From the FAQ: Does Opus make all those other lossy codecs obsolete? Yes. From a technical point of view (loss, delay, bitrates, ...) Opus renders Speex obsolete and should also replace Vorbis and the common proprietary codecs too (e.g. AAC, MP3, ...).
CDs are 1400 Kbit/s and FLAC might be half that at 700 Kbit/s. I have 50 GB or so of music at 260 Kbit/s AAC. If I were to use FLAC instead, I would need 140 GB of storage. Practically speaking, for a MacBook this means going to the 512 GB storage model instead of the 256 GB storage model, which is an extra $300. For my desktop, this means either dropping $200 on another SSD or going with rotational storage for my music library. For my phone, this means downsizing my library by an additional 60%, and I have already had to trim a lot of music out. Or I could buy another phone and spend $200 to get a model with enough storage for me. (Streaming to the phone is not always an option. I spend time in areas without reception.)
However, since the difference between FLAC and 260 Kbit/s AAC is imperceptible to us mere mortals, I can spend my $500-700 on something more interesting than "insanely cheap storage", and I get to enjoy my music.
Yes, let's not have people transcode music needlessly. But stop saying that "storage is insanely cheap" because cheap storage isn't portable.
FLAC is great for archiving music but that's not something that I do. I don't archive music. I listen to it.
Depends on the storage context.
If you want to carry a significant fraction of your music collection on a phone, lossless is not an option. Lossless is what I use for my main storage (a machine with a dozen terabytes-worth of spinning rust), but the music is converted to lossy before being synced to my laptop and phone, and I'd rather have lossy files in these contexts (whose output quality isn't exactly top-of-the-line anyway) than having to juggle external drives.
In that context, converting the lossy versions of the library to OPUS (with master files remaining lossless) makes perfect sense.
My gold library is in FLAC but I transcode to MP3 using mp3fs for various uses.
I'm buying music in FLAC, and store it as master copy. Then I convert it to Opus for actual playback. If better free codec will come out, I'll do the same thing, using FLAC as a source. So FLAC has its usage.
If you want good sound, treat your room. Then buy _good_ speakers or headphones. Then buy a good sound interface. And only after that should you start worrying about signal qualities above 320kbps mp3.
Given as majority of people don't actually hear the difference form the first of these steps, advice about flac is only relevant to a really small amount of people, who are not only capable of hearing the difference, but also have gone through all previous steps.
> rates are internally converted to 48 kHz
> only frequencies up to 20 kHz are encoded.
> In particular, software developers should not use Opus Custom for 44.1 kHz support
There are times when I need to not drop non-audible frequencies. Like when my microphone is on one system and my voice recognition is on another, voice recognition needs the full spectrum for accuracy. There are times when I need 44.1kHz and not 48 kHz, like on an embedded system with everything running at that and no performance left for converting or extra PCM channels for playing with different settings at once.
They keep saying it is designed for the internet, but did they miss the fact that all major players now have voice assistants like Siri, Google Assistant, Cortana? Did they miss the fact that more and more sensors are cheap embedded throwaway IoT devices? It's like it is designed for the internet of the 90s.
> There are times when I need 44.1kHz and not 48 kHz, like on an embedded system with everything running at that and no performance left for converting or extra PCM channels
Unfortunately we live in an interoperable world. And there are _many_ devices running at 48kHz without the resources to resample. Your devices would be unable to interoperate.
Being "opinionated" in this way is specifically to accomplish the goal of guarenteeing interoperability in a world with cheap throwaway IoT devices.
> the fact that all major players now have voice assistants like Siri, Google Assistant, Cortana?
You might want to look at how these systems are sending audio...
MP3s actually tend to have a worse cutoff depending on encode settings, with values from 20.5 down to 16 kHz [1].
[1] https://www.whatinterviewprep.com/prepare-for-the-interview/...
> voice recognition needs the full spectrum for accuracy
do you have a source for this? Voice signals are conventionally low-bandwidth; 16kHz is usually "good enough" for human-human transmission. Formant frequencies top out around 3kHz [1] and upper vocal harmonics are not really important outside musical applications. Consonants are a bit more complicated but I'd be interested to know what voice information is present above 20kHz.
[1] https://en.wikipedia.org/wiki/Formant#Formants_and_phonetics
I long ago stopped caring about lossy formats for music as everything is FLAC on the NAS. For audiobooks and similar squeezing another couple of percent compression just isn't worth the effort nowadays, so they can stay in whatever format they came in, preferably MP3. When you've a few TB of space it mostly doesn't matter any more.
> Opus combines the speech-oriented linear predictive coding SILK algorithm, and the lower-latency, MDCT-based CELT algorithm, switching between or combining them as needed for maximum efficiency.
Note this is from before 1.1.1 and 1.1.3 (above) which each contained ARM optimizations so perhaps quite a bit better?
One of the nice side effects of this is that transport streams are really low-bandwidth. If I set up a JACK master server, and, say, six slave laptops recording input, I can stream/mix over 802.11g without latency, whereas other solutions will fully saturate a gigabit switch.
It appears in some of the graphs on the comparison page. Is that just the closeness to the original audio sample?
So, you could easily have an encoding which sounds worse than a different encoding, even though it's actually bitwise closer to the original.
For example, humans hear very bad above 20 kHz, so you can pretty safely drop any information about sounds above that. A bad codec would keep that information around and instead drop something in the usually audable range. And then it'd sound worse, even though it's really close bitwise to the original, from all the information it didn't drop above 20 kHz.
So, yeah, you actually need humans to rate the quality.
