Hacker News Reader

Comment by adolph | Hacker News Reader

As a counter point, Apple is bringing computation closer to ears. The Sennheiser 650 is an analog cabled experience and will accurately reflect the upstream digital to analog conversion, typically made without regard to the destination device. If Apple can use on-ear computation to make a plastic driver cone sound like a more expensive material, why would you care about the materials used?

https://www.kenrockwell.com/audio/sennheiser/hd-650.htm

[edit]

Think for a moment about what is possible in a device as large as the Airpods Max. It could fit the guts of an Apple Watch, which can already stream audio via cell modem without a host computer (as long as you are using the Apple services).

https://support.apple.com/en-us/HT204691

View on HN

There is simply absolutely no need for on ear computation to make low quality drivers sound better.

In ear headphones have evolved beyond plastic cones for a long time. At that price, you're competing against the very best balanced armature drivers, which basically have perfect frequency responses and harmonics, as well as against planar magnetic and the very best dynamic drivers too.

As for computing audio, yeah that's simply useless. Just have the DSP at the device. The device, iPhone, or Mac, or a PC, or whatever, is aware of which headphones it is connected to, and can trivially run the correction routines itself. In fact, this is already done by many.

In ear headphones aren't going to have enough battery to power a cellular modem.

Is this still true when active noise cancellation is involved?

No, it's not. If using ANC, you do need a chip on the device to do ANC for latency reasons. But that's already a solved problem.

Sorry, but when the DAC can be tuned to the specific hardware, and the engineers know what the hell they are doing, the result can be pretty damn impressive.

This is the reason modern active bookshelf speakers like the Klipsch The Fives punch way above their weight and pull off stuff its hardware shouldn’t be capable of.

Apple can do the same thing, they have the engineers and know-how. And on top of that, the positional audio is something that requires compute power anyway.

I never argued that point. I just said that it makes no sense to make the vast majority of that processing in the headphone when your phone and computer already have specialized hardware to do it.

Also, DACs today are pretty much transparent.

For example, for positional audio and driver/amp compensation, people already do the same as Klipsch does in software on their computer. You just need a good enough microphone and then you can do it by your own using REW and EqualizerAPO, for example.

Sounds absolutely horrible. I'd rather not install some bloatware crap headphone software that injects itself into my operating system's audio stack and introduces extra latency on every device just to get the headphones to sound right.

You already have to install software that inserts itself into the audio stack and introduces extra latency to be able to use your wireless headphones. Those are called "Bluetooth drivers".

In any case, I run processing on my computer to apply a convolution filter on my audio stack with adds around 2.3ms of latency and uses 0.15% of my CPU. Compared to the ~200ms of Bluetooth latency it's completely unnoticeable, and I'm sure Apple can figure this out better than I can.

Hardly the same - they are generic, provided by the OS, follow a standard, don't require specific support to be provided for individual devices.

They do need some pretty wide support per individual device. A proper stack requires something like 6 different codecs, all of which are quite heavy and very different.

In any case, the transformations are not generic at all. You basically need to do one convolution, that's it. The headphones can provide the impulse response with which to convolve at pairing.

For things from Apple, of course, this is trivial.

Yes, basically Bluetooth is already almost too complicated to do well, and you would rather it were more complicated still? And like you say applying a convolution is computationally trivial especially compared to decoding a lossy audio codec and running a Bluetooth stack and antenna, which the headphones are already doing. Offloading this to the device is going to make absolutely no difference to battery life while increasing complexity and unreliability, while also restricting what processing can be done to static convolution with an impulse response. There is no reason to do this on the device.

It certainly would make a difference to battery life. Bluetooth connections as well as decoding are done in hardware using commodity chips that can't do much else. Adding additional audio processing hardware will increase the complexity of these chips which translates to higher prices and lower battery efficiency.

Remember, some of those devices have 20mA of battery. The codecs already have to be made easy to decode.

There is also absolutely no need to limiting processing to a static convolution with an impulse response. That's just the only device-specific processing you have to do. Headphones are minimum phase devices, so except for things like distortion, they can basically be described by their single impulse response.

For the rest, like spatial audio, or EQ, or anything of the sort, there is no need to do it per headphone, it's the same for all of them.

There is simply absolutely no need for on ear computation to make low quality drivers sound better.

Sure there is. If the cost of computation for a required sound quality is lower than the difference in driver cost, then computation will win, especially when it adds additional value unavailable to the non-computational approach. Positioning the computation at the device is advantageous since it allows for a simpler data interface. For example, adjustments to the soundstage based on head movement can be performed on-device instead of making a round trip to source for the adjustment.

In ear headphones aren't going to have enough battery to power a cellular modem.

Sure, and when the first iPod was released it lacked wireless networking and was storage constrained compared to comparably priced products. I can appreciate the pragmatism and be appreciate folks working towards an exciting future.

Pardon me for departing, my kid wants to watch the SN8 livestream together; hope you enjoy some high quality Soyuz laserdisks.

https://slashdot.org/story/01/10/23/1816257/Apple-releases-i...

Read carefully. I said there is no need for on ear computation. I personally even have impulse response profiles for my headphones that I load onto my phone and computer to make them sound much better.

For things like on ear sound stage correction, sure. That can be done. You will pay a heavy cost in sound quality, but it can be done. If your goal is sound quality, you're not going to be doing anything more that is very latency sensitive than ANC.

As for the last point, have fun with your sound streaming device that has 3 bit/s of UI bandwidth and a battery life of 1/5th that it would otherwise. It's not an issue you can fix by trying to fix it in that use case, it's an issue you fix by improving human interfaces, battery technology, and cellular technology, which is already being done. When that will be good enough for what you're describing, we can talk.

As for me, I'll enjoy having headphones with high quality and essentially unlimited battery life, and use the carefully optimized interfaces and computational power of my phone and computer to listen to music.

The last thing I want with my audio is some bullshit "computational experience" fucking with things and telling me what I want to hear.

Mainstream taste in audio is crap. Designing headphones for a target audience that thinks Spotify HQ mode is sufficient is not going to produce a very good product, audio-quality-wise.

He's talking about the computation involved in turning any digital sound into the analog one needed for a speaker. If there's no cable between the digital-to-analog conversion chip and the analog speaker, there's no chance for cable characteristics to have an impact on the sound.

I think you're taking about audio pre-processing stuff that Spotify or others do to qualitatively ""enhance"" audio for their lowest common denominator listeners, which is completely independent from what's being discussed.

The wording implies there is at least some level of extra DSP going on (talk of "dynamic EQ" or whatever notwithstanding). Calling a digital-to-analog converter a "computational experience" seems beyond the pale, even for Apple PR hacks

Pretty mainstream level of snake oil in the audiophile world though.

Whom amongst us has not been told that this fancy power cable will be the second coming of Christ?

Was Monster's marketing ever that aggressive?

So I guess you travel with an orchestra?

Because there aren't any "natural" / "neutral" / whatever speakers / headphones / cochlear implants. They all get a signal and a task: make this sound the way the listener wants.

"Just make it sound as it did when it was (live) recorded"? That's not going to happen even assuming some "perfect" speakers (not headphones), unless you can also supply the room, audience, humidity, and appropriate level of inebriation.

Apple has done a lot of of work on adjusting AirPods to your head and ear, IIRC. That's all software, and exactly the sort of thing they've been good at, even in the last few years.

It's basically iPhone camera vs. Canon & Nikon, the sequel. How many times did people on HN insist phone cameras will never be as good as SLRs, because the latter have the five pound of glass it takes to create good photos?

Turns out software and sensors are a pretty good substitute. While some features of good lenses are still not quite there (bokeh), they've eviscerated them in other, arguably more important disciplines (low light).

Phone cameras still aren't as good as SLRs, and never will be due to sensor size and optics.

Apple could have made these as good as $500 Sennheiser/Grado/Sony cans. But they didn't.

Rants about "traveling with an orchestra" don't even make sense when one listens to music produced exclusively in-the-box. And even if I am listening to recorded material what is wrong with nuance and detail... after all that is what the mixing and mastering engineers heard when they made it. Do you think they choose their microphones and recording hardware willy-nilly?

Software is a terrible substitute for hardware! Hardware is easier to understand, latency-free, and produces a more harmonically interesting result

Yeah man, you live in a different world from 95% of consumers. For typical users the iPhone camera will destroy what they could do with an SLR.

You remind me of a buddy of mine that has hearing loss from the marines. He is a giant headphone nut, and owns numerous $500+ pairs of headphones. All sorts of head phone amplifiers and dacs. But his hearing loss prevents any of that from really even mattering.

Does it? Or are you just saying that? You can't experience hearing through his ears, so how can you possibly say that?

Hearing loss is a weird beast that takes many forms. If he truly cannot tell the difference between Beats and his setup in a blind hearing test then maybe I would agree. Do you know if he's done this?

> Mainstream taste in audio is crap. Designing headphones for a target audience that thinks Spotify HQ mode is sufficient is not going to produce a very good product, audio-quality-wise.

Mainsteam taste is what sells products. Obviously you're not part of the target market and should probably just move on.

Apple could use their position of product leadership to introduce mainstream consumers to a higher standard of fidelity if they wanted to, particularly at this price. But I guess that would cut into their Beats business and would make Apple Music look worse than it already is, and surely the profits aren't as nice as using plastic(!) drivers

You can get Sennheiser HD280's, where the audio quality is likely worlds apart from these, and earcup padding that attenuates outside noise by something like 30db passively, for $60. Or you could be a doof and buy these Apple cans

Spotify HQ mode is not a bad digital source. Do you mean that's all (these) people think is required to get high quality playback?

It's not bad but neither is it good. The compression still butchers a lot of sound. It's easy to tell when compared against the original CDs or FLAC backups for certain kinds of music

Spotify high quality is 320 kps, if I remember correctly. This should be indistinguishable from lossless unless something is weird with the track.

I've done A/B testing with decent headphones (http://abx.digitalfeed.net/list.html) and not been able to tell the difference. Maybe you can, but I'd bet that most people can't, and in any case the result is far from butchery.

320 kbps mp3 is already indistinguishable for most people with most workloads, let alone the much more performant 320 kbps Vorbis that Spotify uses.

A lot of people love to fret about this stuff and convince themselves they can tell for the exact same reasons uninformed people go out and buy $200 cables.

It's also the same as people who install Gentoo so they can tune their CFLAGS to be 0.0000001% faster under certain workloads at the cost of not running on any other machine and crashing inexplicably on occasion.

We all love optimizing things, regardless of the actual real-world benefits.

I will have to listen more closely, though last I remember it was like 256kbps.

I've mostly noticed differences in lower frequencies seeming attenuated in electronic bass music, though it is possible that this might be due to mastering (for stream vs CD) or normalization

I bet you are fun at parties.

And I'd bet you my parties have far better sound than you've ever heard

Reply on news.ycombinator.com