Hacker News Reader

Comment by nicolas_t | Hacker News Reader

I use the optical digital out from the headphone port about 8 hours a day. It's connected to my audio technical dac/headphone amp.

I'll only be interested in wireless headphones once they actually have decent quality. And even then, I already have very good Beyerdynamic headphones (the T90 and the DT880) and I don't see why I'd spend money to change them.

I'd be rather annoyed at not having hdmi or sd card mostly because carrying dongles is a pain.

It's kind of sad but Apple products just do not excite me anymore. I remember before I used to try and follow the keynote to see the new iphone or I looked forward to seeing the first intel macbook when it was unveiled... Now, I look at the iphone 7 and there's really nothing in it that makes me want to buy it. There hasn't really been anything new with their laptop either (ok, maybe the new small macbook is not bad it's a good travel notebook)...

View on HN

Genuine question, is there any particular reason DAC is connected via optical rather than USB? Or was it because AT's DAC do not have USB support?

I asked this question because I initially thought anyone with an high-end headphones are least likely to be affected if Apple removed headphone port from a Mac (because the included headphone port usually isn't powerful enough to drive some of the high-end models anyway, and many would likely to replace the DAC too, in addition to adding an Amp.)

Optical just feels like a cleaner interface. USB requires drivers, and introduces timing and electrical interference issues.

USB also feels like overkill for an audio interface. It allows transfer of data in both directions, and introduces security risks if you don't fully trust the DAC you are connecting to, or worse still if the DAC is connected to a network/the Internet.

I'm not suggesting for a moment there is any audible difference between the two, but optical audio has always struck me as a 'cleaner' way to output digital audio.

I would be extremely annoyed if it were removed from the MacBook Pro.

"USB requires drivers, and introduces timing and electrical interference issues"

On multi-tasking OSes with memory protection, everything that accesses hardware requires drivers; it's 'just' the amount of driver code that differs (which, for USB vs optical, probably is quite a big difference)

Also, originally USB didn't support isochronous channels at all, so one couldn't guarantee that one could send audio out continuously.

Most USB DACs do not require special drivers. The drivers are included with the OS. If you don't trust the device to be plugged in via USB then you shouldn't buy it. Most DACs are fairly simple hardware and tear downs are common. Check Head-Fi and other audiophile sites.

> introduces timing

SPDIF also combines the clock and data in a single signal, it has more or less the same issues with jitter that USB does. Not that it matters since I'm not aware of a single proper study that indicates jitter is audible.

I would be shocked if there wasn't a hardware fifo in that receiver path.

Depends on the device.

Often there is a receiver IC that will perform clock recovery and convert from S/PDIF to an internal digital audio format, I2S most commonly, to send it to a DAC IC. In this simple setup there generally is no more buffering than needed for operation of the pieces.

In more complex situations, either where the S/PDIF signal is being re-clocked or where the data on the S/PDIF line needs to be decoded, then they will be buffering.

To clarify some terms:

Re-clocking - the system detects the clock rate of the signal but does not use that clock as the reference, instead it will generate it's own clock at the proper frequency. Some buffering is needed here to offset drift between the clocks causing sample starvation.

Encoded data - S/PDIF (really AES3 for the most part) has a framed block transmission format not entirely dissimilar to a TCP packet. There is meta-data, frames, sub-frames, etc, it's not just 'pure audio'. Various formats can be stuffed into these data frames, like compressed audio data (5.1 Dolby Digital for example).

USB seems a bit overkill but since it's already there isn't it more overkill to add a dedicated optical port for digital audio?

The digital optical port is baked into the 3.5mm jack on most upper-end laptops and Macs. Though sadly my MacBook Air does not have one.

The port isn't the important bit that is added, it's the supporting circuitry and photo diode to drive it that adds complexity to the design.

And the little physical switch in the port that broke on my last MBP. My port got stuck in optical mode.

Mine got stuck serveral times too. Was alway able to fix it with a toothpick.

I have a similar setup; I use optical anywhere and everywhere I can because it isolates electrical noise. No mains hum, no ground loops, no 'chainsaw' noise from poor switched-mode power supplies. No copper path from source to amp at all.

I try to steer clear of some of the more 'woowoo' audiophile stuff; but if a $10 amazonbasics optical spdif cable can do more to reduce noise than any thousand-dollar unicorn-dust copper cable on the market, I'm in.

It's because the AT's DAC/amp that I use doesn't have usb support but I do have a USB DAC and it's proven more flaky. The advantage of an optical connection is that it doesn't need drivers and so you only need to rely on the constructor being good with hardware and not on them being good on the software side of things (which most Japanese company cannot do).

In this case, I use a DAC combined with the Amp but honesty while I see the difference that an Amp brings, I don't see the difference of using a different DAC than apple's DAC (I've tried doing blind tests). So, I could see myself just using an headphone amp without using a DAC..

Why doesn't an optical connection need drivers?

It needs drivers, they are just the same as the audio jack. At least in my computer (not apple), the same board that drives the audio jack is, at the same time, providing output for the optical plug without any special software.

I've never used an external DAC, but it seems that the parent is weary of companies fucking it up with the usb drivers (or the firmware in the device) and introducing delays or artefacts in the audio stream. An optical to audio converter is easy, probably any third year EE could produce a working prototype.

Ok, to be clearer, an optical connection uses the drivers from apple and will always work well whereas I've seen some usb dacs that used custom drivers and would fail (example AT-HA90USB)

The reason against USB was always latency. Firewire was doing better in that regard but optical was hard to beat. Not sure if that has changed since.

Yes, that's some pretty old information, but I used to think the same thing. The main cause of latency is audio buffer size. At the same buffer size, the difference between FireWire, USB, and optical is not really enough to care about. The main reason you'd use a higher buffer size is if your software can't handle half as much data in half as much time.

There were some problems with old software and hardware that have been resolved for the past ten years or so, but USB audio could be problematic in the early 2000s. Optical was never really any better than FireWire in terms of latency.

> The main cause of latency is audio buffer size.

Right. I'm really not an expert on this but my understanding is that with S/PDIF supposedly you have a control over the buffers on either side whereas with USB audio you are limited by the USB spec.

Eh, that's not a very accurate way to describe things. S/PDIF itself transmits samples of audio one at a time, whereas USB transmits frames encoded in packets, but in practice there are going to be hardware buffers of some size, and the software will process one buffer at a time. The end result means that there are no real advantages to S/PDIF, and a number of serious disadvantages.

There's at least one really nice benefit: because there are no conductive wires in optical cables, electrical noise isn't directly coupled from the source to your DAC. This can make a big difference if your source is simultaneously charging from a noisy source like a car adapter [1]. Similarly, you won't create a ground loop by connecting your source to a wall-powered DAC.

[1] I have literally never seen an isolated [2] USB charger for a car, and I've looked. You can fudge it using an inverter, an isolation transformer, and a 120V charger. I'm not talking about "my golden ears hurt" noise; I'm talking about "egads this music sounds awful and crackly" noise.

[2] By isolated, I mean a charger that provides galvanic isolation between its inputs and outputs. Some switching power supplies are isolated and some are not. Any charger that plugs into a wall plug and isn't polarized MUST be isolated because, if it's not, it'll zap you. The upshot is that your car's electrical ground ends up connected to some piece of copper in or near your amp, and car's electrical systems are noisy as hell. [3]

[3] Fun reading: http://www.newark.com/pdfs/techarticles/tyco/Auto_Network_AN... If you ever visit the NRAO in West Virginia, there's a big area where you're not allowed to bring a non-diesel car because they're so noisy.

USB is asynchronous - need to introduce timing logic. Optical is by most marks, a cleaner interface.

USB has an isochronous mode, which is used for audio streams. See, e.g., http://www.edn.com/design/consumer/4376143/Fundamentals-of-U...

You still need timing logic, you have to extract the clock signal from the SPDIF signal as well.

It just sounds so much warmer without those few microseconds of jitter, y'know.

DAC is only the audio data being sent,modulated as light pulses, USB has some overhead (and timing issues)

i have a similar setup with a consumer level tube dac (highly recommended with a pair of studio monitors, btw).

1. i used all the usb ports on my workstation and don't want a hub

2. i use the usb and line-in analog inputs on my dac for portable devices, the optical one is permanently connected to my workstation.

3. optical feels 'native' in the operating system, it's just another output on the soundcard and doesn't require anything other than native drivers. this is probably wrong, but whatever.

Things might be interesting again if Apple ever tells the world what they're doing with the PrimeSense (designer of the original Kinect) assets they bought. Consumer depth-sensing cameras took a major step backward when they pulled the plug on PrimeSense.

I've been messing around with a RealSense depth cam lately and it's been somewhat interesting. In the past I played around with an old gen1 Kinect using stuff like TouchDesigner and ZVector but had to hold off on picking up a gen2 Kinect when the adapter needed for PC hookup was somewhat pricey and always on back order.

I haven't gotten into the gritty details yet and have only been using it with TouchDesigner and ZVector Touch so far but it's fairly promising as a low cost, decent quality RGB+depth camera.

I presume these assets were bought so they would have a competitor to Google's Project Tango if it takes off.

The first Tango phone from LG is far too big and underpowered but future Android phones could have this stuff built in unobtrusively. Couple that with better AR games (a new version of Pokemon GO, for instance) and Apple will badly need something similar to compete.

Oh yes, that's a good point. I really hope they release something based on that.

Reply on news.ycombinator.com