Is opus the final solution? How do other codecs even compete?

HE-AACv2 because my phone and car doesn't have opus support

>no adoption

RIP

Doesn't support 44.1kHz, which is a massive problem since it adds color to CD rips. Other than that, it's perfect.

Doesn't it have bluetooth? I'm pretty sure you can listen to opus like that. Put your encodes on cloud, sync them on phone and connect to your car via bluetooth.

>no adoption
It's used by Youtube.

No it doesn't niggerfaggot stop spreading FUD

Attached: radeon_vegana.jpg (1280x450, 26K)

>I can't put opus on a USB and play it on my car stereo without aftermarket

If you alter the extension to .ogg it'll more than likely play. Modern hardware can play most any codec including Opus, the media players just can't read the fucking containers because of the extension used.

Change it to .ogg and see what happens.

Opus has one major drawback that pervents me from using it: no bluetooth support. fix that and opus will rule

No standard hardware decoder = useless for 99% of real-world applications.

If you’re just listening to music on your computer you may as well use flac since storage is cheap.

>If you’re just listening to music on your computer you may as well use flac since storage is cheap.

Nooooo, you can only Mp3

No it doesn't what? If it supports 44.1kHz, say so. It doesn't the last time I checked the spec. I wish it did, because there is no ideal sample rate converter.

>vorbis fans have 3 minutes to explain themselves.
Tell Hiromoot to allow Opus on /wsg/ and /gif/ and I'll stop using it.

He's referring to
>which is a massive problem since it adds color to CD rips
I don't know enough about this kind of stuff to argue about it. According to the Opus devs
>The quality degradation caused by any reasonable resampler (SoX, libspeexdsp, libsamplerate, ...) is far less than the distortion caused by the best lossy codec at its highest bitrate.
>If you can't tolerate the quality degradation caused by a good 44.1 to 48 kHz resampler, then you shouldn't be using a lossy codec in the first place.
See wiki.xiph.org/OpusFAQ#But_won.27t_the_resampler_hurt_the_quality.3F_Isn.27t_it_better_to_use_44.1_kHz_directly.3F

using flac is like using lagarith to store your movies
Literally the difference is like 1%
Saying FLAC is better is like saying you can hear the slight change of a drum being a mm thicker when being hit.
Most FLACs have the exact same sample rate and bit depth as most MP3s

No, killing all Jews is the final solution.

>Is opus the final solution?
Not for music at under 64 kbps. HEv2 AAC is still king there. The Opus encoder keeps on improving, so it may close the gap somewhat.

From a broader perspective, we haven't reached the final codec yet. There are diminishing returns, of course, I don't think even Opus will be considered "good enough" 10 years from now. Companies will seek to improve on it for the sake of saving bandwidth.

There is a lot of work left in decoding. Speech encoded with Codec 2 and decoded with a WaveNet decoder sounds amazing at 2400 bps ("bps" - not "k"): auphonic.com/blog/2018/06/01/codec2-podcast-on-floppy-disk/#wavenet-examples.

Attached: codecs.png (933x763, 59K)

>we haven't reached the final codec yet
so much this. I think the way we work with codecs are fundamentally wrong and an absolute mess and we are a long way away from a final codec

>the way we work with codecs are fundamentally wrong
You can't just say that and leave us hanging. Explain why.

but with a flac file you can transcode it into any lossy codec with a minimum of fidelity loss. As well you'd need to, because few devices support opus.

In the lossless vs lossy compression debate, HD's measured in terrabytes are the final solution.

is opus only audio?

AAC is still better in some cases. Opus is really only relevant for phone calls and where you need muh freedums and not pay royalties.

Music cds burned from lossy files sound like shit.

Get off /g if you're not able to make something that solves that problem

This.
Literally any lossy codec forever locks you to it.
Ripped some rare music 20 years ago into mp3 that you can't find anywhere online? Sucks to be you, now you're forever locked to mp3.
Ripped something to wav / flac? Great, you will be able to transcode it to in 2050 without hassle

music has been shit for almost 300 years desu

I'm going to give a basic example as it's easier to explain. You can easily prove it wrong, but it's just to explain.
I fuck around a lot with video and video codecs, so my knowledge isn't that big in audio but some of this still applies I believe.
Let's say you made a video, and now you decide to render it, so you have a lossless AVI and you compress it with H264 to post on youtube or something. You decide to render with only one pass, but after you change your mind and think it'll help increase quality and do a second pass. Why is it that you now have to re-render the entire first pass, why cant we just continue putting information into the first file using the raw file we have. Or if we want to render several versions of a file, why do we have to re-render each file each time, rather than just compressing the next one a little more each time. Sure there are certain types of compression that don't work that way, but there are some that do.

Another issue is our lack of pushing towards lossless quality in lossy files, If you compress a greenscreen video, we still get noise in the green, despite it all being green. Now this depends on the codec, but still, why aren't we pushing to fix this.

We also don't learn from compressing other types of files. We could learn a lot from the compression of vector images and 3D models and pushing similar methods into video and audio allowing for more "modular" compression.

And a bunch of other small issues that we have that can be pointed at being codec dependent, but for every codec that solves a problem, they implement more. One might support 12-bit footage, but it doesn't support matte or depth footage. One might do matte and depth PERFECTLY but doesn't support other bit depths.

One thing that I forgot to mention would be the fact that we could have filters run on top of codecs, so for example, people apply noise to their image for a certain feel or look, but it can completely destroy the image when compressing it. Why don't we instead have codecs work differently in the fact that it processes the overlay in playback. This is just a basic example, but it could apply to other aspects such as brightness and color that gives the viewer more control over playback and helps with compression. We also could have better compression in scenes that are very dark or very bright by scaling that over a different bit depth and then converting in playback. I hate this example, but it's just too good to of an example, is the This is America music video, where at the end, it's almost entirely black, and with youtube compression, it's horrible.

The situation with video codecs and formats is thoroughly fucked and it will take another decade at least until consumers get a transparent video format, for example.
But audio is not nearly as bad. We've had a transparent consumer format for over 30 years (CD) and unless we're talking about some exotic stuff audio is pretty much a solved problem.

>opus fag gets BTFO
>response: music sucks anyways.
>absolutely pathetic.

FLAC is better for quality
MP3 is better for compressed audio
PCM (WAV) is better for legacy support and integration

How exactly do you want to mod a car radio ?

>MP3 is better for compressed audio
(You)

I actually tried Opus out.

I would use it if my iPhone and iPad could play them. I did some transcode tests from FLACs that ranged from 800-2500kbps, with a couple of different genres. I'm not using audiophile equipment. Not even a DAC/amp. Just bog standard AKG headphones, and not even particularly expensive ones. I could hear compression in Opus at 80-96kbps; it wasn't garbage, there weren't instances of severe clipping or crackling. Rather, the highs were a bit lower, particularly on symbols it was noticeable.

It was transparent to me at 128kbps. To compare, the original file was 25MB. The MP3 CBR 320kbps came out at 9MB, while the Opus 128kbps came out at 4.3MB. I could not aptly discern between the two, or between the FLAC original.

Quality:
FLAC = 320kbps MP3/128kbps opus > 256kbps MP3 > 80-96kbps opus

Size:
Opus > MP3 > FLAC
There was no competition to differentiate them.

Usability:
MP3 > FLAC > Opus
Really needs better support.

Attached: literally you.png (614x1211, 145K)

>pushing towards lossless quality in lossy files
FLIF does an interesting thing about this. It follows the policy of "lossiness should be in the encoder, not the format", which IMO very wise. It means that even when an image loses information from the original you can still edit it losslessly. Imagine if the JPEG group had made the same choice.
youtube.com/watch?v=jjhomJ04S18
youtube.com/watch?v=W66XO2pNLkQ

What's the format and how do I make an opus file? Audacity doesn't show

>we could have filters run on top of codecs
If you really push this, won't you be reinventing Flash or just any sandboxed executable format?
>people apply noise to their image for a certain feel or look, but it can completely destroy the image when compressing it
AV1 actually filters out film grain before compressing the video then synthesizes similar film grain on decompression.

>it will take another decade at least until consumers get a transparent video format
Not with the rising resolutions. Personally, I'd much rather watch BD 1080P than YouTube's 4K, but it's always been harder to sell normalfags quality than megapixels. There is also the matter of streaming replacing optical media. Home Internet connections will probably max out at 1 to 10 Gbps for a long time.

Might as well ask here.
So I stumbled upon this bug while working on a script: trac.ffmpeg.org/ticket/5718 . As a workaround I want to use Vorbis, for all audio streams where this bug prevents me from using Opus without additional filters. But I'm not sure how to effectively detect them. Would something like using Vorbis for every audio stream with 6 or more channels work? Or could this happen with less channels as well?

What does frame size do for opus?

It's used by pretty much every major platform that has to stream audio, such as YouTube, Spotify and even Discord. Google pushes Opus support like crazy. If you're a big tech company and your audio related products do not support Opus in 2018 you're really stupid.

>Being this dumb

I could swear it is after that 1.4-something update they were raving about how clear and transparent it is low bitrates
what version of opus is on that graph

meant version 1.2

>the onion used to be funny before it was bought out by univision to be used as a propaganda outlet
fuck politics. We wouldn't have to put up with this nonsense if it weren't for all these bullshit "elections" they keep ramming down our throats every few years.

>AV1 actually filters out film grain before compressing the video then synthesizes similar film grain on decompression.
That's neat as fuck.

>the final solution
I believe it is!

Attached: 1535476392280.jpg (960x846, 119K)

>lossy

Attached: ccc1457ff4e5896995e209459ce0f456_width-600.jpg (600x789, 55K)

You are fucking retarded please do not post about things you know jack shit about retard.

Who are you quoting?

found the autistic "audiophile" who doesn't know jack shit about how information is stored and compared.

If the FLAC and MP3 have the same bit-depth and sample rate, the only difference in quality is the values for the frequencies. The only way to really lose quality in an MP3 is by complicated information that requires more data than the bite rate allows.

>bit depth and sampling rate are the most important metrics to judge the quality of a codec on

>The only way to really lose quality in an MP3 is by complicated information that requires more data than the bite rate allows.
For someone trying to appear knowlegable about a subject you sure are terrible at explaining things.
But I'd go for my earlier statement that is;
You're retarded and don't what you're talking about

Attached: snap.png (846x585, 24K)

For someone trying to appear knowledgeable, you haven't told me how I'm wrong, is this bait?

Yeah no sorry m8 I was wrong both 16/44.1 Flacs and mp3's are the same you won this argument.

spotify uses vorbis though, not opus...

>a USB

The bit rate refers to how much data is required to produce one second of audio. A lossy audio codec reduces the amount of information required to produce one second of audio. This is achieved by removing information from the raw audio stream, and supplementing it with approximate data. Each codec approximates this data in a different manner, each trying to produce audio as close (audibly) to the original stream as possible.

Yes, that's not really different from what I said. However it's more about compression than approximates, and those approximates, like I had said, brought around an approx. 1% difference, and a 1% change is very insignificant. What causes the larger amounts of approximations, is the complexity of the information that needs to be compressed and stored into the bit rate limit of the MP3 file.
We won't see any difference in audio quality from an audio stream at a specific frequency, but once you start compressing far more complicated streams of information, we start getting more approximations due to that limit.