That's a tough one. I'm not sure the track is very well suited to that exercise, since it has so many transitions of its own. I transferred it to CD and played it on my reference system, and I was unable to hear any significant changes. A time-resolved FFT of the data does not show much variation in bandwidth of the kind that is obvious when comparing lower bitrate music samples, so I the compression artifacts are not very apparent.
I have thought about doing it to a purely orchestral piece, but it is a bit time consuming and these days there is Foobar, which can take a source and a compressed version of whatever bitrate is of interest, and conducts a proper double blind trial alternating the sources in multiple instances which the listener chooses between and then spits out a nice report on whether the listener's choices were correct with statistical significance - or not. The user base of Hydrogenaudio.org have been conducting codec trials at varying bit rates for years and have amassed evidence from a lot of such double blind tests with a multitude of subjects.
I assure you the file is as represented and that it does contain sections derived from a compressed version of the source.
I have another file where I did the same thing, but recording the analog output of my iPod and the analog output of my CD player and mixing those together. I have gotten the gripe that my Mac's A/D converter isn't good enough to resolve the difference so I keep that one shelved.
I was not doubting your test - just observing that the variations must be rather subtle. It's easy (FFT or just listening) to tell a lower bitrate MP3 from a lossless version, but neither method showed up the variations in your sample. I'll try a couple of other techniques and see if I can pull anything out.
mieswall wrote: »
I agree that the difference with higher quality sampling rates is subtle, but to say that mp3 files actually sound better than lossless files is just incredible. Try a good chamber music and listen, for example, to plucked strings resounding in the auditorium; try Talk Talk's 'Spirit of Eden' highly dynamic range tunes; listen to Gentle Giant's 'In a Glass House' and try to identify and position in space the many overdubbed instruments. Do all that and just check how MP3 just kills that music.
My equipment isn't that expensive: the DAC/amp is about $300, close to the cheapest you can get. And the headphones cost about $100, $300, and $900 (which, btw, is not that much for somebody that really, really loves music). The differences among those three cans are simply enormous, and bear in mind that the three of them are considered very, very good in their respective price categories. Use one or another, and it is a different space or singer what you begin listening to.
What I did say is that a 24/96 audio file differences are subtle, compared with a very good 16/44 recording (still, the sound is more bodied, silkier, deep). The article, instead, says an MP3 can sound better than a lossless file. That's simply impossible. It is fully equivalent to say that a medium resolution monitor is better than a retina display, because colors look more saturated. Things with hifi audio are the same. Some kind of music/listener may benefit from compression (nothing too loud or too soft), or false exaggeration of bass, for example, but for most serious, artistic performances, subtlety, tonal accuracy, micro detail, spaciousness, etc, are absolutely key factors in the listening pleasure. There is huge difference, btw, between "passive" and "active" listening, between casual listening and instead to put your whole concentration (and emotional) capabilities in that. The latter, once you fully experience it, may well open a whole new world to you. Imho it is, in fact, almost a life-changing experience.
Happy listening, too.
The conditions of the test aren't very clear but if they compared the same file played back on both devices it makes no sense.
The real test would be comparing identical tracks purchased on both stores, because that would be real life use conditions. Still, I'm not sure there's enough of a market out there of people who care enough about the ultimate audio quality for this to be a successful business. It's not the first attempt at providing higher quality audio, and previous ones have failed miserably. Good luck, Pono.
Sony do one as well - http://www.sony-asia.com/product/nw-zx2
I have no trouble admitting that I don't know how to listen any more than someone else doesn't know how to see.
"How to Listen is a desktop software application developed by the Harman International R&D group for the purpose of training and selecting listeners used in audio product research, development, and testing. The software consists of a number of training exercises where different kinds of timbral, spatial and dynamic distortions commonly found within the recording and audio chains are simulated and added to music.
Wow, amazing how backwards computer people get audio. "Common sense" seems to be getting worse.
Of course ponoplayer sounds better than iPhone, or you are deaf and/or listening to horribly produced music.
PP has a far better DAC than Apple uses.
PP has discreet signal chain from DAC to analog phase.
PP has no radios or sensors on the board introducing noise and distortion.
PP can drive line-out and balanced setups.
Therefore, understanding basic signal chain, PP is going to play even MP3 files better. It sounds amazing playing 16/44 FLACS, and perhaps is the best digital I've ever heard when playing 24bit files.
Apple spends about $4 on the audio components in an iPhone. You are telling me that $4 sounds the same as $400? Sure it does. Deaf leading the dumb.
Blind tests don't work for audio. Neither does that horrible radio shack switcher this guy used. Seriously? Plus, he used modern music, crushed to hell and back, and the guy is old enough for us to know he doesn't actually listen to that stuff.
Check out the Leo Leporte review for a good mainstream look:
and check out my review of the PP plugged into all manner of regular gear here:
Finally, don't listen to mathematicians about music and don't listen to physicists about how your ears work. Listen to your own music, the musicians that make it, and your own ears. Stop following the herd, even the late Steve Jobs couldn't believe how many people accepted the substandard audio quality of the iTunes/iPod world.
I'm going to post this and back away slowly.
http://www.stereophile.com/content/ive-heard-future-streaming-meridians-mqa (contains actual details on how MQA works)
I also use the excellent dragonfly DAC in my case with audioengine 5+ speakers. I have found that arguing quality sound is pointless with people that don't know what to listen for. I hand them a pair of grado ps1 headphones to plug into there idevice and they tell me they are crap cause they hear all these other sounds in there music. Better fidelity always reveals the flaws in poor quality recordings and devices while also revealing the true quality of better sources. Junk equipment and files are more than adequate for most clueless consumers. If you care about your music and take the time to truly listen to your material a quality source is magic. Its a lot like fine wine, only appreciated by educated people who taste the nuances, while lost on everyone else. So, if you are happy with your awesome apple earbuds and lower fidelity sources enjoy!
So according to you, both blind listining test and equipment measurements test doesn't work for quantifying music? This is exactly what the Emperor get no cloths story is all about.
Do you know which DAC the Pono and the iPhone uses before claiming a winner?
Do you know the iPhone and the Pono electronic schematic before claiming the Pono is better?
Besides, Leo Leporte reviews is far from objective, He is to happy too show of is Pono "signature" edition given to him to say anything bad about it.
It's amazing to see how much "audiophile" people get scammed by snake oil corp like Monster Cable and Beat
BTW here is a great iphone 5 audio quality analysis with factual results made with proper equipments: http://www.kenrockwell.com/apple/iphone-5/audio-quality.htm
It's triangular because they used cheap as can be electrolytic capacitors in the back.
Not worth the $399. If the iPod Classic were still around I'd readily recommend it over this. Even then...buy a used Classic, do some internal mods, and you'll have a better player for less.
Thanks for posting the teardown.
The best iPod click wheel was the 4th gen / photo - it was the last that used the good Wolfson DAC in it. But the 5th gen and laters are easier to modify to flash storage.
What mods would you do?
mstone wrote: »
16 bit vs. 24 bit is not an argument that relates to dog vs. human hearing, because everything that is within human hearing is also improved due to the smaller steps.
<img alt="" class="lightbox-enabled" data-id="54996" data-type="61" src="http://forums.appleinsider.com/content/type/61/id/54996/width/350/height/700/flags/LL" style="; width: 350px; height: 213px">
Bit rate and sample rate are two different things. In standard RedBook CD recording, the bit rate is 16 bits. The sample rate is 44.1KHz.
The sample rate is how many times per second the analog waveform is sampled ("reading taken"). A sample rate of 44.1KHz can provide a frequency response up to 22KHz. Most adults can't hear much over 15KHz. Most LPs, which are all the rage now because of their supposedly superior quality, are filtered above 16KHz. Personally, I'd avoid any LP longer than 20 to 22 minutes per side, because it's going to be level limited and frequency restricted in order to fit that length. And most CD releases today are more in the 60 to 70 minute range, so they're really not made with LP in mind.
Bit rate determines how many different levels you can have. 16 bit recording provides 65,535 different levels. As I wrote in another post, that's far more resolution then a sound mixer can ever hope to achieve by moving a fader. If you have a level that's in-between two of those 65,535 levels, the A/D converter averages the signal up or down to the nearest whole amount. That's called quantization error.
Most high resolution recording is performed at a sample rate of 96KHz and a bit rate of 24 bits. That enables a frequency response up to 48KHz and 1.67 million different levels. Obviously, with 1.67 million different levels, there's far less quantization error. But that doesn't mean you can hear any difference.
More isn't always better. If you're trying to play back the frequencies from 22KHz to 48KHz that you can't hear anyway, your amp is consuming far more power and the speaker cones are heating up. This all can limit the quality of the sound that you hear in order to attempt to reproduce frequencies that you can't hear anyway. Any portable player is an iffy situation because they're putting out less than a quarter of a watt. It's amazing they sound as great as they do.
So on the one hand, all this high-resolution stuff may be useless. On the other, I was listening to the Sony high-resolution player at a trade show and it sounded pretty damned good to me, but the improvement may have come from better electronics and higher quality headphones. A standard RedBook CD might have sounded just as good on that system, but without such a test, there's no way to know for sure. I'm certainly not opposed to better sound if it's real.
There is plenty of misinformation posted. Usually by those who simply don't understand exactly how digital recording and playback works and they frequently don't understand analog recording either. So instead, they basically have formed a religion to preach certain mantras on faith. So they think higher bit rates and sampling rates must be better because more is better and they want their amp to go to "11". These are the same people who used to think that marking the edge of a CD with a felt tip marker could improve the sound. All that's on a CD are pits and lands representing 1's and 0's. Those pits and lands either get read or they don't. You can't make them better (or worse). The quality difference between players is mostly the differences in the filtering circuitry and the quality of the playback preamps.
One of the biggest problems in all electronics, especially over time, is leaky capacitors. Posters are making fun of the triangular shape of the Pono player, but someone posted a photo of the interior and the shape was to make room for large capacitors. Someone said those capacitors were junk, but I don't know how they could tell that. The Pono player may indeed have better playback circuitry. Or not.
If it's truly better, I hope it succeeds. If it's not, I hope it fails.
I'm a lossless guy too.
Years ago I encoded music at the standard settings for iTunes and was dissatisfied. I just kept increasing the bit rate until I arrived at the Apple Lossless setting and I've never looked back.
It didn't even take top rated headphones and speakers. I can easily tell the difference (double blind) with my 30+ year old B&W DM-10's or on my equally vintage SenheiserHD420's. These are good quality equipment, much nicer than similarly priced stuff today (IMO,) but they were never seen as more than decent inexpensive equipment in their day.
As for what people prefer? Preferences have nothing to do with quality and everything to do with familiarity. What are most people familiar with today? — music that has been crappily transcoded, over-compressed, and played on sh1tty speakers or headphones. Even naive listeners can learn to tell the difference, it's just that most people have become more comfortable with distorted sound than accurate sound.
BTW: The big difference probably isn't in the player. It's in A.) the original transcoding and B.) the speaker or headphone used to listen.
I think you are confusing several different concepts, at least by name. Bit rate is bits per second, not bits per sample. 16 bit describes bits per sample, and is bit depth, not bit rate. Bit rate is samples per second multiplied by bits per sample. Your description of sample rate is correct, and it is often quoted as a frequency (in kHz) although it is technically more accurate to use samples per second.