Apple: New MFi Headphone Initiative Puts Limit at 48 kHz
In another sign that Apple is not interested in high-resolution audio, they have introduced a new headphones specification that uses the Lightning port of your iPod/iPhone/iPad to output and input audio…digitally at 48 kHz. According to 9to5 Mac, a website dedicated to all things Apple, the new MFi (Made for iPod/iPhone/iPad) specification removes the requirement to use the traditional “mini jack” (the 3.5 mm stereo analog plug that has been in use for decades) and pushes audio bi-directionally to and from the headphones.
The standard output will carry stereo audio at 48 kHz and allow a microphone or other digital signal to travel the other way (a microphone or other transducer). I couldn’t find any reference to the word lengths in the piece by 9to5 Mac but they do point to a Wolfson DAC converter as a component that manufacturers will have to support (at least at the basic level of the connection…more below). The datasheet on the DAC lists audio performance at 106 dB SNT (A-weighted’), which translates to 18-bits…I found it curious that the datasheet didn’t actually state the word length but it does handle all “common audio sampling rates between 8kHz and 192kHz”.
So why is Apple limiting the resolution to 48 kHz when the chip that will reside inside new headphones will clock higher? Perhaps Apple is not so interested in high-resolution audio after all.
The connection to your headphones will switch from analog, using the 3.5 mm jack, to digital through the Lightning connection. This is a fundamental switch that dooes much more than offload the conversion of digital to analog signals from the phone. This opens up tremendous opportunities for headphones manufacturers…like Apple, now that they have acquired Beats. Despite the fact that Beats headphones are not highly regarded for sound quality, they do dominate the headphone market and upgrades can be expected.
For example, the new connection could enable remote control of volume, transfer of health related data (i.e. heart rate, hydration level) or other functionality through iOS. They haven’t activated the new capabilities but iOS 7.1 and higher devices will benefit from this update in the future. I also believe this would open the door to high-resolution audio headphone sets, provided the device has the right type of files.
There will be two different configurations of Lightning connected headphones. The Standard Lightning Headphones is the base package with the Wolfson chip mentioned earlier. The Advanced Lightning Headphones specification adds additional DSP processing to the equation. Think noise cancellation, surround processing, speaker or headphone modeling (make your Beats headphones sound like Stax…now we’re talking), upconversion and dynamics control. Once you get into the digital domain and pair it with a powerful DSP and some clever algorithm, you’ve moved way past the traditional model.
These are just some of the benefits of swapping a long-standing analog connector for a digital pathway. The connection also allows power from the Apple device to drive the headphones (if necessary) and also the other way around. As I have learned over the past couple of days in Cambridge, using my iPhone a lot drains the battery…I’ve resorted to using an external power charger. The new headphones could have high capacity batteries on board. The headphones would also be able to receive firmware upgrades if connected digitally.
This is a much bigger deal than virtually every report I read on the topic. It wraps the entire sound reproduction ecosystem under a single digital model that happens at the last possible point in the signal path. Meridian did it years ago with “digital” speakers and now Apple is leading the way with headphones. In this light, the Beats purchase makes more sense.
The bad news is that the Standard Lightning Headphones configuration, which will likely be the dominant flavor for a while, won’t handle better quality audio. Although, you wait and see…the 48 kHz number and 18-bit word lengths will have the labels and download sites claiming that everything at that spec is “high-resolution”.
Having been in the Apple cult for years, they’ll market 48 khz as better than CD, or “high” resolution. Then, after selling a gazillion more Beats headphones, they’ll open the bandwidth and sell “ultra” high resolution. The old Jobs m.o., market and milk every incremental step.
Android is far more popular than Apple mobile devices.
Not to mention, it is most likely we won’t see any devices that do not have standard interface (a.k.a USB port) in Europe starting 2016. Even if MFi is successful, the success will be limited to American customers with idevices, which are quite small compared to rest of mobile population.
But Apple has proven over and over again to be an innovator, which other companies has copied. If the MFi digital connection becomes a success…others will follow.
Yes, and none of these companies will use the Lightning port even if MFi actually succeeds (which I doubt about it). They will use usb ports… and nothing will really change the current situation other than the fact that people will have to carry a small headphone amplifier to use traditional headphones.
There are a lot of people who listen to music on the smartphone in very same way.
Here is a typical Head-Fi nut’s typical portable setup.
A smartphone -> USB OTG cable -> portable USB DAC/amp -> their choice of headphones.
Some extreme cases, a separate DAC and an amp is used, which make portable setup far less portable than my own desktop setup. For me, a small Sansa clip+ is more than enough for portable use.
The biggest difference is that the most current version of android system does not have restriction on bandwidth of the audio signal unlike apple portable devices (ALL of apple portable devices’ hardware components are restricted to 24/48, hence limit on 48). There is no limit on USB port on the android smartphones… you can listen to DSD256 and/or DXD with right hardware (like those from Light Harmonics GEEK) from your smartphone as a source.
Also, not just HTC M8, but Samsung’s Galaxy Note 3, S5 and LG G2 and G3 all support at least 24/192, straight from headphone jacks of the smartphones. All of android flagship phones are fully ready for 24bit music future. Both from their own headphone jack and USB out.
You should not worry, Mark. There are far more potentially harming issues for digital music such as complete lack of a centralized album art picture archive, for instance.
*Yes, I just finished rebuilding my music library. Finding missing album art… more importantly, finding high-res version of album art is even more absurdly hard.
**Now to think about it… Someone should really voice a concern about this lack of a place to find high-res album art files in general…..
106 dB A-weighted delivers 17-bit not 18
If I did my math correctly, 17 times 6 comes to 102 dB…that’s why I threw another bit in the mix. It doesn’t really matter…the point is that the new connection type will come up short of high-resolution. We need 24-bits.
Hi Mark,
This DAC must do at least 24-bit processing, so we are talking that it could deliver 10 dB of dynamic range above the potential of Redbook specifications. While I agree that 106 dB is not the full potential of 24 bits, no DAC device, I believe, can achieve the full potential of its chip. For comparison with some of the top performer devices, the Benchmark DAC2 HGC output has 126 dB SNR and the Oppo BD105 output has 116 dB SNR, both then not that close to what 24 bit allows (144 dB). I am not sure what the chips in the Benchmark are, but the ones in the Oppo are 32-bit chips and still the output is 116 dB, which is just over the potential of 19 bits.
I think the 106 dB SNR potential will make many HR Audio fans happy for playing recordings done at 24-bit word length. If Apple is going to max out its output to 48 kHz sampling rates, I hope Apple at least would allow access to the DAC by third-party apps so 96 kHz and maybe even 192 kHz sample rate files can be sent to the DAC without downsampling to 48 kHz.
By the way, I have not been able to find the specs of the DAC in the HTC One Harman-Kardom phone or the SNR of its audio output, does anyone have further info about this?
I’m trying to figure out how they are going to make this work. Files stored in iTunes are not necessarily limited to 48kHz. Are they going to downrez everything to 48kHz or less when transferring to the iDevice? Is the iDevice going to downrez during playback? Does the required headphone dac have the built in ability to down rez? (and if so, why?)
I have never been able to confirm just what I can playo in an iPod and this doesn’t help any.
The world of Apple and files coming from the major labels are derived from either CD or the original masters at 96 kHz/24-bits (if they labels are subscribing to the Mastered for iTunes initiative). Yes, you can use iTunes to do 96 kHz/24-bits but they don’t sell them.
Hi Mark,
As an audio signal processing architect with 7 US Patents, I feel that distributing the MIPS load across multi-processors (i.e. the ARM in the phone and another redundant DSP-enabled processor in the headphones) is not a benefit to the consumer. The NEON instructions in the phone can already do the job. BTW, why manipulate properly recorded HD audio anyway – what is the benefit ?
Jim…any additional processing would be for the personalization of the audio. Such as Smyth’s “Room Realiser” or speaker modelling. I know Beats and others are already developing this type of capability.
Hi Mark,
Despite portable audio getting much better, I don’t think High-Resolution audio playback will be portable any time soon. As you mention above, the 17 Bits worth of performance is rather poor, and even though the effective number of bits of the best DACs is still a tad better than 21Bits – like the Benchmark DAC2, the Weiss DAC202, the Resonnessence Labs Invicta/Mirus, or the very competitively priced Anedio D2 -, listening to HRA with roughly 16Bits of digital performance and most likely an inferior SNR and THD+N from the analogue output, makes it a rather feeble promise.
The typical portable setup mentioned above: smartphone, portable DAC/Amp and headphones, does get better, but still falls short of HRA. The best performance available now is the CEntrance hifi M8, that achieves a SNR of 113 dB (A-weighted), and a THD+N of 0.002 % (mid gain position) in the analogue domain, which is what you actually listen to. With the noise that surrounds you on the bus, train, street and the shortcomings of portable digital and analogue audio performance, HRA will still not be portable for a while.
I think that the HTC M8 phone could in best of cases come close to the performance of the CEntrance hifi M8, but it will not be up to delivering HRA at the level of some of the best DACs and Amps. We might have CD quality audio with our best portable gear, but that’s pretty much it. It’s of course not bad at all, and should provide a very satisfying listening experience, but definitely not HRA.
Cheers
The audio fidelity of portable players is very close to the quality of home servers and devices. In fact, many people use their Astell & Kern players to supply the music for the home system. The environment is a very important factor but more important than that is the lack of material that takes advantage of the new fidelity.
I totally agree with your on the huge lack of recorded material that takes advantage of HRA delivery containers like 24/88. 24/96, etc., but I also agre with John Siau (Benchamrk) regarding HRA playback:
“… an outboard DAC is only a partial solution to the High-Resolution Audio dilemma. A second key part of the problem is the performance of the audio power amplifier. A 24-bit audio system is useless if it passes through the typical power amplifier. It is nearly impossible to find power amplifiers that can deliver an SNR higher than about 102 dB. This is the equivalent of 17 bits (adequate for CD applications, but definitely not adequate for High-Resolution Audio). Anyone who thinks they can hear the difference between 16-bit and 24-bit digital audio through a “17-bit” power amplifier is fooling themselves.”
(http://benchmarkmedia.com/blogs/news/13174001-what-is-high-resolution-audio-part-1)
There is on the one hand, the HRA delivery container that is not being used in its entire potential, i.e., the misuse of the technology available to record High-Resolution audio degrades recordings below HRA (“the potential of delivering 20-20,000 Hz AND a dynamic range of 130 dB”). On the other hand, there is a bottleneck that doesn’t allow HRA to be played back, i.e., the analogue – and in many cases digital – components aren’t capable of delivering a HRA signal to the speakers.
You are also right that “the audio fidelity of portable players is very close to the quality of home servers and devices”, because that’s what the large majority of Amps out there deliver – as John Siau clearly points out -, but that does unfortunately not make portable audio any better or get it closer to HRA.
If we consider that HRA is a recording that has been captured during an original session using equipment capable of matching or exceeding the capabilities of human hearing, then we certainly must have playback equipment that can deliver that recording in its integrity. Otherwise, as John Siau well puts it: “If your playback system can’t resolve anything better than CD quality, then “High-Resolution Audio” will remain an illusion.”
The use of digital I/O and headphones opens a real possibility for binaural recordings too!
I totally agree, Carlos. I think binaural recordings have been largely overlooked, and I have to say that the very few ones that I have – mainly David Chesky’s – sound more natural and convincing to my ears than what I have heard through the Smyth realizer, and also the iX samples from AIX records.
Again, it is a question of personal taste and preference, but I believe binaural recordings have a huge potential. I really love the sense of space and overall feel of binaural recordings. Regardless of all the attempts at surround sound through headphones and development of Head Related Transfer Function by Harman, I still find binaural recordings more natural and convincing. I’m sure technology will only get better, but we’re definitely not there yet, and I still think binaural has a good chance with all the cans being sold out there.
Cheers
Binaural recording have a place among the various “sonic documentary” format available to recording engineers. Having received my doctorate in music composition based largely around using the Neumann Fritz “Head”, I can attest that it can produce amazing results. However,to compare the Smyth “Room Realiser” and Headphones[xi] to a binaural recording is like comparing apples and oranges…they are attempting to accomplish two distinctly different things. Binaural recording is trying to capture the sense of your head in a specific place and the others try to capture the sound of a specific room and set of speakers…very different things. Again, it’s personal taste. I personally find binaural recordings to be too distant and lacking in intimacy and intensity. There are not a lot of people doing binaural recordings these days. Chesky is an exception.
Mi Mark,
Yes you’re right, comparing binaural recordings to those using the Smyth Realizer technology for room simulation, is unfair as they respond to two completely different objetcives. Nevertheless, I think this is also in some way – and perhaps in a minor manner or just only from my paranoid perspective – a battle of “formats”, since Harman started to do their research with HRTF (Head Related Transfer Function) and developing target response curves for headphones.
This research is moving forward with clear objectives in terms of commercial applications, and the results could once again become the predominant way to listen to music and to standardize sound perception (like the compressed MP3 “sound” has become favoured by younger people and even selected as the HRA sample in blind tests, for example).
I find the research of Dr. Sean Olive et al very interesting, and certainly worth study and a read, but with the underlying premise or hypothesis being that: “since stereo recordings are optimized to sound good through loudspeakers in a room, they will only sound good through headphones that simulate the response of a loudspeaker system in a room”, I believe it narrows the headphone listerning experience down a bit too much. They also claim that: “This study provides empirical evidence that this premise is well grounded.”
The effect of binaural is not listening to speakers in a great room, but actually being in the room with the performing musicians. This is a completely different listening experience and model, and as you well point out, accomplishes a different objetcive.
I listened carefully to the AIX stereo samples and compared them to the Headphone Xi samples. I did several series of comparative listenings and I also had the headphones on for the entire Xi compilation twice in order to adapt to the paculiar Xi sound. I still have to say that the music I heard from the Xi samples didn’t manage to convince me in the sense of my perception being able to accomodate the desired room effect, and at times it sounded confusing and even slightly distorted or muffled in comparison to the corresponding stereo tracks. (I used my Violectric V100 Headphone Amp with a pair of AKG K702s, and a friend’s Benchmark DAC1 HDR as D/A converter).
One curious thing I noticed was that when switching to IEMs, the room effect got noticeably more convincing, and I also felt that the Nitty Gritty Surround album Xi album sounded more convincing and straight out of the iTunes samples. I don’t know why the IEMs sounded more convincing, but since the Pinnae of each Smyth Realizer owner are modelled, maybe this could be a factor. The difference I heard with the iTunes samples doesn’t account for this difference though, since Ithey sound equally – miore – convincing with headphones and IEMs.
Comparing the Xi effect with Binaural recordings is unfair from the perspective of Xi and Binaural being two different technologies and two approaches with different outcomes in mind. But since the premise of Harman’s research is that headphones should sound like listening to speakers in a room, I believe it is also fair to disagree with that hypothesis, and state that headphones should sound like being in the room or stage with the musicians. From the general perspective of how headphones “should” sound, I believe we can debate the merits and preference of both listening experiences without them being equivalent. Especially since the Harman model, pretty much in tune with what the Smyth realizer does, could result in products that will be a massive industry and thus turn into a predominant way of listening to music and mold our perception as iPods and MP3s did, and excluding the interesting and potent binaural sound headphone experience even more.
Anyhow, too long post, too many questions and open thoughts without full clarity yet, but at least enough for a discussion.
Cheers