply looking at the same picture at another scale.
The "visual equivalent" to 192kHz sounds would be recording colours outside the human range. Like recording also infra-reds, microwaves, ultraviolets, and X-Rays.
Things that can't never been seen, because human lack the corresponding apparatus. The only way to get someting out of this extra data would be to transpose it into the visible domain. Thus use pseudo-colours to display levels of low infrared (heat), etc.
Just like the "zoologist" use-case above, there are a lot of scientific use-case where that could actually make sense (as an exemple, think about all the data collected by astronomers).
But in no way is it useful to record X-Rays to enjoy a painting by some known artist. The painting was done by a human painter, for human public, using colours chosen for their effect on an un-aided human visual system, disposed on a canvas in a way which is pleasing to the eyes.
(Well, okay. I know that some scientist use infra-red or X-ray image of paintings to analyse how they were done, what are the layers underneath or if there's even another picture over which the current one was painted. But these are scientist analysing the paint, so we're agin on the "scientific analysis" use-case).
24/192 makes sense as an intermediate format to avoid rounding errors, aliasing during filtering, etc.
There could be also some scientific value to keeping
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
I also spent 4 years studying an EE degree, and although it was not especially focused on signal processing, I now work for a large pro audio company.
Some of the issues pointed to in this and other posts regarding oversampling and AA filters are not really relevant to the subject at hand, given the technology currently in use. A statement like 'oversampling at 192 kHz' shows a lack of knowledge regarding the kinds of audio converters that have been in use for a good while now. A Delta Sigma ADC running with an Fs of 48 kHz might often be oversampling at 3.072 MHz or 6.144 MHz. Anti aliasing filters that many people have mentioned are implemented digitally inside the converter (no need for external analog filters, which may well exhibit many of the problems mentioned), and actually have extremely good pass band ripple.
Look at datasheets for converters from manufacturers such as TI (burr brown), cirrus [page 36 here has detailed plots of 48, 96, and 192 kHz pass pand characterisitcs for the device, highlighting the fact that increasing the sampling rate does not improve pass band ripple for this device (also note the scale is 0.02 dB/div)], AKM, Wolfson micro You will find pass band pass responses that are flat to within less than +/- 0.05 dB over the audible range, and stop band attenuation in excess of 100 dB, whether sampling at 48 kHz or 192 kHz. If you can find anything in actual converter datasheets that points to better converter performance from selecting a higher sampling rate, I would be interested to see it.
All in all, the basics of sampling theory don't really help people to understant the real world issues in designing a moden high end audio device. And in the end, surely the proof of the pudding is in the blind tests, that never seem to show that anybody can tell any difference when moving to higher rates? Even if there were a few people who could hear this difference in some perfect listening envirmonment, would it really make sense for everyone else to go out and buy 192 kHz equipment?
Only if your definition of "perfectly good" is "so convoluted that nobody EVER uses it". ;)
Let's be honest here, verisimilitude exhibits a superlative and ostentatious preponderance of syllables.
"Mind, as manifested by the capacity to make choices, is to some extent present in every electron." -Freeman Dyson
The problem with low-pass filtering was resolved eons ago with a concept called "oversampling."
Only the earliest and ruddiest of CD players (and a lot of computer sound cards) had a brick-wall filter at ~22.5 KHz. The rest of them resampled the input by 4x or 8x, or converted the original signal to PWM, and then applied the anti-aliasing filter at a frequency several octaves above the range of human hearing.
This hypothetically pushed the nastiness inherent of a steep filter to a realm well outside such that humans could hear, and at least far beyond the limited confines of a CD.
Welcome to 1985, where your stated concerns are both accurate and already solved.
Kid-proof tablet..
BS. If the overtones of a flute high C and a piccolo high C are both under 22Khz, then sampling at twice that will catch all the overtones, and replaying the sample at the same rate will perfectly reproduce them.
And if the overtones are over 22Khz, but their lower-order harmonics aren't, the sampling will pick up the harmonics and reproduce them perfectly, even without the existence of the original overtone.
There is no subjectivity in that. An oscilliscope will show you that the overtones and/or their harmonics are all there.
The only step that decides whether or not the overtones have any influence is the quality of the low-pass filter. At 44Khz that can be a bit iffy, so using 48Khz to get a little more headroom is nice, but in practice you won't be able to hear a difference with anything above that.
"I know I will be modded down for this": where's the option '-1, Asking for it'?
You willfully leave out nerds, geeks, dorks, and spazzes? Obvious /. bias! ;)
I8-D
I wish you guys would get this right. There is absolutely no way you can tell the difference between a 15kHz sine wave, square wave, or sawtooth wave (apart from amplitude, perhaps).
Sawtooth waves have even and odd harmonics, and square waves only have odd ones. This means that the first harmonic of a 15kHz sawtooth wave would be at 30kHz, and the square's 3rd harmonic would be at 45kHz. As you pointed out, even if you could hear them, you'd have to have damn good speakers to reproduce.
Three samples is enough to reproduce the 15kHz fundamental per Nyquist.