Hacker News new | past | comments | ask | show | jobs | submit login
Aura: shipping software that was delayed for 4 years because of patent licensing (alastairs-place.net)
121 points by josephpmay on Feb 23, 2018 | hide | past | web | favorite | 25 comments

The cherry on top is that in the four year limbo between this software's completion and release, Apple removed the optical audio port that it requires to function from all new Macs.

It sounds like people have a love-hate relationship with Apple. If you hit the sweet spot, you are minting money, but they can cut you off at any time by exercising their control over the ecosystem.

Yeah, that was just a tragic kicker to the story.

>AC-3, like the competing DTS standard, is a non-optional part of various other standards, including ATSC, ...

At this point it appears that Dolby has managed to worm its way into the new ATSC 3 broadcast TV standard as well in the form of something called AC-4.

Dolby started as a tape hiss reduction system that actually was useful. Since that time it has more and more become just a way to extract license fees from companies making electronics that involve audio.

For the TV standard they could of just used the open Opus codec, but since the people who are creating the standards are the ones getting the licensing revenue there is no real incentive to do something like that. These "industry standards" have become a kind of a scam.

So no, Dolby is not all that interested in licensing software implementations of their codecs for random programs. That is simply not where the big money is and it isn't really their business.

I wonder if the big players creating the hardware are part of these patent collectives and are also receiving some of these licensing fees. That would be a good way to keep smaller competitors out of an industry.

I work for a rather large company that works with audio and video and we used Opus but have since switched to a Dolby based audio stack as it offers much better sound quality, reduced transmission bitrate and is generally more robust. However this needs a direct involvement with Dolby and not just of the shelf stuff.

So Dolby has some sort of audio compression method much better than the state of the art that they are keeping secret? Why would they do this?

I thought this is a "shipping software". But it's a music related app for Mac (Which seems to be pretty cool).

So, I am thinking this post deserves a better title.

Does the optical input for the amplifier only suport AC-3? Why don't manufacturers use a more naive patent-free method that might be slower? Are they really starving for bandwidth enough to justify the cost of an AC-3 license? Even uncompressed PCM times 6 channels could easily be transmitted over USB 2.0, and I was under the impression that these optical formats were even faster.

Usually the four codecs supported over optical are uncompressed (limited to 2 channels), AC-3, DTS, and MP2.

The patent status of MP2 seems to be hard to determine.

DTS seems to be still patented.

Weird, so manufacturers just haven't implemented uncompressed 6-channel optical audio? Is it bandwidth?

The current standard tops out at a clock of 6.144MHz which is exactly enough to carry 24-bit 48Khz 2-channel uncompressed audio in a 32-bit, differential Manchester encoded frame that includes sync & signalling bits. Running AC3 and DTS over this is a bit of a hack, but it's a standardised hack.

HDMI supersedes S/PDIF so comprehensively that the S/PDIF standard will likely never be substantially revised. There's no compelling reason to invest in the R&D.

Uncompressed stereo 16 bit 44.1khz sample rate (cd quality) is <150 KBps so it's definitely not bandwidth.

It's bandwidth. More specifically, standard TOSLINK only has the bandwidth to carry stereo audio in uncompressed form. If you want full surround sound it needs to be compressed to fit into the same bandwidth that stereo audio would use.

I thought the problem wasn’t toslink (the physical optical layer) but rather s/pdif (the software layer) which is constrained by its support for the rca physical layer?

IIRC, toslink 1.0 (circa ‘82 or ‘83) had a limit of 2 or 3 mbps while modern toslink is well over 100mbps (constrained more by distance than speed).

SPDIF is not a software layer, but a limited version of AES3 for consumers. In the OSI model, AES3 would be considered layer 2.

Layer 2 is not considered software, although it can be performed by software in systems where such a concept makes sense. However, in AES3, this does not make sense, and has always been handled in devices via a chip that merely turns AES3 into I2S to be fed to either a DAC or another interface chip (ex: conversion between AES PHYs is often done by just bridging the receiver of one type to the transmitter of another type semi-blindly via I2S).

Also, the AES3 protocol is incredibly simple. It only has a very limited set of flags (very limited in today's sense, more than enough for what was intended in 1985, see https://en.wikipedia.org/wiki/AES3#Channel_status_word ), and was never actually meant to do anything but stereo audio at conventional rates. Doing encoded bitstreams across it is largely a hack, and is limited to the base versions of Dolby and DTS.

Using >3mbit over Toslink is not done involving an AES3-purposed PHY, but instead with other unrelated protocols, such as Digital ADAT. Distance on typical Toslink is already constrained to about 30 feet.

I have seen other people quote this mythical 100mbit number, but no one has ever been able to cite such a use in the real world: no standard I have ever seen that uses Toslink as a PHY has supported such a transmission rate: ADAT maxes out at around 12mbit, and MADI does not use Toslink at all.

Also, in further irony, in use, I have found unbalanced AES3 (such as via RCA or BNC) to be more reliable than Toslink; and virtually all of professional AES3 usage is done by 3 pin XLR because of how unreliable Toslink is...

Although a lot of professionals have moved on from AES3 entirely, and use MADI or Dante, or don't use digital transmission pipes inside of their studio at all and run it entirely on a single PC.

Very interesting reply, thank you. I meant to say “protocol layer” rather than “software layer” but misspoke. The correction is appreciated.

I think I’m going to do some more research on that 100mbps number. Perhaps they meant Bd/s?

CD quality is 1.4 mbps

16 * 44,100 * 2 = 1,411,200

Lower case "b" - so bits. 1,411,200bps = 176,400Bps ~~ 172KBps I wasn't that far off.

What happens if you connect a Mac to a receiver using an HDMI cable? (And then chain your monitor to the receiver.)

HDMI can carry uncompressed 7.1 channel audio - will a Mac send that? If it does, does any playback software support that?

Yes, just tried it from a 2015 MacBook attached through HDMI dongle to a Denon AVR. Playback of Dolby test videos and the receiver shows 7.1 audio with correct perceived sound position. Tested with VLC, Quicktime and Elmedia player. It's possible the audio stack is remixing on the way through, because I tried a 5.1 AC3 file as well and receiver was still showing 7.1 except with Elmedia player configured for audio pass-through, upon which it showed 5.1 audio (and issued a noticeable click when the change of format occurred)

This deserves a better title. HN's title submission rules are going to hurt chances of it being up voted despite the good story. Something like:

Aura: delayed 4 years by "Reasonable and Non-Discriminatory” licensing

edit: Thank you both folks who replied. I posted this comment, emailed the mods, then saw your replies. :-)

Agreed. Thanks! and thanks for the email.

They aren't rules. They are guidelines. If you feel it really needs another title, you can email the mods. They are typically very helpful where there is a good article with a terrible title

The rules are bent when deemed useful. Suggest a better title to the mods via the Contact link in the footer. I've seen it done in the past.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact