Hacker News Reader

Comment by tonmoy | Hacker News Reader

Having worked on Audio Codecs and even though my livelihood depends on it, I would like to say 3.5mm jacks suck! You need to figure out how much resistive and how much capacitive load any headphone has. You need to detect if it has mic, if it does is it US or China standard. Need to make sure that the DAC and amplifier don’t consume too much power yet have a good dynamic range for arbitrary cable length. Let the phone send digital audio data along with power using USB-C and have the earpieces play the audio, it’d be much simpler IMO

View on HN

Those all sound like good things to optimize to make a really good headphone driver, but they aren't strictly necessary are they? "Standard" line level analog audio has been a thing for what, 90 years? I could explain to an 8 year old how it works. It's very simple.

Embedding an IC into a cable that talks USB 2.0 doesn't seem like a simplification to me. I'm a staff level embedded engineer, and I wish an 8 year old could explain to me how that all works.

Any DAC still ideally does all the things you listed optimally, so you're just externalizing the same complexity by leveraging the marvels of modern technology. It's a pain to have to carry another dongle. They feel like they're going to damage my USB-C port when plugged in inside my pocket. Do buttons even work at all with those adapters?

I'd rather the smartphone engineers build an awesome DAC and charge me an extra $15 for the phone, rather than have to buy a $15 dongle of questionable quality on Amazon, or a $50 dongle of even worse quality from Best Buy or something.

The USB-C audio mode just pushes analog audio over the pins, so the hards DAC stuff stays inside the phone and just gets harder with all the multiplexing required.

And fully agreed on USB-C ports being even more fragile in the pocket than 3.5mm.

Lol bluetooth sucks a 1000x more. The protocol stack is itself a nightmare.

Bluetooth is terrible. My gripes are: 1. Re-connect from mobile to desktop takes minutes. 2. Batteries need to be charged (and are always not charged when it matters)

Sony xm4 and xm5 have dual connectivity.

Same with Bose. But I have two desktops and a phone, and never know which two are connected. Why limit it to two? Seems arbitrary and unnecessarily limiting.

Generally multiple bluetooth signals seems to degrade the overall audio quality.

The AirPods are the only product in the same class that can sort of link to more than 2 devices... but all of them need to be apple.

They don't support multiple Bluetooth audio connection at the same. They basically quick switch between devices in the background.

I'd be fine with that. The Bose headphones seem to be doing something similar as only one device can play audio at a time.

Dual connectivity on the XM4 is horribly broken for me. As soon as you try to use the mic, nothing works reliably anymore : the headset randomly stays blocked on one profile on another and you invariably end up with either crappy sound for your music or no voice for your zoom call.

Now I use it exclusively on my phone for music and I have another headset for voice calls on my computer. Which is a shame given how much money I had to throw to buy this thing.

Still, I have to acknowledge, sound is awesome, so that’s why I keep it anyway.

AAC and LDAC should be good. I even use them with PipeWire. Once you switch the mic on though, you end up with a different (low quality) audio profile. It should switch back to high quality profile after but I guess it does not?

Also does dual connectivity even work with LDAC? I feel like the best way is to use software to connect manually. You do have to specify the address and such, and execute the action, and you have to disable Bluetooth when you don't use it. Advantages are less tracking and interference.

As for XM5 it is expensive compared to XM3 and XM4. If you want to spend that extra money there are better targets (victims of current war for example).

The biggest thing that bothers me in Bluetooth is the lossy audio compression

3. Manufacturers forcing you to install their (adware infested) mobile apps to remap or use the hardware buttons

For developers, yeah, Bluetooth is like a torture device, but for consumers, it's wonderful.

I take my Soundcore buds out of the case. They connect in seconds. I put them back, and put the case on a wireless charging pad. There is no interruption to any of my other Bluetooth items.

I wouldn't even look at anything non-wireless. It would have been nice to have pro audio grade latency, but it's fine for music and not particularly noticable in calls.

As a user I don't love having the sound keep dropping out because I went too close to the train station, or because my phone was rotated slightly wrong in the pocket.

I noticed when stepping out of a train BT drops for a brief moment, what causes this interference? The powerlines ?! Interference from other commuters?

...where on earth are you guys that this happens? I'm commuting every day with Bluetooth headphones and never, ever had a dropout while listening to podcasts and stepping on/off trains.

> For developers, yeah, Bluetooth is like a torture device, but for consumers, it's wonderful.

Let me respectfully disagree. Most of the times it's a pain, from my personal experience.

Sure it it's good enough if you're just playing audio, but it's garbage if you need to keep the audio in sync with the video (if you can't do latency compensation), so it's useless for games. And the other major problem is that the audio quality drops down to garbage if you want to use the microphone, as it needs to switch to HSP or HFP. I find it ominous that the marketing for LE audio has completely ignored these problems, instead focusing on minor improvements like streaming directly do both buds.

Wonderful when it works. It's always a crap shoot which device will connect to what. I don't have particularly weird kit.

Bluetooth is fine, it's just the codecs that suck.

As soon as you enable the microphone on a headset, the codec switches to "American cell phone circa 1990" -mode and everything sounds like shit.

Airpods Max seem to circumvent this with the W1 chip though and it's a complete game changer. I can actually use noise cancelling while having the mic enabled.

That's a shame, but at least you can't run DRM over a 3.5mm cable.

I suspect that's why everyone's dropping it.

I tend to doubt that. Even ten years ago on enthusiast music piracy sites it was quite rare to see pirated analog recordings of DRMed digital audio. There isn’t even widely used end-to-end audio DRM like HDCP, is there?

10 years ago, there wasn't a good reason to do that. 10 years from now, who knows... The more lock-in, the "better"!

Yeah let's dump decades of electronics relying on different sized Jacks to the landfill for your convenience. And let's not forget that you can upsell the mini usb-c dongle for a decent margin!

There's a middle ground working for everybody without going into audiophile territory.

>You need to figure out how much resistive and how much capacitive load any headphone has.

Why would you? That's part of the headphone's characteristics. People who buy high impedance headphones should know when to use an amplifier. That's not the DAC's job.

So hard yet millions of devices, cheap and expensive, had no problem providing it for decades.

I remember being a little shocked to discover that I had better battery life on my phone with Bluetooth than I did with a wired connection.

The dac is running off the batteries in your Bluetooth headphones, so that's a pretty natural outcome.

And the power amp which probably consumes more power than the DAC.

The size of batteries in two piece earbuds is miniscule and would not make a difference in phone runtime.

...and people still want to manually disable bluetooth on their phone "because it eats up their battery".

Dude. You posting that message to FB took more power than your idling bluetooth will spend all week :D

I turn it off because historically it's just less secure WiFi, and, bonus, junk to use more often than not.

Your Bluetooth headphones have their own batteries. They also have terrible sound quality compared to equivalently priced wired IEMs because powering multiple drivers/speakers per ear for maximum quality takes more power.

You really want to take up your only usb-c port for headphones? Would be nice if they gave an extra USB port to compensate for the loss.

Also USB can do analog audio supposedly...

>Also USB can do analog audio supposedly...

That's how the USB-C to 3.5mm adapters work for if you need to use headpones which don't have a USB-C cable.

I'm pretty sure USB-C adapters have tiny little DAC ICs embedded inside them. They show up to the phone as a USB sound card.

A lot of USB-C cables have little ICs built into them. It's required as part of the PD spec. A port isn't supposed to provide a laptop with 60W or whatever unless the cable itself communicates that it's rated for it.

Like everything USB-C, there's several incompatible ways to do it.

There is both. Some phones, e.g. OnePlus, have analog audio output on USB-C and can use a passive adapter. If the phone lacks analog output, e.g. Samsung, it needs an active adapter with a USB controller and DAC like you mentioned.

Passing an analog signal fractions of a millimeter from a bunch of digital signals is a recipe for very crappy sound. I have yet to find an analog USB-C audio adapter that didn’t have terrible quality and usually some hissing or other noise too.

I have some balanced armature earbuds, and they are sensitive enough to pick up noise from poor digital grounding even in the 3.5mm jacks in a lot of phones. Additionally, they are often too loud even on the lowest digital volume setting. 8 years ago I actually made an inline attenuator, which solved the problem on my phone at the time.

My current phone is a pixel 6. I got a USB-C adapter with a DAC built into it, and it actually sounds pretty good. The volume levels are reasonable, and there's no noticable noise other than perhaps an unnoticeable hiss.

Years ago, I had a desktop case front IO panel with a really bad ground loop. I ended up cutting the molded over plastic open and cutting the PCB traces between the audio ports and the USB ports, and the problem went away.

USB-C is also a minefield - there are multiple incompatible ways to do audio over USB-C (including outputting it in analog for cheap adapters!)

> figure out how much resistive and how much capacitive load any headphone has

This is not and has never been a problem for consumer based devices. All headphones are made +/- with the same impedance.

> Need to make sure that the DAC and amplifier don’t consume too much power yet have a good dynamic range for arbitrary cable length

If cable length is an issue for you then you're using a really bad cable. I mean, really bad. Or have a severely underpowered output amplifier.

Analog is simpler than digital. And if you think your cheap headphone will do sound better than the phone manufacturers I have a bridge to sell to you

> This is not and has never been a problem for consumer based devices. All headphones are made +/- with the same impedance.

A quick search for "high-impedance headphones" might be enlightening on this topic.

For example, Beyerdynamic offers 32, 80, and 250 ohm versions of several of their models.

> If cable length is an issue for you then you're using a really bad cable. I mean, really bad. Or have a severely underpowered output amplifier.

A cable has both resistance and capacitance, which means every cable is also a low pass filter.

If you are tuning the DAC/amplifier with power consumption as one of the criteria, then actually, yes, you will want the output power as low as possible -- which means cable length is a factor in the optimization.

> For example, Beyerdynamic offers 32, 80, and 250 ohm versions of several of their models.

Yes and you know what happens when you plug a higher impedance phone in a low impedance output right? (and the point is moot because we're talking about consumer devices)

> Headphones with impedance greater than 100 ohms are typically older or professional studio specific design

(or valve outputs - because of course it's "audiophile stuff") https://www.headphonesty.com/2019/04/headphone-impedance-dem...

> A cable has both resistance and capacitance, which means every cable is also a low pass filter.

A cable is a https://en.wikipedia.org/wiki/Transmission_line and while it is technically a low-pass filter you have to show me actual numbers to convince me that it will significantly affect, let's say > 10kHZ at usual headphone cable lengths.

That is unequivocally false. Amplifiers (inside the DAC) are not ideal voltage sources at all loads and frequencies. Depending on the capacitive loads the amplifier might need to adjust compensation to make sure it remains stable in all situations. For resistance you will need to limit the output voltage to ensure you’re not going over the over current limit, or you can get hurt on dynamic range for high resistance loads. To make a good product then you’ll need to dynamically detect the resistance and adjust the voltage. All of these problems are trivial and would increase efficiency greatly if you know exactly how much load there is connected. With all the crazy 3.5mm jacks there are it is inefficient to make a product that fits all. And don’t even get me started on splitters and all the types there are.

Headphones are mainly inductive charges. A short cable won't have enough capacitance to change that. (maybe in audiophile crazy-land, that is).

> amplifier might need to adjust compensation

You're pushing audio, not a cable TV signal...

> For resistance you will need to limit the output voltage to ensure you’re not going over the over current limit, or you can get hurt on dynamic range for high resistance loads.

What crappy device are you plugging into a headphone jack that's even an issue? Short protection is common in most ICs.

Or, you know, use a regular headphone with an impedance between 32 an 100 ohms and that should be fine.

> crazy 3.5mm jacks there are it is inefficient to make a product that fits all

What's so crazy about it? Yes, stereo and mic inputs might be unpredictable, but for usual stereo audio it's pretty predictable

People are saying that like if most USB headphone vendor is not going to plug the cheapest DAC they could find and push whatever to the headphone.

The 3.5mm jack/cable is much more reliable than a usb-c port.

Reply on news.ycombinator.com