It's not only the difference (or none) between 16 or 24-bit, but it's mostly about the quality of the DAC that is being used - in this case a very high-end one.
Then again, it's been proven time and time again that the average person can't even hear the difference between mp3 or FLAC.
Considering you'd also need proper high-end monitor headphones, i believe this is more about marketing, the brand and the pitch than it is about what you actually get or can actually use.
+ it can only hold about 2000 FLAC files even with expansion card.
without the big name attached I doubt it would have had much traction. Its merely a very good marketing exercise. Snake oil for the 21st century. No better than all the joke techies used to make about HDMI cables.
I always like Alan Parson's quote, audiophiles use your music to listen to their equipment
A 24-bit sample depth is not completely useless and silly in the way that gold plated HDMI cables are. You can get a real improvement in dynamic range with 24-bit samples, allowing a song to have quiet parts and loud parts with no loss in quality in either, which is hard to do with 16-bit samples without losing quality for the quiet parts or using dynamic range compression and having everything sound roughly the same.
Of course, few people actually want impressive shifts in volume in their music, because prog is dead and no one wants their headphones to suddenly blow their eardrums out. Gotta compress the shit out of that dubstep for the kiddies! Somewhat ironically, greater bit depth in audio makes a bigger difference for film than most music because of this.
It's not portable in the sense of an iPod or an iPhone where you carry it around with you in your pocket. It's portable in the sense that you keep it on your desk at home, grab it and take it to work, then sit it down on your desk at work.
The design is not for everyone, but I think when viewing it through this use case, it has a much more successful design than a flat form-factor like most portable music players
(i realize this is two days old, but i just saw your reply)
But... they could have just made it flat, and then it fits the use case you laid out AND the pocket one. Like, there is zero advantage to the bulky shape.
Another problem is that "hear the difference" is so often defined as "hear the difference between 30 second samples repeated a dozen times." The testing impacts our experience immensely.
It isn't equivalent enough to real music, and the human auditory sensory perception and recall systems just aren't perfect enough to make accurate judgements in those test cases. When testing you end up looking for discrepancies you can identify with words and identifiable momentary observations, things like "I can definitely hear this passage the notes are more muddled together in the compressed file." But you miss so many things that show up as minor feelings of unidentifiable hunches. Of course, you could change the format of the test, but even still, "identifying" is not "listening."
Things you can not identify, or talk about, or remember, or form sentences about, can still impact your musical experience. Our ears and brains are complex, and the range of input that is perceived subconsciously is astounding. We accept this for vision, for taste, for emotions, for memory, but for some reason not for audio. The simple feelings that you can't put your finger on can be important to the experience even if they can't be identified. But if you start to talk about "realism" and "emotion" and "feeling," you are immediately blacklisted by the empirical measurement mafia.
There is so much truth to the fact that scientific measurements and empiricism are important in determining the decisions you make about your audio storage and listening. You shouldn't pay money for things you can prove won't make a difference, and there are so many things out there that fit that bill. But we shouldn't throw out entire possibilities of discussion just because they influence parts of our experience that are subconscious or unidentifiable in an A/B test.
Again, I'm not saying we should disregard scientific evidence, just that listening tests set a limit that might not be accurate. We're finding only things that can be tested, and many parts of the audio experience don't fit the test.
In this case, there are more bits down there, higher resolution at lower levels. It basically means you can turn up the amplification and receive more resolution regardless of amplified volume. But it also means you have more information, and that's worth something.
One thing that is alluded to but perhaps not explored fully is the use of noise-shaping dither techniques which rely on pushing the noise out into the ultrasonic region. Xiph mentions it in the context of 16 bit resolution, but I believe it interacts with the discussion of 192 kHz sampling.
Significant noise power at ultrasonic frequencies (or even a skewing towards higher audible frequencies) represents a problem for tweeters & amplifiers both in terms of the intermodulation distortion mentioned in Xiph's article and in terms of power dissipation at the business end. IIRC Dell and VLC are having a falling-out because VLC's soft clipping is damaging the speakers in certain Dell laptops.
> Significant noise power at ultrasonic frequencies (or even a skewing towards higher audible frequencies) represents a problem for tweeters & amplifiers both in terms of the intermodulation distortion mentioned in the article and in terms of power dissipation at the business end. IIRC Dell and VLC are having a falling-out because VLC's soft clipping is damaging the speakers in certain Dell laptops.
Dell probably didn't bother putting an analog reconstruction filter on their DACs, assuming they used an average Realtek codec the digital filter in the DAC has a minimum cutoff point around 28khz but could be higher when operating with higher sampling rates. They probably also sent the signal into a filterless class D amp to drive the output. Those two things together add up to a lot of ultrasonic crap in the signal.
VLC allows amplification of the INPUT above the sound that was decoded. This is just like replay gain, broken codecs, badly recorded files or post-amplification and can lead to saturation.... snip ...At worse, this will reduce the dynamics and saturate a lot, but this is not going to break your hardware.
Except it can - because the saturation skews the power distribution towards higher frequencies which weren't designed for. This is very, very well known - it's the reason why one always chooses an amplifier with higher output rating than the speakers, often by a factor of two. It's counter-intuitive but driving a low-rated amp into saturation can overheat and destroy tweeters. (Guitar amps get away with it by having massive voice coils on a speaker which will only ever need to reproduce up to about 5kHz.)
I was talking about soft-clipping, not clipping in general. I know that the volume control in VLC can increase the gain of the signal beyond full scale.
Here's the discussion on that Dell issue. https://news.ycombinator.com/item?id=7205759 "Simply said, the sound card outputs at max 10W, and the speakers only can take 6W in, and neither their BIOS or drivers block this."
That's actually good design - the intention being that any clipping can only be due to software (which you can fix), not underpowered amplification hardware (which you can't).
You can definitely "experience" the sub-audio range, but you don't need 24 bits to do that.
I have very sensitive hearing, I hear things most others cannot. I can detect 192kbps vs 320kbps mp3 encoding. I cannot have any switch-mode AC/DC transformers in my bedroom because the switching noise actually keeps me awake (I charge my phone in the kitchen, most people think I'm mad when I complain about the "noise" :)
Still, I've never been able to detect frequencies above 20Khz or hear the difference between 16 and 24 bit. I think anyone claiming to is full of crap.
A lot of music is felt through the body. That's simply a result of low frequency at high levels. That's just one reason why live performance is a more dynamic experience than listening over headphones, but adding dynamic range won't create this effect over headphones.
24-bit is pretty much useless for playback. While human hearing range may extend past 16-bits of dynamic range, that doesn't mean that full range is musically useful. It is, however, quite useful to record and mix in 24-bit for the headroom.
Weird that the author uses a 'ordinary MP3' as source for the test. Garbage in, garbage out. Right? Would be a better test with a very high rez audio file.
I suspect it's heavily compressed audio.
His argument is that you'll have a hard time hearing it at all, so if anything's going to help you hear it in the first place, it's garbage in. garbage recordings are the only ones which will survive this test.
Anyway, nobody here is both qualified and willing at the same time to tell you what you need to know about this topic. The article referred to, written by Monty at Xiph, should give you a very good overview of how this works .
The author is likely using a 16-bit audio card to listen to the result, which is going to clip the least significant bits off anyway. If you undo the shifting with sox -v 4096, the quantization noise in the 16-bit version is audible, even on laptop speakers.
This is my only issue with the article - it's alluded that the -v and -b flags only perform multiplication and bit shifting.
By default, this is not true because sox will use a higher bit-count and perform dithering when converting to a lower bit format (ref http://blog.beatunes.com/2014/04/does-24-bit-audio-matter.ht... ) in order to remove the extra bits while reducing the impact on dynamic range.
In all fairness though, this is the same argument used in the Xiph article ( http://people.xiph.org/~xiphmont/demo/neil-young.html ) in an attempt to justify the use of 16-bit over 24-bit, the prior allowing for a comparable dynamic range with the use of appropriate dithering techniques. I just think it's important to note that the process used with sox is doing more than simple arithmetic bit-shifts/multiplication.
The author admits that "16 bits does not quite cover the entire theoretical dynamic range of the human ear in ideal conditions". But then the author concludes "let's not use 24 bits anyway because it doesn't really help anyway, and it wastes space"
Wastes space? In an age where we stream gigabyte movies? Do we need to remind the author that we're no longer in 1982 and that we have evolved beyond floppy disks?
>> Wastes space? In an age where we stream gigabyte movies? Do we need to remind the author that we're no longer in 1982 and that we have evolved beyond floppy disks?
Clever way of leaving out just how much space it takes up (takes up 6 times the space [0][1]) also as both this author and Xiph concluded you can't really hear the difference. When we stream HD (As you put it "Gigabyte movies") we are getting a clearly superior product. I can easily see the difference between 480/720/1080 whereas I seriously doubt I can hear the difference between 16 and 24bit audio. Yes space is cheaper than it used to be but it's neither free nor unlimited, not to mention the primary way people listen to music nowadays is probably on mobile devices where space is still at a premium or they enjoy their music via a streaming service a la Spotify/Rdio/Google Music where bandwidth is at a premium (And storage factors in here as well due to either being able to store less songs locally or in cache due to larger file size).
Also the point of both articles is that: You can't hear the difference.
Look at MP3 vs WAV (or FLAC), MP3 won out (for consumers) because of it's filesize being so much smaller. Look at the compression differences:
>> Uncompressed audio as stored on an audio-CD has a bit rate of 1,411.2 kbit/s,[note 2] so the bitrates 128, 160 and 192 kbit/s represent compression ratios of approximately 11:1, 9:1 and 7:1 respectively. [0]
At the best quality listed here (192kbps) MP3 is 7 times smaller than it's WAV counterpart. I just don't see people paying for 6x the space for something they can't tell the difference between. If anything history has proven this not to be the case.
The six times is also on the wrong side of your equation, I guess.
But if you need a source, I can cite what you cited:
> Unfortunately, there is no point to distributing music in 24-bit/192kHz format. Its playback fidelity is slightly inferior to 16/44.1 or 16/48, and it takes up 6 times the space.
They were talking about both sampling rate and bit depth.
Oh, and just for the record, I do agree with the general idea that it's useless to ship audio to consumers as 192/24, for the same reason it's unwieldy to publish photos online as 20+ MiB raw files (even if browsers could display them). I was just noting that the size comparison included both sampling rate and bit depth and not just one of them.
It also only applies directly to raw audio like pcm. Once you encode it (to, say, flac) some of that size increase could disappear depending on the waveform.
I don't know how mp3 is defined for higher dynamic range, but I know frequency range/sampling rate doesn't matter, because a 128kbps mp3 is 128kbps whether the frequency cutoff is 14kHz or 20kHz. Isn't a 192kbps mp3 from a 16/48 master the same size as a 192kbps mp3 from a 24/192 master?
Approximately? Either it does or it doesn't. What I'm thinking is that a 4x increase in sampling rate doesn't necessarily lead to a 4x increase in waveform complexity except for certain types of random noise.
With the references to 192kbps and mistakes, it wasn't clear what parent was talking about.
>Wastes space? In an age where we stream gigabyte movies? Do we need to remind the author that we're no longer in 1982 and that we have evolved beyond floppy disks?
Well, he probably also knows that we now increasingly use SSD disks for our laptops, that are more expensive and sold in lower capacities than HDs on average.
Or that we also listen to music on mobile phones, with like 16 or 32 GB or space (which we also want for apps, photos, videos and other stuff).
Or that we might have "evolved beyond floppy disks" but we have also evolved beyond having 20-100 albums in our collection. In fact digital music collections of 1000 or 10.000 albums are not uncommon.
Now, 1000 albums of 500MB each would make it 500GB for an music lover's collection. That's a non starter in 99% of modern laptops, and impossible on an mobile phone.
I'd rather have it compressed more or in less than 24bits, and fit more stuff to have with me to listen.
(Also keep in mind that music lover is not the same as audiophile).
I think the definition of "wastes" is very important here.
Don't use space that is literally wasted, because it adds zero information.
It's equivalent to a database which adds a sequence of 32 bits of 1's after every field. It adds nothing for any purpose, yet it uses space. This space is truly wasted.
This test is flawed in a few ways. The and the error lies in the source material. The why would there be a perciveable difference between the same inferior input shown to both encoding formats.
Changing between encoding formats is a topic of its own but the problem here is similar to getting a 640x480 compressed jpeg and putting it into a 1920x1080 lossless format and the same source into a 1280x720 lossless file and saying. What's the difference? They both look crappy!
Many people have an opinion on high res audio formats but they are so many places to go wrong.
Audio should be given a little more credit for it immersive ability. When done right it can take you places. Something to consider is that our eyes can only see 1 octave of information our ears can hear 10 octaves. Whilst these high res formats aren't needed for every application they do make a big difference when the source material can take advantage of it. I feel in 2014 music should be released in the best format available from the studio and if u want a crappier version so u can shove 5000 songs on your iWhatever that's your choice.
It's not the same as the same inferiour source material.
When you apply negative gain to the 16-bit signal, it has less amplitude in absolute terms, but it should have all the resolution of a whole 16-bit sample in that space.
The point is that whether that 16-bit signal is represented with the lower 4 bits of a 16-bit sample, or the lower 12 bits of a 24-bit sample, you can scarcely hear anything at all, let alone a qualitative difference between the signals.
192KHz is not a higher quality format, it is a production format. The purpose is to reduce the cost and finality of anti-aliasing filters when sampling the signal, it is just cheaper to manufacture. There's a good reason for this being the standard in professional audio, they need to buy a LOT of audio interfaces, many studios will have thousands of such inputs.
Thanks to the basic laws governing signals, we know with near certainty that not only is 192KHz overkill, but so is 48KHz, and so is 44.1. Without significant new evidence showing humans hearing signals with frequencies greater than 24KHz, you will not make any convincing argument as to why we should go with any sample rate higher than 48KHz for human listening.
As for 24-bit, it is another production interchange format. It's there so that you don't need to stand around adjusting gain knobs on audio interfaces so that you get decent fidelity but also don't clip. With 24-bit you can just sample your audio once, and assuming it's within a reasonable range, you can adjust the gain in the discrete signal. There is some indication that 16-bit sampling is less than completely ideal. The very best ears in humanity(newborn ears) distinguish about 21 bits in the safe ranges of amplitude, 24-bit may make sense in an audio system for newborn babies.
I feel that in 2014, music should be released in the best format available from the studio, and if you want a crappier version so that you can shove only 100 songs on your iWhatever, that's your choice.
The methodology might be flawed, but the point stands: We can't use the full dynamic range of a 24 bit signal for audio. Great converters have 20-21 bits of range and if we fully use this range, hearing damage will start at maximum volume levels. If the volume is kept the same between the two files, there is a 48dB drop in level - not inaudible, but there are very few source materials that will actually use the extra dynamic range. On top of that due to fletcher munson curves we will only hear content in the 1-4kHz range at that low volume level.
Great converters have 20-21 bits of range and if we fully use this range, hearing damage will start at maximum volume levels
but that doesn't have anything to do with number of bits? I mean, an audio player system is basically a digital stream sent into a DA converter outputting an analog signal which is then fed into an amplifier where the signal is amplified and then sent into speakers of some sort. No matter if you have an 8bit audio stream, or a 128bit audio stream, the stages afterwards play a much larger role in the hearing damage caused. Also not all DA converters ouput the same level to begin with. Some are -5V to 5V while others (single power supply) can be 0V to 5V, etc etc.
Hearing damage as a result of loudness; the electronics have little to do with it. If you set up a 96dB (16 bit) system such that you can hear the lowest bit toggling, your maximum volume will be 96dB above that. 96dB SPL is loud but not painful for average listening.
If you set up a 20 bit system (120dB) so that you can hear the lowest bit toggling, 120dB above that will most certainly cause hearing damage if used for very long.
144 dB (24 bit) or more SPL will cause near instantaneous damage and pain.
It would take highly dynamic music to really hear more bits. The quiet passages is where it makes the difference. Most music is compressed beyond the need for anything other than 40db of dynamic range.
I barely can tell the difference between high sample rate 8 or 12 bit audio and 16 bit audio. In some circles I hear people talk about that "gritty" 12 bit sound, but its usually not properly dithered, heavily filtered, 20khz or less sample rate and generally destroyed on purpose.
Audio is generally the best arena for snake oil since subjectivity is so high. I cannot believe that Focal Grande Utopia EM loudspeakers(180,000 usd MSRP) exist in quantity, but I know two people who own a set in their (comparatively) modest living room. I have heard them in action, they sound great, but Im not sure that I heard anything in any of the music that I hadnt heard with good headphones or speakers. At what point is it "good enough"? I own plenty of recordings that no amount of hifi audiophilia will help.(many that I recorded myself ;) )
For the test the quality of the input does not matter.
You can conduct the same test with a beautiful 24-bit recording as source, you would still not be able to hear much in the end.
And that's the point.
When you print a photo in size 9x13cm, does it matter whether you print it in 10,000dpi or 20,000dpi? Hardly, unless you always run around with a microscope.
And it does not matter, whether the photo was originally a good picture or not.
> When you print a photo in size 9x13cm, does it matter whether you print it in 10,000dpi or 20,000dpi? Hardly, unless you always run around with a microscope.
It can matter when you put it in a scanner though. I tried scanning some old photos, they look horrible compared to even my puny 8mp camera. Then I tried scanning a piece of cloth, and was blown away by the detail..
I have pretty good headphones (Sony MDR-V6). I can hear decently enough (although getting older). Like I noticed that the Deutche Gramaphone 111 years music collection I bought (256kbps mp3s) has taped performances from the 70 & 80s which have some tape his.
Honestly as an end user its really good enough. If I was remixing this stuff and processing the audio I would want the non-compressed. (similar to shooting Raw photographs ). But at 256kbps honestly it really is amazingly good. thinking back to cassette tapes and records, its astonishing.
As for the tape hiss on the classical music, the performances are so good I just ignore it. Sometimes I think people get worked up about the quality of the recording and ignore the quality of the performance.
re: "It's useless because most people can't hear the difference."
Sure, but when sampling and then processing things, you can consider it "headroom". Just like it's not hard to edit a photo in a few steps that the limitations of 8 bits per pixel begin to show. Though I suspect sampling frequency is a lot more important there, 24bit can't possibly be "too much".
Yes, I understand that this is not a good argument for consumer products to be more expensive and need more storage, just so musicians of the future, or pirate musicians of today, can have a better time. So I think the best plan is to encourage everybody to become hobby musicians, then it would be an easy sell.
There has been a lot of talk about quantative aspects of but little about the qualitative. What does a 18khz square wave look like wher recorded and reproduced in 44.1 kHz. How is phase affecte?. Phase of an audio signal reaching the ears helps you perceive distance and position. How many samples are used to represent the frequencies at the higher end of the audible range. What is the quantization error difference between 16 and 24.
And if the author can't hear the difference between 4bit audio and 12bit audio I question what he/she is listening on. The aliasing would be HUGE.
An 18 kHz square wave sampled at 44 kHz looks like an 18 kHz sine wave, everything after the fundamental frequency is well outside the Nyquist limit and will have been thrown away by anti-aliasing filters. And furthermore you couldn't hear it even if it wasn't. Fourier decomposition of a square wave gives the sum of odd multiples of the fundamental, the next frequency is 3 x 18 kHz = 54 kHz.
Okay now actually look at the response on Matlab. As you get closer to the nyquist freq there are less samples to describe the wave. And while you get a bastadised 18k signal it's a far throw from what you put in. Anf while most are crying your splitting hairs the frequency response and dynamic range arguments pale in comparison to making the exact wave you put into the encoder come out. On the right system with the right recording the tiniest nuance in a room reverb helps trick your brain into believing the sound as actually happening. And this is not important to all listeners. But people saying that it makes no difference and is not important as a format to release music in are thinking only of their current needs and experiences. If you have actually heard a good recording on a good system in a good room you will know what I'm talking about. If listening experience has been laptop speakers, headphones and your mum and dads mini system then I totally agree any more than 320kbs mp3s are overkill. But to say that high res formats gave no place in consumer land is to show your lack of understanding of different people and different needs. In the age of iTunes you surely you buy the album and download it in whatever format u see fit. If u think audiophiles are wackos go get ya 320's. I would like to get it as I came fr the studio. The best it can be is as it came from the studio.
First off, forget about MP3; I'm not arguing for a lossy standard.
You can't get an 18 kHz square wave out of a system with 44 kHz sampling. You need at least 1 harmonic before it'll even LOOK square, and that requires a frequency response out to 54 kHz, ie. a sampling frequency of 108 kHz. You CERTAINLY won't find one in a reverb tail, even assuming you had a generator for one in the first place (you might JUST get one from a cymbal crash, but I don't think the physics works)
The point being, your source material can't contain an 18 kHz square wave either since it's been through a studio production system with the same antialiasing filters.
Since you know nothing about me but seem to be making assumptions anyway, here's some background. I've worked in broadcast audio; I own studio recordings in 24 bit / 192 kHz (Linn release of Mozart's Requiem, studio master series). I also own studio equipment that can actually play it. Audiophiles are, by and large, cash cows for companies with no scruples.
Good an audio nerd. I am making the point about the square wave close to the nyquist to point out the short comings of a format for accurately reproducing an input. Square waves in the real world are rare but I am arguing for a format that produces the most accurate representation of the intended signal. Imagine the situation where i have my guitar cab set up and I have a square wave (distortion) coming out of it and I far mic it up so as to capture the room a give a feeling of space. To really feel like your there you would want the resulting complex wave made up of the 18k direct sound from the cab and the room response a recording medium that can't do that accurately is second rate especialy when the formats are out there. And the higher the sample rate the further from the nyquist that 18k is and the more samples that can be used to describe the resulting wave an the more convinced my brain is that sound is real.
...right up to about 20 kHz, whereafter YOU CAN'T HEAR THEM. Hence 44 kHz sampling.
Seriously, A/B test this, you might be surprised.
Also if you think that's anything like a square wave coming out of a guitar speaker (or that that is even desirable in the most case), I've got a bridge to sell you. And yes, I do play.
Okay here is a picture of what I'm trying to explain. And the author of this picture used a frequency much further inside human hearing range. This is transient response test I guess. My main argument is for the verbatim capture of the input wave. It will make the sound at 10k but it isn't the same wave that went in.
> What does a 18khz square wave look like wher recorded and reproduced in 44.1 kHz.
It looks like an 18khz sine wave, possibly with slightly reduced amplitude depending on the anti-aliasing filter rolloff and fc, but not enough to be audible (18khz isn't audible for a large portion of the population anyway).
> How is phase affecte?.
Probably delayed a bit by the anti-aliasing filter.
> Phase of an audio signal reaching the ears helps you perceive distance and position.
Not at 18khz, the wavelength is too short for your ears to notice any realistic group delay. High frequency localization is most done by ILD and effects caused by ear shape.
> What is the quantization error difference between 16 and 24.
Quantization error in a DAC just defines the noise floor. 16 and 24bit DACs are usually within a few dB of each other in dynamic range, it's really not audible.
> And if the author can't hear the difference between 4bit audio and 12bit audio I question what he/she is listening on. The aliasing would be HUGE.
http://www.eirec.com/DPimages/digisqwvtest.jpg
Here is an example of higher sampling rates being useful INSIDE the human hearing range. This example shows transient response. Frequency response and dynamic range are low hanging fruit in high res audio debates. To argue that you can't hear it is a mute point when accuracy is the point, this is a reference quality format. Temporal resolution is the next area to persue if accuracy is of concern. If file size is of concern then high res is not aplicable. What is everyone arguing here? That if they any hear it on their setup it is of no use?
At the last year's IFA (http://b2b.ifa-berlin.com/en/Home.html) there was a demonstration of a 24-bit audio player with corresponding high quality audio files and headphones. The results, even though it was a single test, were very good. Of course the issue is, the whole chain has to be 24-bit, if you're going to pump your 128kbps MP3 there the difference is probably negligible (not to mention the 1-bit DACs commonly used, that's one of the elephants in the room)
About the MP3 files, yes, the difference is audible. But of course you need good headphones.
So I converted a few tracks to 8-bit with noise shaped dither. Have a listen and remember that the noise floor is now 48dB higher (8-bits) than it would be with 16-bit.
More to it than frequency response and dynamic range. The goal should be for the most accurate reproduction of the input waveform. Here is a pic of transient response for different sample rates.
http://www.eirec.com/DPimages/digisqwvtest.jpg
Frequency response and dynamic range are low hanging fruit in high res audio debates. They are not the only factors in accurate reproduction of an input wave. What is rarely considered is the temporal resolution of the format. Which is the formats ability to describe change over time. To plot a graph but depth is the y axis resolution and sample rate the x axis. Again I'm talking to temporal resolution. Transient response is directly effect by this. The link below shows the advantages of hi res audio formats inside the human hearing range. So if accuracy is of concern high res formats DO hold more information about the original wave form. If file size is of concern hi res is not applicable.
I tried to find the context for that image but there wasn't anything on the site. The image is most certainly incorrect -- or more likely, looking at the output of the analog electronics after the DAC.
Sampling theory says that a perfect square wave can be represented at any frequency below Nyquist. That doesn't mean that the codec or the analog electronics are capable of responding instantly at those frequencies, but that has nothing to do with the fact that a 1kHz square wave can be perfectly sampled with a 44.1kHz sample rate. The image is simply incorrect.
Transient response in the real world is generally limited only by the acoustic transducer response of the system, because everything else has the ability to respond much faster than audio rates. With Pono, this means that the earbuds or headphones you use with the player will have a greater affect on the transient response than the electronics inside.
This image is from RME a sound card maufaturer. And the image is of the analog output after dac. So still in the electrical domain. Short comings of a playback system has what to do with the accuracy of a recoding medium? There are SO many factors that stop and audio signal reproduction from being perfect. But the signal domain is the easiest to make better. Transducers that have to fight the law of physics for accuracy are the obvious weak point. But as with anything rubish in = rubbish out. Before you even get to the weakest point of an audio system the transducer (the speaker) transistors and amplification have their own short comings the accuracy of an amplifiers transient response is measured by its slew rate. A square wave is impossible for a speaker cone to reproduce in theory the rising and falling edge is describing an instantaneous movement. A speaker cone can't do this, inertia says no. And this is called distortion. As every part of the system introduces its own distortion the accuracy gets less and less. To accept distortion at the signal level in the persuit of accuracy is counterintuive. But if your system can't take advantage of this extra accuracy, go ahead and use a lesser format, no sense in the extra file size. If you want a reference quality signal, use the high res formats.
To be clear here I'm not saying I want to listen to digitally encoded square waves, a square wave is a tool for showing the accuracy of input=output. A square wave being the hardest analog waveform to capture and reproduce. I am objecting to the idea that hi res formats have no advantage over what we are used to as consumers. It is the content distributors that are taking advantage saying that 24bit/96k is more expensive. The bandwidth and storage does not make this a more premium product you should be paying for the album not the format. But why limit choice because the masses can't see the benefits. If you buy the album you as the consumer can choose the format. I'm not sure if all this arguing stands up to pick whatever format suits your needs. This argument should be if I bought the album why can I choose the format most applicable to my playback needs/desires. And if your concern is that company's charge more for the high res, that refeclts poorly of the distribution company. If you think it's a waste of space just download the lesser format. If I had a choice I'd download it as it left the studio/mastering house.
Not sure what point you're trying to make here, but I was responding to your post implying that 192kHz sampling rate improves transient response. It doesn't improve it at all. While there is sampling "distortion" by way of quantization error, this error is inaudible because we're sampling at 24 bits, beyond the dynamic range of human hearing. High sample rates actually REDUCE audio quality because they sample ultrasonics that will interact within the audible range and potentially damage speakers or pick up EMI interference.
It has been proven over and over via ABX testing that high res formats are completely indistinguishable from a 24b/44.1kHz master. And further studies have proven most audio engineers can't distinguish between lossless audio and 320kbps MP3. That's what the Xiph link elsewhere in the thread shows.
Did anyone else actually try this? I thought the differences were quite noticeable. I ran it through Foobar 2000's ABX Comparator and I got the right answer 5 out of 5 times.
76 comments
[ 3.2 ms ] story [ 146 ms ] threadThen again, it's been proven time and time again that the average person can't even hear the difference between mp3 or FLAC.
Considering you'd also need proper high-end monitor headphones, i believe this is more about marketing, the brand and the pitch than it is about what you actually get or can actually use.
+ it can only hold about 2000 FLAC files even with expansion card.
I always like Alan Parson's quote, audiophiles use your music to listen to their equipment
Of course, few people actually want impressive shifts in volume in their music, because prog is dead and no one wants their headphones to suddenly blow their eardrums out. Gotta compress the shit out of that dubstep for the kiddies! Somewhat ironically, greater bit depth in audio makes a bigger difference for film than most music because of this.
The design is not for everyone, but I think when viewing it through this use case, it has a much more successful design than a flat form-factor like most portable music players
But... they could have just made it flat, and then it fits the use case you laid out AND the pocket one. Like, there is zero advantage to the bulky shape.
It isn't equivalent enough to real music, and the human auditory sensory perception and recall systems just aren't perfect enough to make accurate judgements in those test cases. When testing you end up looking for discrepancies you can identify with words and identifiable momentary observations, things like "I can definitely hear this passage the notes are more muddled together in the compressed file." But you miss so many things that show up as minor feelings of unidentifiable hunches. Of course, you could change the format of the test, but even still, "identifying" is not "listening."
Things you can not identify, or talk about, or remember, or form sentences about, can still impact your musical experience. Our ears and brains are complex, and the range of input that is perceived subconsciously is astounding. We accept this for vision, for taste, for emotions, for memory, but for some reason not for audio. The simple feelings that you can't put your finger on can be important to the experience even if they can't be identified. But if you start to talk about "realism" and "emotion" and "feeling," you are immediately blacklisted by the empirical measurement mafia.
There is so much truth to the fact that scientific measurements and empiricism are important in determining the decisions you make about your audio storage and listening. You shouldn't pay money for things you can prove won't make a difference, and there are so many things out there that fit that bill. But we shouldn't throw out entire possibilities of discussion just because they influence parts of our experience that are subconscious or unidentifiable in an A/B test.
Again, I'm not saying we should disregard scientific evidence, just that listening tests set a limit that might not be accurate. We're finding only things that can be tested, and many parts of the audio experience don't fit the test.
In this case, there are more bits down there, higher resolution at lower levels. It basically means you can turn up the amplification and receive more resolution regardless of amplified volume. But it also means you have more information, and that's worth something.
Significant noise power at ultrasonic frequencies (or even a skewing towards higher audible frequencies) represents a problem for tweeters & amplifiers both in terms of the intermodulation distortion mentioned in Xiph's article and in terms of power dissipation at the business end. IIRC Dell and VLC are having a falling-out because VLC's soft clipping is damaging the speakers in certain Dell laptops.
Dell probably didn't bother putting an analog reconstruction filter on their DACs, assuming they used an average Realtek codec the digital filter in the DAC has a minimum cutoff point around 28khz but could be higher when operating with higher sampling rates. They probably also sent the signal into a filterless class D amp to drive the output. Those two things together add up to a lot of ultrasonic crap in the signal.
VLC allows amplification of the INPUT above the sound that was decoded. This is just like replay gain, broken codecs, badly recorded files or post-amplification and can lead to saturation. ... snip ... At worse, this will reduce the dynamics and saturate a lot, but this is not going to break your hardware.
Except it can - because the saturation skews the power distribution towards higher frequencies which weren't designed for. This is very, very well known - it's the reason why one always chooses an amplifier with higher output rating than the speakers, often by a factor of two. It's counter-intuitive but driving a low-rated amp into saturation can overheat and destroy tweeters. (Guitar amps get away with it by having massive voice coils on a speaker which will only ever need to reproduce up to about 5kHz.)
But I always wonder how much of the sound we hear by feeling (skin, hair vibrations). Because maybe I can't hear 1Hz, but I can very well feel it.
21-bit would make for an awkward file format. If we can hear 17 bits of range that's enough to justify storing music as 24-bit.
I have very sensitive hearing, I hear things most others cannot. I can detect 192kbps vs 320kbps mp3 encoding. I cannot have any switch-mode AC/DC transformers in my bedroom because the switching noise actually keeps me awake (I charge my phone in the kitchen, most people think I'm mad when I complain about the "noise" :)
Still, I've never been able to detect frequencies above 20Khz or hear the difference between 16 and 24 bit. I think anyone claiming to is full of crap.
24-bit is pretty much useless for playback. While human hearing range may extend past 16-bits of dynamic range, that doesn't mean that full range is musically useful. It is, however, quite useful to record and mix in 24-bit for the headroom.
Anyway, nobody here is both qualified and willing at the same time to tell you what you need to know about this topic. The article referred to, written by Monty at Xiph, should give you a very good overview of how this works .
http://people.xiph.org/~xiphmont/demo/neil-young.html
Maybe people who work in professional (not consumer) audio? Who design these systems for a living?
By default, this is not true because sox will use a higher bit-count and perform dithering when converting to a lower bit format (ref http://blog.beatunes.com/2014/04/does-24-bit-audio-matter.ht... ) in order to remove the extra bits while reducing the impact on dynamic range.
In all fairness though, this is the same argument used in the Xiph article ( http://people.xiph.org/~xiphmont/demo/neil-young.html ) in an attempt to justify the use of 16-bit over 24-bit, the prior allowing for a comparable dynamic range with the use of appropriate dithering techniques. I just think it's important to note that the process used with sox is doing more than simple arithmetic bit-shifts/multiplication.
http://people.xiph.org/~xiphmont/demo/neil-young.html
Wastes space? In an age where we stream gigabyte movies? Do we need to remind the author that we're no longer in 1982 and that we have evolved beyond floppy disks?
Clever way of leaving out just how much space it takes up (takes up 6 times the space [0][1]) also as both this author and Xiph concluded you can't really hear the difference. When we stream HD (As you put it "Gigabyte movies") we are getting a clearly superior product. I can easily see the difference between 480/720/1080 whereas I seriously doubt I can hear the difference between 16 and 24bit audio. Yes space is cheaper than it used to be but it's neither free nor unlimited, not to mention the primary way people listen to music nowadays is probably on mobile devices where space is still at a premium or they enjoy their music via a streaming service a la Spotify/Rdio/Google Music where bandwidth is at a premium (And storage factors in here as well due to either being able to store less songs locally or in cache due to larger file size).
[0] http://people.xiph.org/~xiphmont/demo/neil-young.html
[1] http://blog.beatunes.com/2014/04/does-24-bit-audio-matter.ht...
Look at MP3 vs WAV (or FLAC), MP3 won out (for consumers) because of it's filesize being so much smaller. Look at the compression differences:
>> Uncompressed audio as stored on an audio-CD has a bit rate of 1,411.2 kbit/s,[note 2] so the bitrates 128, 160 and 192 kbit/s represent compression ratios of approximately 11:1, 9:1 and 7:1 respectively. [0]
At the best quality listed here (192kbps) MP3 is 7 times smaller than it's WAV counterpart. I just don't see people paying for 6x the space for something they can't tell the difference between. If anything history has proven this not to be the case.
[0] https://en.wikipedia.org/wiki/MP3#Bit_rate
But if you need a source, I can cite what you cited:
> Unfortunately, there is no point to distributing music in 24-bit/192kHz format. Its playback fidelity is slightly inferior to 16/44.1 or 16/48, and it takes up 6 times the space.
They were talking about both sampling rate and bit depth.
Oh, and just for the record, I do agree with the general idea that it's useless to ship audio to consumers as 192/24, for the same reason it's unwieldy to publish photos online as 20+ MiB raw files (even if browsers could display them). I was just noting that the size comparison included both sampling rate and bit depth and not just one of them.
filesize(192kHz + 24bit) = 6 * filesize(48kHz + 16bit)
It also only applies directly to raw audio like pcm. Once you encode it (to, say, flac) some of that size increase could disappear depending on the waveform.
I don't know how mp3 is defined for higher dynamic range, but I know frequency range/sampling rate doesn't matter, because a 128kbps mp3 is 128kbps whether the frequency cutoff is 14kHz or 20kHz. Isn't a 192kbps mp3 from a 16/48 master the same size as a 192kbps mp3 from a 24/192 master?
With the references to 192kbps and mistakes, it wasn't clear what parent was talking about.
Betamax didn't lose to VHS because it was worse.
Well, he probably also knows that we now increasingly use SSD disks for our laptops, that are more expensive and sold in lower capacities than HDs on average.
Or that we also listen to music on mobile phones, with like 16 or 32 GB or space (which we also want for apps, photos, videos and other stuff).
Or that we might have "evolved beyond floppy disks" but we have also evolved beyond having 20-100 albums in our collection. In fact digital music collections of 1000 or 10.000 albums are not uncommon.
Now, 1000 albums of 500MB each would make it 500GB for an music lover's collection. That's a non starter in 99% of modern laptops, and impossible on an mobile phone.
I'd rather have it compressed more or in less than 24bits, and fit more stuff to have with me to listen.
(Also keep in mind that music lover is not the same as audiophile).
Don't use space that is literally wasted, because it adds zero information.
It's equivalent to a database which adds a sequence of 32 bits of 1's after every field. It adds nothing for any purpose, yet it uses space. This space is truly wasted.
Same with storing audio we cannot possibly hear.
When you apply negative gain to the 16-bit signal, it has less amplitude in absolute terms, but it should have all the resolution of a whole 16-bit sample in that space.
The point is that whether that 16-bit signal is represented with the lower 4 bits of a 16-bit sample, or the lower 12 bits of a 24-bit sample, you can scarcely hear anything at all, let alone a qualitative difference between the signals.
192KHz is not a higher quality format, it is a production format. The purpose is to reduce the cost and finality of anti-aliasing filters when sampling the signal, it is just cheaper to manufacture. There's a good reason for this being the standard in professional audio, they need to buy a LOT of audio interfaces, many studios will have thousands of such inputs.
Thanks to the basic laws governing signals, we know with near certainty that not only is 192KHz overkill, but so is 48KHz, and so is 44.1. Without significant new evidence showing humans hearing signals with frequencies greater than 24KHz, you will not make any convincing argument as to why we should go with any sample rate higher than 48KHz for human listening.
As for 24-bit, it is another production interchange format. It's there so that you don't need to stand around adjusting gain knobs on audio interfaces so that you get decent fidelity but also don't clip. With 24-bit you can just sample your audio once, and assuming it's within a reasonable range, you can adjust the gain in the discrete signal. There is some indication that 16-bit sampling is less than completely ideal. The very best ears in humanity(newborn ears) distinguish about 21 bits in the safe ranges of amplitude, 24-bit may make sense in an audio system for newborn babies.
I feel that in 2014, music should be released in the best format available from the studio, and if you want a crappier version so that you can shove only 100 songs on your iWhatever, that's your choice.
but that doesn't have anything to do with number of bits? I mean, an audio player system is basically a digital stream sent into a DA converter outputting an analog signal which is then fed into an amplifier where the signal is amplified and then sent into speakers of some sort. No matter if you have an 8bit audio stream, or a 128bit audio stream, the stages afterwards play a much larger role in the hearing damage caused. Also not all DA converters ouput the same level to begin with. Some are -5V to 5V while others (single power supply) can be 0V to 5V, etc etc.
If you set up a 20 bit system (120dB) so that you can hear the lowest bit toggling, 120dB above that will most certainly cause hearing damage if used for very long.
144 dB (24 bit) or more SPL will cause near instantaneous damage and pain.
I barely can tell the difference between high sample rate 8 or 12 bit audio and 16 bit audio. In some circles I hear people talk about that "gritty" 12 bit sound, but its usually not properly dithered, heavily filtered, 20khz or less sample rate and generally destroyed on purpose.
Audio is generally the best arena for snake oil since subjectivity is so high. I cannot believe that Focal Grande Utopia EM loudspeakers(180,000 usd MSRP) exist in quantity, but I know two people who own a set in their (comparatively) modest living room. I have heard them in action, they sound great, but Im not sure that I heard anything in any of the music that I hadnt heard with good headphones or speakers. At what point is it "good enough"? I own plenty of recordings that no amount of hifi audiophilia will help.(many that I recorded myself ;) )
And that's the point.
When you print a photo in size 9x13cm, does it matter whether you print it in 10,000dpi or 20,000dpi? Hardly, unless you always run around with a microscope.
And it does not matter, whether the photo was originally a good picture or not.
It can matter when you put it in a scanner though. I tried scanning some old photos, they look horrible compared to even my puny 8mp camera. Then I tried scanning a piece of cloth, and was blown away by the detail..
Honestly as an end user its really good enough. If I was remixing this stuff and processing the audio I would want the non-compressed. (similar to shooting Raw photographs ). But at 256kbps honestly it really is amazingly good. thinking back to cassette tapes and records, its astonishing.
As for the tape hiss on the classical music, the performances are so good I just ignore it. Sometimes I think people get worked up about the quality of the recording and ignore the quality of the performance.
This is the main problem with audiophiles.
Improve your equipment until it makes your music awesome, then just enjoy it.
Sure, but when sampling and then processing things, you can consider it "headroom". Just like it's not hard to edit a photo in a few steps that the limitations of 8 bits per pixel begin to show. Though I suspect sampling frequency is a lot more important there, 24bit can't possibly be "too much".
Yes, I understand that this is not a good argument for consumer products to be more expensive and need more storage, just so musicians of the future, or pirate musicians of today, can have a better time. So I think the best plan is to encourage everybody to become hobby musicians, then it would be an easy sell.
You can't get an 18 kHz square wave out of a system with 44 kHz sampling. You need at least 1 harmonic before it'll even LOOK square, and that requires a frequency response out to 54 kHz, ie. a sampling frequency of 108 kHz. You CERTAINLY won't find one in a reverb tail, even assuming you had a generator for one in the first place (you might JUST get one from a cymbal crash, but I don't think the physics works)
The point being, your source material can't contain an 18 kHz square wave either since it's been through a studio production system with the same antialiasing filters.
Since you know nothing about me but seem to be making assumptions anyway, here's some background. I've worked in broadcast audio; I own studio recordings in 24 bit / 192 kHz (Linn release of Mozart's Requiem, studio master series). I also own studio equipment that can actually play it. Audiophiles are, by and large, cash cows for companies with no scruples.
Seriously, A/B test this, you might be surprised.
Also if you think that's anything like a square wave coming out of a guitar speaker (or that that is even desirable in the most case), I've got a bridge to sell you. And yes, I do play.
Okay here is a picture of what I'm trying to explain. And the author of this picture used a frequency much further inside human hearing range. This is transient response test I guess. My main argument is for the verbatim capture of the input wave. It will make the sound at 10k but it isn't the same wave that went in.
It looks like an 18khz sine wave, possibly with slightly reduced amplitude depending on the anti-aliasing filter rolloff and fc, but not enough to be audible (18khz isn't audible for a large portion of the population anyway).
> How is phase affecte?.
Probably delayed a bit by the anti-aliasing filter.
> Phase of an audio signal reaching the ears helps you perceive distance and position.
Not at 18khz, the wavelength is too short for your ears to notice any realistic group delay. High frequency localization is most done by ILD and effects caused by ear shape.
> What is the quantization error difference between 16 and 24.
Quantization error in a DAC just defines the noise floor. 16 and 24bit DACs are usually within a few dB of each other in dynamic range, it's really not audible.
> And if the author can't hear the difference between 4bit audio and 12bit audio I question what he/she is listening on. The aliasing would be HUGE.
What does aliasing have to do with bit depth?
(and I'm inclined to agree)
About the MP3 files, yes, the difference is audible. But of course you need good headphones.
https://www.dropbox.com/sh/1h4z0quoybo60c0/O1SwT35Nve
http://www.eirec.com/DPimages/digisqwvtest.jpg This is an example if transient response of different sampling rates.
Sampling theory says that a perfect square wave can be represented at any frequency below Nyquist. That doesn't mean that the codec or the analog electronics are capable of responding instantly at those frequencies, but that has nothing to do with the fact that a 1kHz square wave can be perfectly sampled with a 44.1kHz sample rate. The image is simply incorrect.
Transient response in the real world is generally limited only by the acoustic transducer response of the system, because everything else has the ability to respond much faster than audio rates. With Pono, this means that the earbuds or headphones you use with the player will have a greater affect on the transient response than the electronics inside.
It has been proven over and over via ABX testing that high res formats are completely indistinguishable from a 24b/44.1kHz master. And further studies have proven most audio engineers can't distinguish between lossless audio and 320kbps MP3. That's what the Xiph link elsewhere in the thread shows.