Was the Pono Predestined to Fail?

The Pono player was a much talked about portable digital music player that was released in late 2014; with its parent company closing down three years later at the end of 2017.  A fascinating history of the company and the evolution, success, and failure of its product can be read in the book “To Feel the Music”.  We review “To Feel the Music” here.

The Pono received a lot of favorable publicity for two reasons.  The first reason was that the entire undertaking was a labor of love by the well-known rock musician, Neil Young, who used his bully-pulpit as one of the megastars of the late 1960s and 1970s rock scene to proclaim that modern music recording/storage/playback processes were “losing” a lot of the musicality and emotion of the music they were handling.  On the face of it, surely a musician is best qualified to comment on music quality, and his related contention that we as consumers were somehow being cheated of better quality music by greedy recording studios had a visceral appeal to it.

The second reason, no doubt linked to the first, was that the company raised money for the development of their Pono player via a Kickstarter campaign in early 2014.  Their campaign was spectacularly successful.  They had published a fundraising goal of $800,000, and ended the 40 day campaign having raised an extraordinary $6.2 million, making it the second most successful Kickstarter campaign to that point, eclipsed only by a $10 million campaign in 2012 for the Pebble watch.  (Pebble went on to set another record on Kickstarter in 2015 when it raised a further $20.3 million for a new generation of watches, only to then close down completely in late 2016.)

One wonders why Pono, when money became tight, and with a desire to develop a second model Pono with added capabilities, didn’t also return to Kickstarter to fund its further existence.  A former Pono executive suggested that by the time they considered such a move, the company was already in dire straits, and the money raised from any Kickstarter campaign would have needed to be used to pay creditors rather than to fund the development of a new product, which would have been dishonest and wrong.  One has to admire their ethical standard, something clearly demonstrated throughout the company’s brief life.

The extraordinary success of Pono’s Kickstarter campaign suggested a market receptive to Young’s messaging, and caused a lot of positive excitement and anticipation for the release of the Pono, which finally started its deliveries at the end of 2014, more or less in line with the promised timeframe during their Kickstarter campaign.

So it seemed the stars were aligned in a positive pattern.  What went wrong, and why?

We feel their mission was Quixotic and unlike to succeed for two main reasons, and assorted other reasons too.  While it is easy to make this analysis with the benefit of hindsight, we did attempt to share some of our misgivings at the time the Pono was nearing release with them, but they were too engrossed in their Koolaid to listen to alternate viewpoints, and the fundamental problems that caused Pono to founder remain timely and contentious, even today.

The Right Product at the Wrong Time

The first reason is perhaps the less contentious of the two.  The company decided to develop a high-end portable digital music player at a time when the market was shifting from the concept of buying/owning music – either on CD or track by track purchased from sources such as iTunes, and instead starting to embrace the concept of paying flat monthly subscriptions to access a large library of online music that could be streamed via your playback device, which as likely as not could be a regular computer or phone.  It no longer needed to be a specialized device.

The smart phone was dooming dedicated music players the same way it was killing off dedicated notepad type devices (who even remembers the Palm Pilot these days) and marginalizing dedicated cameras.

The Pono concept not only had an unpopular hardware item (a relatively expensive “high end” music player in a market that was dominated by less expensive devices that the market did not perceive as lower quality) but also had a slightly confusing approach that seemed to be sort of like streaming, but was not real streaming (ie with no local storage of tracks), at a time when the growing focus was on medium quality streamed music.

A bit like how the Edsel started life as an excellent concept, but by the time it finally made the market, the market had changed and Ford’s much-vaunted “signature” car failed, we feel that if the Pono player had instantly appeared back in 2010 when Young was first considering the concept, it may have been a much greater success than its eventual release at the end of 2014 allowed it to be.

To be fair, it was not only Pono that suffered at the hands of a shift to streaming.  Apple’s iPods were in a period of serious decline, too.  With Apple’s iPod in terminal decline, what chance did a new entrant have?  And, to be clear, the Pono was a high quality version of a digital music player.

A Flawed Underlying Premise and Promise

The second reason is much more contentious.  The most fundamental underlying concept of the Pono player was flawed.  The Pono concept, in a nutshell, not only attacked MP3 music as inferior (which it is) but went further and claimed that CDs were not a high quality music source either.  Pono offered access to what was claimed to be a better-than-CD music source.

The kindest analysis of this is to say that, unlike video where you can actually see the difference between good and bad video quality and therefore be somewhat encouraged to spend more to get a tangible improvement in viewing experience, 99% of all people listening to a Pono could not and can not hear any sound difference at all between the Pono and most other portable music players, and those who did appear to detect a difference were as likely to prefer the other product as they were the Pono.

We’ve no doubt that if there were a clear difference that could readily be heard between Pono music and regular music, the Pono would have sold much more strongly.  Video is undeniably a positive tale of ever better quality, and its improvements are clearly describable and visible to even the most “untrained” eye, making the appreciation and sale of DVDs, then Blu-ray, of 330 line, 480 line, 720×1280 pixel, 1080×1920 pixel, and now even 2160×3840 pixel based monitors easy to justify.  But there isn’t any perceptible (or real) difference in sound between what a Pono provides and what a CD provides, and that makes the underlying “benefit” of a Pono player (better quality sound) very hard to successfully sell.

Neil Young started off his crusade with excellent fundamentals, but he got carried away and went too far.

We too have certainly observed the curious evolution of music, and how lower quality music formats and listening experiences have displaced higher quality formats (we write about the lowering of music recording quality here and elsewhere).  Most other digital and computer based products have steadily improved, but the extraordinary reality is that music reached its zenith with the CD (first released in 1982), and since then, new digital formats such as MP3 and distribution via streaming rather than CD have jointly robbed stored music of much of the quality it could have.  So we agree with Neil Young’s cry from the heart about the debasement of music quality.

But when Neil goes beyond complaining that MP3 and AAC type compressed music streams are inferior to CD, and goes on to suggest that CDs are also in turn inferior to some other form of better music recording/storage/playback, that’s where we – and all industry experts who don’t have a vested interest in promoting products that actually convey no extra sound quality – totally diverge and disagree.

We hasten to add that we have totally no doubt whatsoever that – unlike many in the  “high quality” music game, Neil is totally and utterly sincere in his beliefs.  But being sincere doesn’t make you right.  Our article about the difficulty in truly establishing if you are hearing better, worse, or the same sound helps explain some of the problems associated with this.

The Facts About CD Quality

The facts are simple.  The human ear can hear sounds from about as low as 20 cycles per second (Hertz) up to about 20,000 Hertz.  As you get older, the upper limit of the sounds you can hear steadily drops (there are a couple of links to tests in the article we mentioned immediately above so you can test your own upper frequency hearing, if you are interested.)

Even if we can hear all the way to 20kHz, most of the sounds we want to hear don’t go that high.  Most of the sounds we hear are comprised of a base frequency somewhere under about 2,000 Hz, and then some “harmonics” above that base frequency.  A concert grand piano has its lowest note at 27.5 Hz and its highest note at 4186 Hz.  A soprano singing a “High C” is singing at 1024 Hz.

Now, you know that a piano sounds different to a violin and different to a trumpet, and so on.  This is because as well as the actual note being played, the instrument generates a mix of higher frequencies as well – perhaps some twice the fundamental note, some three times, some four times, and so on.  Each of these higher parts of the sound tend to be successively quieter than the main fundamental frequency sound, and so become of less and less importance.

Very few of these harmonic sounds actually go much over about 12 kHz (12,000 Hz).  And almost none ever go over 20kHz, and of course, those that do go over that point can’t be heard by nearly every adult in the world.  Not only that, but microphones, recording equipment, and playback equipment and speakers get weaker and weaker at recording, storing, and playing back such high frequencies too.

A compact disk perfectly records and stores all sounds up to just over 20 kHz.  That’s not an accident.  It was designed, free of the constraints of previous analog recording methods, to be a perfect, free from any compromise, method of storing music in the same quality exactly as it was played.  If the people designing CD’s in the late 1970s and early 1980s had seen any value in allowing for higher frequencies beyond 22.05 kHz, they’d have tweaked the specification to include them.  But, at the time when they were free to settle on any quality at all, they decided that 10% higher than the generally accepted “best case” highest upper limit of hearing was more than overkill already.

They also had a second design choice to make.  How much “dynamic range” did they want to allow the CD to hold – that is, how much louder did they want to be able to record the loudest bits, compared to the quietest bits.  One of the big limitations of records and tapes back then was they had a limited dynamic range.  Tape hiss and record surface noise set a minimum volume that had to be exceeded to avoid being drowned out in the background/surface noise, and the physical limitations of groove size and how far and fast the playback needle could move limited a record’s loudest recording, and the electrical ability of tape to store more charge limited the ability of tape in a similar fashion.  A typical dynamic range for records and tapes was in the order of 40 – 50 dB; and even the very best “demonstration quality” recordings probably never exceeded 60 dB.

This was adequate for most types of music, but was not excellent.

The CD designers decided to go “wild and crazy” and increase the dynamic range by more than 1000 times, ie, more than 30 dB.  A CD would have a dynamic range somewhere in excess of 90 dB – a theoretical limit of 96 dB, and depending on various techniques, an actual and perceived range that might be slightly higher or lower.  Whatever the actual number is, it is theoretical, because there’s nothing that would make sense to be recorded with that much dynamic range.

For example, a library has a background sound level of about 44 dB – a mixture of air conditioning sounds, pages rustling, books being moved, people walking, and so on.  If you played music in a library, and set the minimum volume as say 10 dB over the background noise (54 dB) so as to clearly hear the quiet sections, then any maximum level sections would be at 150 dB.  How loud is that?  Loud enough to instantly burst your ear drums and cause permanent hearing damage.  A military jet aircraft, taking off from an aircraft carrier with full afterburners, heard from 50 ft away is about 130 dB.  150 dB is 100 times louder.

So, the CD created a sound that had all the frequencies we could ever hear, and more besides, and a wider range of volumes from quiet to loud than we’d ever need.  As we said, it was designed to be better than the best, and a specification that would never need to be improved upon.

Could you hear the difference in quality between an analog recorded/stored/played back piece of music and one which had been digitally recorded, stored in CD format, and played back?  Yes, you could.  No doubt about it.  And by every possible acoustic standard, the CD version was not only “better” but it was also “truer” to the music as it came out of the instruments and singers on stage.

Most people rejoiced and celebrated at the new purity of sound from CDs.  But a curious thing happened.  Some people, who had become so used to the sound limitations of analog recording, decided they preferred to listen to a scratchy hissy recording, with reduced sound dynamics and more distortion.  They struggled to justify this, other than claiming that because the music was analog all the way through the process, it was therefore “purer”, whereas something was “lost” when the music was converted to digital and then converted back to analog again.

Why Do Some People Believe They Can Hear a Difference?

There are really two cases, here.  The first case is whether there is a difference between an lp record and a CD.  Yes, there truly is a difference, and it is relatively easy to tell the two apart, particularly – of course – once the record gets some scratches and dust damage on it.  As to why people think the lp version is better, there are two answers to that.  The first is they prefer the lp because it is what they are expecting to think, and because it is what they have become used to, and so their unconscious bias (as we discuss in this article) causes them to confirm that which they already want to believe.

The other answer is slightly more complex.  The background noise on an lp record is sometimes thought to make the sound more “warm” and more “live”.  Some people describe CD’s as “too cold”.  Maybe even the greater distortion on an lp contributes to its “warmth” and the stunning clarity of a CD contributes to its “coldness”.  So from a psycho-acoustic perspective, some people prefer lp sound, even though it is less clean and less pure and less accurate than the CD sound.

The second case however relates to why do some people claim that a “high definition” digital recording (ie 96/24 or 192/24) sounds better than a CD digitized at 44.1/16 settings?  We don’t believe there is any difference at all – in theory.  Indeed, that theory can be and has been proven.  Take a 192/24 digital file, down-sample it to 44.1/16, then upsample it back up to 192/24.  Compare the files.  Chances are you’ll see them to be exactly the same.  In other words, nothing was lost when the file was downsampled.

But, note we said that there is no difference between a 192/24 and 44/16 recording in theory.  In practice, it is almost never that the two music files have everything else identically in common, with just the sampling rate being different.  It is quite likely that the two recordings were done on different machines, and perhaps with different settings, and maybe even one version was taken from one set of masters and mixings, and the other from a different set of masters and mixings.

So even if the digital quality is identical, other parts of the total process between the musician playing and the sound emerging out of your speakers or headphones are not identical.  Maybe even, in the last step to your music output device (headphones or speakers) one sound source is infinitesimally louder or quieter than the other, and generally, people associate louder music with better quality.

All we can say is that when absolutely laboratory style clinical double blind A/B testing is done (and we describe how you can fairly easily do this yourself, here), the listeners fail to come up with any level of statistically significant preference for the “better” version, always assuming they can tell the difference at all.

Correcting a Common Misperception About Digitizing Music

The suggestion that something had been “lost”, while vague, had the benefit of being intuitively acceptable.  Worse, some people who either didn’t understand or deliberately chose to misunderstand how the digitizing process worked proceeded to come up with “explanations” why digitally recorded music would be worse than analog music.  Here’s an example, taken from the “To Feel the Music” book, which seems to be correct but in truth is totally wrong.  They say, on pages 16/17, under the heading “Sampling Rate”

… digital audio is “manufactured” from the original performance. The original recording is chopped up into pieces of data (sampled) to represent the recording. How often it takes these samples and how precise the samples are, affects the sound quality. The more frequent the sampling, the higher the quality; the less frequent, the poorer the representation of the original recording.

….

Sampling music is much like sampling a moving object to create a video or movie; the higher the sampling rate the better the representation. Images in old-fashioned movies flickered and were jumpy because the sampling rate was just a few frames per second (fps). With digital cinema, HD video, and Blu-ray, sampling is up to 60 fps, resulting in a smoother and much higher quality representation of the original.

This illustration is attached to show what they are describing.

As you can see on the illustration, it suggests that the nice smooth sine wave shown in its pure analog form is made into a rough series of steps to approximate the same wave form when digitized.  Obviously, the more the steps, the closer to the original wave form, and the fewer, the more inaccurate the digital wave form is.

This all makes easily understood sense.  Unfortunately, it is also totally and completely wrong.  That is not how a CD digitizes music.

Think again of the sine wave.  Sure, you can store/recreate it with a complicated series of digital “steps”, but you can also do so mathematically with a simple very short formula that describes the frequency (length) and volume (height) of the sine wave.  You just need to know these two pieces of information, and you have digitally captured the entire sine wave in a single equation, and can recreate it in perfect identical smooth form.

It gets more complicated for a musical sound which has many different frequencies all added together, but you still end up with a combined sine wave, and it’s entire analog complexity can still be reduced to a simple mathematical formula through a process that is called the Nyquist-Shannon sampling theorem.  Click the link for more explanation, and you’ll quickly see how complicated the explanation and process is, and why its complexity makes it easy for people to misunderstand, and then mistakenly assume a totally different digitizing process is used.

It is unfortunate that a company that is based on an assumption that there’s a problem with CDs didn’t better understand the technology right from the get-go.

The effectiveness of the Nyquist-Shannon digitizing process is not even controversial.  It can easily be proven, which is totally unlike the claims of people who say they have come up with something better.

Simply take any analog recording of music, and then digitize it and then un-digitize it back to analog again, and compare the waveforms of the music from the original analog recording and the version that has been converted to digital and back to analog again.  Are they the same or different?  If they are essentially the same, then case proved.  If they are significantly different, case proved the other way.

As far as we are aware, this testing has never revealed any differences of any significance.  Yes, we’re not saying the two versions would be 100.000% identical, because the simple act of feeding the signal through wires to the recorder, and back through wires to the decoder, and so on and so forth, introduce small changes to the analog signal due to the capacitance (and lesserly the inductance and resistance) of the wires, and the nonlinear impacts of the electronic circuitry.  But if you also took the same signal, then rerecorded it in some sort of analog format, then played that back, you’d again see waveform differences, and of equal or perhaps much greater nature than the digital “roundtrip” would create.  In other words, any changes in waveform are not so much because of the digitization but because of the general electronic process.  The least impactful part of the entire process is the digitizing and decoding.

This is the first huge fallacy that Neil Young so enthusiastically espouses.  He says that compact disks only provide “25% of the musicality”.  25% of what, exactly?  We can’t start to guess how we arrives at this number, or what research it is based on.

Is There Something Better than CD Recording?

He then goes on to advocate for a “better” type of digital encoding.  It would have two or even four times more sampling and higher frequency recording capability (ie 96 kHz or 192 kHz sampling, with maximum frequencies being half those numbers), and instead of a 96 dB maximum dynamic range (16 bit), it would have a 144 dB dynamic range (24 bit).

Oh yes, remember when we calculated that a 96 dB dynamic range would burst your eardrums at maximum volume?  Guess what a 144 dB dynamic range could do?  Well, it is hard to say for sure, but this table suggests a 180dB sound level would be like having one pound of TNT detonated 15 ft from where you were standing.  And, yes, that does mean you’d suffer not just hearing loss but physical harm – the same table says that 183 dB creates a 6 psi pressure wave, strong enough to totally destroy buildings.

What is the point of recording sounds we can not hear?  None, whatsoever.  Indeed, not only would this process be recording sounds we could not hear, but there’d be no sounds emitted by the instruments at those frequencies to record in the first place!  We discuss the zero benefit in higher music sampling rates in this article, and urge you to try it for yourself and come to your own conclusions.

What is the value in having a dynamic range which, if fully used, would destroy our house from the sound pressure waves so created?  That’s not a benefit.  That’s a serious danger!  We already have problems with people playing low-dynamic range music too loudly and destroying their hearing as a result.  What would happen if they could go even louder and louder?  Besides which, just like no musical instruments go up to the high frequencies Neil wants to record, neither do any normal or conceivable concert performances have dynamic ranges even half of the “standard” 96 dB, let alone an entire 144 dB.

So this was the foundation of and justification for the Pono player.  A player that would store sounds that didn’t exist and which we couldn’t hear, and which could handle sounds so loud that they could kill us.  As for ordinary normal music, it would play that too, but in no measurable/relevant way better than a CD would.

The Pono Promise Meets the Marketplace

Sure, there were dozens of claims that people could hear or in some more spiritual way “sense” the better musicality of the Pono player, but where were the double blind scientific studies proving that?  Where was the “Pepsi Challenge” for people to listen and hear the difference themselves?  Apart from the confirmation bias and suggestibility that many people seemed to suffer from, there wasn’t a single shred of evidence to support their claims, either the subjective ones or the “scientific” ones about their music being so many times “better” than CDs.

If there was a clear and obvious difference, like in higher resolution television sets, presumably the Pono would have sold well.  Although exact numbers were never released, our sense is that most of the entire three years of Pono sales happened as a result of their Kickstarter funding, which resulted in people pre-buying, sight-unseen and most definitely sound-unheard, 15,000 Ponos.  When these were all shipped, did they seed the market, and did an army of 15,000 new converts to the “better music” cause start shouting the praises of the Pono, leading to gradually accelerating product sales?  Apparently not.

The failure to gain traction in the market seems to point to a failure of the underlying claim that the Pono allowed for a genuinely better musical experience than other musical storage/distribution/playback systems and services.

Pono also suffered from another problem.  Its same capabilities, in terms of playing 96/24 and 192/24 FLAC files, were offered in any number of other high-end music players on the market; some costing very much more, but others costing very much less.  I was advocating Fiio units in August 2014 prior to the Pono’s release, and in October 2014 (still prior to the Pono release) was writing about a new Fiio unit that cost $100, compared to the Pono at $400.  These days, there are Fiio units for as little as the M3K, available for $60 on Amazon, and plenty of other branded products at sub-$100 prices too.

The Pono player was simply too expensive.

There was another problem, too.  Other digital audio players are “standalone” players – the owner loads them up with music from whatever source they wish, however they wish.  Pono decided to “add value” by promoting their own on-line store of “higher quality” music as a key part of the Pono product offering.  Sure, people could still load other music files onto their Pono any way they liked, but a key part of the Pono concept was “we have the best music for your player, and you can’t trust other music as being similarly good”.

Through no fault of Pono’s, the company they’d contracted with to provide the digital music was bought by Apple and then summarily closed down.  Pono suddenly found itself without the music half their product offering.  It also seems they’d been losing money for some time, perhaps on both their hardware sales and music sales, so when they lost the music side of their business and couldn’t find a fast and satisfactory way to recreate it, and with the hardware sales alone being insufficient to keep them afloat, the end was unavoidable.

And So?

So, the lessons of Pono?  Perhaps Neil Young’s contention that fans would appreciate “better” music was proven wrong, as also was his bemoaning that “high quality” music could/should be sold for low prices.

The good news is that CD quality music is as brilliantly close to the real thing as ever will be possible, there’s a vast library of well-mastered CDs to choose from, and wonderfully low priced portable music players and memory cards that allow you to now have a library of as much CD quality music as would fit on, perhaps, 500 – 1000 CDs, all for little more than $100 (ie the cost of a player and a 256GB – 512GB memory card).

Neil Young moved on to a related project, another “proof of concept”, this time placing his entire personal archive of music online, available in multiple high quality formats, and offering unlimited streaming for $1.99/month or $20/year.  This was to prove to the studios and other musicians that there was no need to charge a premium for higher quality music streaming.

Just a couple of days ago, this new venture issued an announcement.  There is to be “a small bump” in the cost of this streaming service.  The small bump is actually a 50% rise – the service will now cost $2.99/month or $30/year.

We’re not going to second guess Young again, but we will observe with concern the implications of what he is doing.  Imagine if the only way to access music was to subscribe to every separate artist’s own personal streaming service.  $30 here and $30 there; add it all up, and soon you’re talking hundreds of dollars a year for little more than a handful of artists.  We’ll stick to buying CDs and then ripping them to our digital audio player.  Much cheaper and easier.

And if you do want high quality streaming, Amazon has now announced a high quality streaming service that provides access to over 50 million music tracks, all for $13/$15 a month (Prime/nonPrime rate).