mp3 licensing ends for Music services

[vanished]

Just now, -TesseracT- said:

Yes, OGG is better. When I was deciding on what format and encoding settings to use for my locally stored music on my phone, I found that I could notice an enormous difference between AAC ~ 160 kbps and OGG ~ 160 kbps. The OGG sounded like the FLAC whilst the AAC sounded like an old 128 kbps MP3. It was incredible to hear a difference as I didn't believe that there would be one. OGG is clearly better!

I'm not sure everyone would describe the three of them in that manner, but I have heard that ogg is better than mp3.  However, it's not alone in that.  What I like about it is it's open and free

[Misanthrope]

4 hours ago, M.Yurizaki said:

https://arstechnica.com/gadgets/2016/11/netflix-4k-streaming-pc-kaby-lake-cpu-windows-10-edge-browser/

 

You can blame the studio industries for coming up with some asinine DRM requirement that requires both hardware and software support. Which is why Kaby Lake is currently the only processor capable of doing 4K.

 

I guess Microsoft not backporting support to Windows 7 and 8 for Kaby Lake is also a thing, but that's also partly the fault of Intel and other manufacturers for not wanting to support those OSes either.

 

tl;dr, solely blaming Microsoft is silly.

Sorry but I doubt companies said "make it compatible with as little devices and configurations as possible here, we're trying to avoid sales after all!" 

[Zodiark1593]

50 minutes ago, Misanthrope said:

Sorry but I doubt companies said "make it compatible with as little devices and configurations as possible here, we're trying to avoid sales after all!" 

The PC market is a much smaller market for targeting video to, and historically, where the weaknesses are found in DRM schemes (software based solutions must have the encryption keys in memory). I don't see studios losing much sleep over dropping the PC for 4K video, unless the influential powers in the PC industry were to convince them otherwise. 

[Bleedingyamato]

12 hours ago, M.Yurizaki said:

https://arstechnica.com/gadgets/2016/11/netflix-4k-streaming-pc-kaby-lake-cpu-windows-10-edge-browser/

 

You can blame the studio industries for coming up with some asinine DRM requirement that requires both hardware and software support. Which is why Kaby Lake is currently the only processor capable of doing 4K.

 

I guess Microsoft not backporting support to Windows 7 and 8 for Kaby Lake is also a thing, but that's also partly the fault of Intel and other manufacturers for not wanting to support those OSes either.

 

tl;dr, solely blaming Microsoft is silly.

Do you mean using its iGPU or 4K some other way?

[Misanthrope]

7 hours ago, Zodiark1593 said:

The PC market is a much smaller market for targeting video to, and historically, where the weaknesses are found in DRM schemes (software based solutions must have the encryption keys in memory). I don't see studios losing much sleep over dropping the PC for 4K video, unless the influential powers in the PC industry were to convince them otherwise. 

If that was the case they wouldn't even bother to try and get 4k content on pcs. Sorry but enough people have things like laptops and such that there is such a market. Microsoft was just probably fairly adamant at pushing this with a litany of fucking excuses as to why it needed to be only newer hardware because well, they benefit from the Windows 10 thing and intel and AMD benefit from the stupid hardware requirements too.

 

In fact what you say makes it more plausible: since the content creators care a bit but are not too concerned as you say they probably said "Yeah whatever" at MS and Intel schemes.

[SCHISCHKA]

mp3 sucks. iv been using FLAC to store CDs for while. The sound quality is so much superior and most cheap chinesium portable music players can handle them.

 

https://en.wikipedia.org/wiki/FLAC

[Trixanity]

15 hours ago, Misanthrope said:

Sorry but I doubt companies said "make it compatible with as little devices and configurations as possible here, we're trying to avoid sales after all!" 

Sorry but it's true. If they can avoid one instance of piracy they'll do anything because they're out of touch.

 

They dictate the rules and have their HDCP 2.2 requirement (and more). To be compliant every piece of hardware in the chain must support it and the stars must align, then you might be able to play DRM content. There is little doubt Microsoft probably orchestrated the Windows 10 requirement so they could force users onto the newer platform but it's probably also more 'DRM-friendly' so I'm guessing their partners are fine with it. Microsoft could probably backport it but won't. Only Kaby Lake, Pascal (and certain Maxwell) and Polaris are compatible in hardware; then there's the software that needs to be updated and adhere to strict guidelines to be compliant.

 

Read the Anandtech article on it. It explains it quite well.

[mariushm]

They stopped licensing MP3 because all patents have expired so they don't have any rights to ask for money anymore.

AC3 patents will also expire soon (this year if I read correctly, probably in a few weeks or months) but AC3 is somewhat worse than MP3 (but unlike mp3, it supports 5.1)

 

MP3 could now be a legitimate output audio codec for open source applications when you want most compatibility with older hardware... there's nothing stopping you from using MP3 at 256-320kbps  in some old portable player that can't decode flac/opus/aac or to upload your videos to Youtube

You could also use FLAC if you have the bandwidth.  So you have good enough and lossless compression covered.

 

An 128 kbps AAC is pretty much the same as 192..224 kbps MP3 and it's not that much extra data.

 

Don't forget there's also Opus audio (open source lossy codec better than mp3 or aac at bitrates like 128-160kbps) and Youtube and most software players support it already.

 

[mariushm]

 

@Ryan_Vickers

@-TesseracT-

 

Ogg is considered basically "deprecated", everyone is supposed to replace it with Opus audio which is much better.

 

Ogg as a file format is also kinda lousy, with lots of quirks and silly design choices. Shouldn't be used anymore

 

Based on various tests, AAC still produces better quality compared to Ogg at low bitrates (for example as you can see in the picture, better at 48kbps per channel or 96 kbps for stereo audio) but then the difference in quality is much smaller the closer you are to 96-112 kbps per channel (192 kbps - 224 kbps for stereo encoding)

 

See http://opus-codec.org/comparison/

 

 

 

Here's another test , scroll down to the first chart for comparison of various codecs at 96 kbps VBR : http://listening-test.coresv.net/results.htm

Opus is better than AAC and Ogg  and between AAC and Ogg, Ogg wins some but loses some as well (it depends on music genre)

 

[LAwLz]

Just want to tell everyone that:

OGG is not an audio codec!

 

An ogg file can contain several different files. For example you can put FLAC audio in an .ogg file if you want. You can even put video (like Theora) inside it.

When talking about audio formats you should preferably refer to the codec such as Vorbis or Opus.

 

If you ask me, I think everyone moving to Opus for everything would be great. One format for all applications (except FLAC for lossless of course). Vorbis is great and all, but the things it has over Opus are just unnecessary, while Opus got some massive benefits (mostly latency).

 

19 hours ago, Taf the Ghost said:

Hopefully this is a prelude for mp3 to become an actual "open" standard in a few years.  Not that it's going anywhere, anytime soon.

Sadly, this will probably just make MP3 even more popular. Before the patents expired (not that they mattered outside of the US anyway) one of the (several) drawbacks of MP3 was that you had to license it. Now that drawback is gone.

You still shouldn't use it because it's bad, but now we have one less reason to avoid it.

 

 

 

17 hours ago, ashypanda said:

seems like MP3 will be dead, as spotify and google play music both use OGG

I am fairly sure Google Play Music actually uses MP3. At least that's what you get when you download the songs.

17 hours ago, ashypanda said:

apple music is using AAC (obviously)

Apple did not make AAC if that's what you are implying.

 

 

 

17 hours ago, M.Yurizaki said:

You can blame the studio industries for coming up with some asinine DRM requirement that requires both hardware and software support. Which is why Kaby Lake is currently the only processor capable of doing 4K.

It was Microsoft that developed the DRM to begin with...

17 hours ago, M.Yurizaki said:

I guess Microsoft not backporting support to Windows 7 and 8 for Kaby Lake is also a thing, but that's also partly the fault of Intel and other manufacturers for not wanting to support those OSes either.

There is 0 indication that Windows 7/8's lack of support for Kaby Lake and Ryzen is caused by Intel/AMD refusing to support those OSes. Everything seem to indicate that Microsoft put themselves in the middle and blocked support, which caused Intel and AMD to say they don't support it.

17 hours ago, M.Yurizaki said:

tl;dr, solely blaming Microsoft is silly.

I'd say Microosft is 80% to blame. So not 100%, but the vast majority.

They enabled it to happen and carried it out. They might have done so at the behalf of movie studios, but that does not mean they should not get the majority of the blame.

[-TesseracT-]

2 hours ago, mariushm said:

An 128 kbps AAC is pretty much the same as 192..224 kbps MP3 and it's not that much extra data.

I find 128 kbps AAC to sound much worse than 192 kbps MP3.

[vanished]

4 hours ago, LAwLz said:

Just want to tell everyone that:

OGG is not an audio codec!

 

An ogg file can contain several different files. For example you can put FLAC audio in an .ogg file if you want. You can even put video (like Theora) inside it.

When talking about audio formats you should preferably refer to the codec such as Vorbis or Opus.

Yeah sorry, I should have said vorbis... I know that and the ogg theora for video but haven't used it in a while  

 

Didn't know about the flac thing tho, interesting  

[mariushm]

Here's a long list of reasons why OGG as a container is really "the suck" : http://hardwarebug.org/2010/03/03/ogg-objections/

 

Yeah, the most often used codec in an ogg is Vorbis audio codec... my bad.

 

FLAC as a container for FLAC audio codec is also not that great but it's way better than ogg. Basically, developers of the FLAC format stupidly decided to store some information in the header of each FLAC audio frame as bit fields or in 1-2 bytes (instead of just using 4 bytes) so some parameters are limited to very small valid ranges, instead of allowing for extensions to the standard in the future without affecting the frame header structure of each audio frame. This imho was stupid and pointless considering a header is something like 32-128 bytes out of 10-50 KB.. they saved a few bytes but messed up the audio container format.

 

Opus streams are technically in ogg containers, but you can repackage the raw stream in MKA (matroska audio), MKV or WEBM easily and avoid the problems of ogg containers..

[Andreas Lilja]

Meh, not into re-ripping my hundreds (400+) of CDs. 320kbps sounds good enough, tbh.

[LAwLz]

4 hours ago, Raskolnikov said:

Meh, not into re-ripping my hundreds (400+) of CDs. 320kbps sounds good enough, tbh.

That's why you should rip to FLAC straight away.

[kirashi]

On 5/12/2017 at 11:46 AM, Ryan_Vickers said:

why not just OGG?  Stop jumping from one sinking ship to another...

On 5/12/2017 at 0:26 PM, -TesseracT- said:

Yes, OGG is better. When I was deciding on what format and encoding settings to use for my locally stored music on my phone, I found that I could notice an enormous difference between AAC ~ 160 kbps and OGG ~ 160 kbps. The OGG sounded like the FLAC whilst the AAC sounded like an old 128 kbps MP3. It was incredible to hear a difference as I didn't believe that there would be one. OGG is clearly better!

On 5/12/2017 at 0:27 PM, Ryan_Vickers said:

I'm not sure everyone would describe the three of them in that manner, but I have heard that ogg is better than mp3.  However, it's not alone in that.  What I like about it is it's open and free

Spotify uses OGG Vorbis for streaming audio, which is supposedly better quality at the same kbps as MP3. This holds true as many people also state that a 160 kbps Vorbis file can sound as good as a 320 kbps MP3 file. Of course, this is subjective to the listener, but I'm also inclined to agree here since I switched from my own personal library of CD rips encoded at MP3 V0 to Spotify streaming.

 

If you have Premium, it gets even better because the stream bitrate goes up to 320 kbps OGG Vorbis, which is supposed to be in between the quality of 320kbps MP3 and 768 kbps FLAC encoded files. I don't have a proper stereo setup to test this theory, and almost all headphones (save for expensive good quality Studio Monitors driven through a neutral sounding DAC and AMP) will never allow us to hear the full fidelity of FLAC anyway, but it does make sense when you compare the spectrograms of each file format.

[vanished]

2 hours ago, kirashi said:

Spotify uses OGG Vorbis for streaming audio, which is supposedly better quality at the same kbps as MP3. This holds true as many people also state that a 160 kbps Vorbis file can sound as good as a 320 kbps MP3 file. Of course, this is subjective to the listener, but I'm also inclined to agree here since I switched from my own personal library of CD rips encoded at MP3 V0 to Spotify streaming.

 

If you have Premium, it gets even better because the stream bitrate goes up to 320 kbps OGG Vorbis, which is supposed to be in between the quality of 320kbps MP3 and 768 kbps FLAC encoded files. I don't have a proper stereo setup to test this theory, and almost all headphones (save for expensive good quality Studio Monitors driven through a neutral sounding DAC and AMP) will never allow us to hear the full fidelity of FLAC anyway, but it does make sense when you compare the spectrograms of each file format.

here's a question for you: If FLAC is lossless, why do people quote the bitrate?  It wouldn't matter would it?  

[kirashi]

17 minutes ago, Ryan_Vickers said:

here's a question for you: If FLAC is lossless, why do people quote the bitrate?  It wouldn't matter would it?  

Technically correct - a proper FLAC+CUE CD rip using default settings is fine at whatever FLAC bitrate the CD ripping software decides is good enough since FLAC is still technically a compressed format. So for lamens people, bitrate of FLAC files really doesn't matter.

 

However, for technically minded people, you happen to rip an acoustic guitar CD, there's a lot more room for compression because the audio won't be peaking as much, so it makes sense that the CD ripping software might be able to choose a lower bitrate for certain parts of songs. (FLAC is usually variable bitrate, but isn't always reported as such.) When your music playback software decodes the FLAC files, it technically plays back at 1411 kbps (same as CD PCM WAV audio) but all those gaps during the quiet parts of acoustic songs don't need to be decoded because there was nothing there to begin with. This is hugely oversimplified, but the way I understand FLAC to work.

https://www.head-fi.org/f/threads/why-do-my-flac-files-show-different-bitrates-in-winamp.548655/

https://sound.stackexchange.com/questions/24680/why-does-a-flac-file-have-a-lower-bit-rate-than-the-cd-file

[-TesseracT-]

2 hours ago, kirashi said:

Spotify uses OGG Vorbis for streaming audio, which is supposedly better quality at the same kbps as MP3. This holds true as many people also state that a 160 kbps Vorbis file can sound as good as a 320 kbps MP3 file. Of course, this is subjective to the listener, but I'm also inclined to agree here since I switched from my own personal library of CD rips encoded at MP3 V0 to Spotify streaming.

 

If you have Premium, it gets even better because the stream bitrate goes up to 320 kbps OGG Vorbis, which is supposed to be in between the quality of 320kbps MP3 and 768 kbps FLAC encoded files. I don't have a proper stereo setup to test this theory, and almost all headphones (save for expensive good quality Studio Monitors driven through a neutral sounding DAC and AMP) will never allow us to hear the full fidelity of FLAC anyway, but it does make sense when you compare the spectrograms of each file format.

320 Mp3 > 160 ogg. Whilst listning to music encoded in 160 ogg yesterday, I could, atleast I think I could, hear some compression artifacts. I may test this soon and jump over to 192 ogg.

[vanished]

Just now, -TesseracT- said:

320 Mp3 > 160 ogg. Whilst listning to music encoded in 160 ogg yesterday, I could, atleast I think I could, hear some compression artifacts. I may test this soon and jump over to 192 ogg.

It's easy to be swayed by what you think will be better.  To really test accurately, you need to A/B the tracks, record what you think, then check what they actually were, and only if you can reliably get the same result can you actually tell the difference.

[kirashi]

4 minutes ago, -TesseracT- said:

320 Mp3 > 160 ogg. Whilst listning to music encoded in 160 ogg yesterday, I could, atleast I think I could, hear some compression artifacts. I may test this soon and jump over to 192 ogg.

2 minutes ago, Ryan_Vickers said:

It's easy to be swayed by what you think will be better.  To really test accurately, you need to A/B the tracks, record what you think, then check what they actually were, and only if you can reliably get the same result can you actually tell the difference.

Exactly this - A/B tests are the only way to get meaningful results. If you know which format you're listening to, you'll always have some form of bias for or against it. That said, you also need a wide variety spanning multiple genres of music for testing too, since different styles of music have different frequency ranges.

 

Oh, also, don't be like Tidal and provide an A/B test that uses the EXACT SAME OGG files for both A and B version of the song. That's just a slimy way to convince people that Tidal Hi-Fi is better.  I hope they've smartened up since then, but back when they first launched the A/B test site to demonstrate why you should use Tidal, you could look in your browser's element inspector and grab the links for the A/B test OGG files they used, and see that they were identically the same.  

[vanished]

1 minute ago, kirashi said:

Oh, also, don't be like Tidal and provide an A/B test that uses the EXACT SAME OGG files for both A and B version of the song. That's just a slimy way to convince people that Tidal Hi-Fi is better.  I hope they've smartened up since then, but back when they first launched the A/B test site to demonstrate why you should use Tidal, you could look in your browser's element inspector and grab the links for the A/B test OGG files they used, and see that they were identically the same.  

 what!?  Really?  So how did that work, you listen to both, then click the one you think is better, and it told you that was tidal?  Like holding something behind your back and asking someone to pick which hand it's in, which obviously gives the opportunity to switch without them knowing before you show  

[ApolloFury]

On 2017-5-13 at 4:41 AM, Thony said:

Why ? Most of the population cant even notice a difference between 256kbps mp3 and FLAC....

People need to at least try 256kbps AAC compared to 320kbps MP3. A lower bitrate AAC actually sounds much cleaner with similar file size, I have trouble telling the difference between a 320kbps AAC to a medium quality FLAC.

 

http://www.kanyetothe.com/forum/index.php?topic=471869.0 

 

I don't like Apple products but when iTunes sells you 256kbps AAC for the same price as a Google Music 320kbps MP3 song, it makes Google Music look like a scam.

But if you want both 320kbps MP3 and AAC included in your purchase then that's where you buy from HDTracks for Americans, 7Digital for Europeans or ZDigital for Australians. They also sell FLACs of the same song for only a few cents more.

 

On 2017-5-13 at 5:26 AM, -TesseracT- said:

Yes, OGG is better. When I was deciding on what format and encoding settings to use for my locally stored music on my phone, I found that I could notice an enormous difference between AAC ~ 160 kbps and OGG ~ 160 kbps. The OGG sounded like the FLAC whilst the AAC sounded like an old 128 kbps MP3. It was incredible to hear a difference as I didn't believe that there would be one. OGG is clearly better!

 

19 hours ago, -TesseracT- said:

I find 128 kbps AAC to sound much worse than 192 kbps MP3.

No dude.... http://www.kanyetothe.com/forum/index.php?topic=471869.0 

You must not have a good AAC source in the first place.

 

22 hours ago, mariushm said:

An 128 kbps AAC is pretty much the same as 192..224 kbps MP3 and it's not that much extra data.

Exactly this ^

 

In the website I mentioned above (ZDigital or 7Digital) when you buy a song they give you both the MP3 and AAC versions. When you open both files in a program that analyse music files such as http://spek.cc/ you can see that the MP3 version often has a harsh cut at 15 kHz even at the same bitrate.

 

 

Even though that is a 256 MP3, a 320 MP3 file wouldn't be much an improvement, it will always have that harsh drop off at 15 kHz which makes higher pitch sounds such as vocals or classical instruments sounds muddy with a detailed headphone such as the ATH-AD series or AKG K series. And again, for the same price you can get 320kbps AAC when a 256kbps AAC is already very good.

[mariushm]

Your human ears won't hear anything above around 18kHz (unless you're a baby or very young or you recently went to the doctor to clean the wax from your ears) and your speakers can't accurately reproduce higher than maybe 20kHz .. majority of speakers are mid-range focused, they don't have tweeters capable of reproducing sound above 16-18kHz properly.

Everything about around 18kHz you mostly "feel" it, you don't hear it, just like with the bass.

 

The MP3 above was encoded with that filter enabled which reduces the "importance" of sounds above 15kHz in order to allocate more bits to more important parts of the sound.  The filter can be disabled but at a fixed 256kbps bitrate, the quality of the sound would be reduced.

It's probably disabled automatically by the encoder if you choose to encode at VBR 320kbps.

 

IMHO it's smarter to apply that filter and use the bitrate to retain more quality where the majority of human ears would actually perceive the quality loss.

AAC and MP3 are psycho-acoustic encoders, they work by dropping "detail" from the sound where they assume human ears won't easily detect those losses, so in theory what mp3 encoder did in that picture is actually good.

It's the same as with h264 encoders like x264 vs hardware encoders .. if you use PSNR or SSIM measurement tools to measure the quality of the output, in some scenes x264 may score worse but humans watching the video will prefer the x264 output because human eyes are not computers, they have flaws, and x264 is simply better than a hardware encoder at determining where your eyes would focus on and what detail would be missed by your eyes if it's removed or degraded to save bits.

 

Quote

here's a question for you: If FLAC is lossless, why do people quote the bitrate?  It wouldn't matter would it?

 

Indeed, it only matters in the context of streaming or when thinking how much music you'd be able to store on a hard drive or music player compared to the case when using other audio encoders.

 

Think of FLAC like a 7zip archive or RAR archive of the uncompressed wav file, but with some additional smartness inside like detecting if left and right channel have the same data and compressing that only once.

 

FLAC basically takes chunks of sound and compresses them losslessly into tiny "archives" that can be decompressed to recreate the sound. If there's no sound or just a person talking, that chunk of sound will compress better than let's say a second of hard rock music, which is more "random", harder to compress.  That's why the bitrate of a FLAC file varies.

 

Typically, a FLAC's file will be around 60-75% of the original bitrate - for 44100 kHz 2 channels 16 bit , that's 2 bytes x 2 channels x 44100 = 176400 bytes x 8 bits = 1411200 bits or 1411 kbps.

For stuff like audio books where often both audio channels have the same data (and sometimes they're encoded at 22050 kHz because human voices don't go over 10kHz and you only need 2x the frequency to properly encode audio), it's not uncommon to have FLAC files with bitrate in the range of 150-400 kbps.

 

 

[Zodiark1593]

17 hours ago, Raskolnikov said:

Meh, not into re-ripping my hundreds (400+) of CDs. 320kbps sounds good enough, tbh.

I rip to FLAC, for archival, then if desired for a mobile device, encode to a lossy format. 

 

1 hour ago, mariushm said:

Your human ears won't hear anything above around 18kHz (unless you're a baby or very young or you recently went to the doctor to clean the wax from your ears) and your speakers can't accurately reproduce higher than maybe 20kHz .. majority of speakers are mid-range focused, they don't have tweeters capable of reproducing sound above 16-18kHz properly.

Everything about around 18kHz you mostly "feel" it, you don't hear it, just like with the bass.

 

The MP3 above was encoded with that filter enabled which reduces the "importance" of sounds above 15kHz in order to allocate more bits to more important parts of the sound.  The filter can be disabled but at a fixed 256kbps bitrate, the quality of the sound would be reduced.

It's probably disabled automatically by the encoder if you choose to encode at VBR 320kbps.

 

IMHO it's smarter to apply that filter and use the bitrate to retain more quality where the majority of human ears would actually perceive the quality loss.

AAC and MP3 are psycho-acoustic encoders, they work by dropping "detail" from the sound where they assume human ears won't easily detect those losses, so in theory what mp3 encoder did in that picture is actually good.

It's the same as with h264 encoders like x264 vs hardware encoders .. if you use PSNR or SSIM measurement tools to measure the quality of the output, in some scenes x264 may score worse but humans watching the video will prefer the x264 output because human eyes are not computers, they have flaws, and x264 is simply better than a hardware encoder at determining where your eyes would focus on and what detail would be missed by your eyes if it's removed or degraded to save bits.

 

 

Indeed, it only matters in the context of streaming or when thinking how much music you'd be able to store on a hard drive or music player compared to the case when using other audio encoders.

 

Think of FLAC like a 7zip archive or RAR archive of the uncompressed wav file, but with some additional smartness inside like detecting if left and right channel have the same data and compressing that only once.

 

FLAC basically takes chunks of sound and compresses them losslessly into tiny "archives" that can be decompressed to recreate the sound. If there's no sound or just a person talking, that chunk of sound will compress better than let's say a second of hard rock music, which is more "random", harder to compress.  That's why the bitrate of a FLAC file varies.

 

Typically, a FLAC's file will be around 60-75% of the original bitrate - for 44100 kHz 2 channels 16 bit , that's 2 bytes x 2 channels x 44100 = 176400 bytes x 8 bits = 1411200 bits or 1411 kbps.

For stuff like audio books where often both audio channels have the same data (and sometimes they're encoded at 22050 kHz because human voices don't go over 10kHz and you only need 2x the frequency to properly encode audio), it's not uncommon to have FLAC files with bitrate in the range of 150-400 kbps.

 

 

Where I work has a 20 KHz ultrasonic welder. Sounds very horrible, though those next to me tend not to be bothered. Like anything else, hearing frequency probably varies to an extent, though I also dislike loud noises and music (including concerts and such), so it's probably just that I haven't blasted my hearing.