Mix Challenge

Hello MC,

With the recent patent expiration of MP3 and somewhat bigger spread of lossless audio formats like FLAC, I've seen myself looking for lots of info about different formats. I'm also trying to use as much FOSS and Open Source software as I can, and I have been releasing all my sounds in FLAC exclusively. Now I have just read the Advanced Audio Coding (AAC) Wikipedia page, as it's said that the MP3 creators are now pointing towards that format since it has a much better quality at low bitrates.
I would like to know your opinion on that format, do you think it could be as long-lasting as MP3 has been? Supposedly AAC is 20 years old already and it's one of the first times I have seen it in the wild (and actually I thought it meant Apple Audio Codec ). My CD rips are just too big to store in my portable devices, and I'm looking for a format which will be somewhat future-proof and with a decent quality, also to distribute my music in.

PD: I have tried OGG Vorbis and Opus (I don't remember which of those two was the free one), and well... Certainly not the best format I have used.

Regards,

Jorge

Funny that you bring that up, because there is currently a lot of confusion along with (sadly often wrong) information floating around on the internet.

Let's start with the "death of MP3"/patent expiration of MP3 first - which can be misleading (I can agree to that). In fact, the "license program for the MP3 format" and "certain related patents" ran out. MP3 is now considered "patent free". Meaning, the format is not dead at all - it will continue to flourish. In fact, we might see more and more tools using that format, since no licenses can be "addressed" anymore, which in turn means that "everyone" can use MP3 without problem now. Which was the big debate back in the days if LAME was actually considered legal or not.

I found these articles in recent days:
https://motherboard.vice.com/en_us/arti ... s-not-dead
https://marco.org/2017/05/15/mp3-isnt-dead

And this is the official announcement by Fraunhofer
https://www.iis.fraunhofer.de/en/ff/amm ... s/mp3.html


Fraunhofer does recommend to go to AAC though. Especially since AAC is common-place in portable devices, it's been used by Apple iTunes for years, and AAC i's part of the MPEG-4 video/audio stream.

The future for Broadcast streaming however, is MPEG-H - a new format that starts to kick off end of May in South Korea. It offers up to 16 audio channels, object based surround and full control over the audio stream (with presets that can be select on the playback side, and further interacted with). This was one of the topics of the AES Convention #142 in fact.




OGG (the "evolution" of that is OPUS1) is an open source format. Both OGG and OPUS1 are mainly used by Google for Youtube now. HOWEVER... only Google Chrome plays that back natively. All other browsers fall back to the MP4 format. At least to my knowledge/tests.



FLAC is a great "Lossless" CODEC, however it's sadly not widely supported - however, more so than WAVpack (which can actually go up to 96/32, while FLAC can only go to 96/24 - both are surround capable though). Apple for example seems to refuse to incorporate it into their OS (CODEC wise), and there are not many hardware players that can actually play back that format. PC wise, and with the right apps on Android however, absolutely no problem.




So which format is ultimately the best?

If you're after Lossless encoding and wide-spread usability, then I guess it's HD-AAC. Why? This stream has both a lossy and lossless version of your audio file in one. If the playback device can't understand the HD-AAC stream, it falls back to lossy AAC. HOWEVER... creating HD-AAC is not easy (so far, I only know of Fraunhofer Pro-Codec and Wavelab 9 and up), and there are still players that can't understand the HD-AAC format (I've yet to test HTML5 players for homepages), many hosts even don't interpret the AAC format right (decoding).

But to me, it's the natural evolution of MP3. "Plain" AAC (as in lossy AAC), there are many open source encoding tools available at this point. So in the long run, I think MP4 Audio (aka AAC) is the future.

In the near future, META DATA will also be a huge topic. By that I mean that the CODECs can store values like "Loudness in LUFS". But I can't comment on that yet - I have to read myself into that topic as well, what's possible currently.





One question might pop up now - which format should we use for the Mix Challenge?
Personally I'd say, we can stick to MP3, FLAC and WAV.
If we want to use AAC, is up for debate.

Hey Mister Fox, thanks for the reply! Some info there that I have to re-read, understand and research about... Thank you again.

Regarding the last paragraph, I didn't intended to suggest using AAC for the MC, I understand that even if we all use the latest tools available, there may be DAWs that don't support AAC (I don't remember seeing it in Reaper and FL Studio), and MP3 is good for the moment. The metadata discussion is also really interesting, we could store loads of useful data in the tracks and have each program interpret it and modify other stuff like equalizer curves and so on.

But ultimately the multitrack aspect of these new audio file formats is what brings me the most joy. I fancied the Native Instruments' Stems format when they released it and even tried to transform some tracks of mine to that specification, but the resulting product had a really awful audio quality. I believe in a future four-track (I think it's the best number of tracks) format which is enjoyable for musicians, producers and advanced listeners alike and doesn't sacrifice the size while keeping a decent audio quality. Maybe those efficient AAC encoding algorithms can help?

Regards,

Jorge

A couple of things to note:

- MP3 won't die out any time soon. However, the more and more popular AAC (or M4A) is getting after setting the patent free, the more likely that we will see AAC renders from anywhere on this planet. In fact, for the sake of compatibility, the "Mix Challenge" is using MP3 and WAV, with FLAC only for those being on dial up.

- META DATA doesn't mean that we can control EQ's or other audio related tools - at least not in the way you might imagine. The META DATA is only for certain tasks (like giving a value for Loudness Normalization during playback), or additional explanations of the file (song text would count to that)

- the "multi-track" capability of AAC (and MPEG-H for broadcasting) is not your old "4-track" solution. Theoretically, it could work for MPEG-H (according to what I got to know via AES #142). But in reality, we talk "surround channel usage" or alternate voice over.