So it seems Spotify is going to get hi-fi-quality streaming.
According to multiple news sites, the music streaming site is trialling lossless CD-quality sound, which it will ultimately offer to subscribers for an extra fee.
Sounds great, right? Well possibly – because this wouldn’t be the internet without an angry argument…
Indeed. Allow me to make some corrections to this article. (And it’s not surprising that the journalist whose words need to be corrected works for What Hi-Fi? magazine… One of such integrity.)
Spotify currently offers files at 320kbps. The bitrate of a CD is 1411 kbps. So lossless, CD-quality audio means a lot more information, which means more detail, better dynamics – all the good stuff that makes your favourite tunes sound, just, well, better. And now that extra information looks set to come to streams on Spotify.
This mathematical argument is often cited, but it’s simply incorrect. (Neil Young notably uses an argument like that to tout his Pono player.) Yes, a CD is 1411 kbps, and yes, Spotify uses 320 kbps MP3s. But that doesn’t mean that the quality is equal to 320/1411. That CD at 1411 kbps has no compression. Take an average lossless compressed file, and you get between about 500 and 800 kbps. (The range is because the amount of compression possible depends on the volume and density of the music.) So that 320 is a lot closer to the lossless bit rate. Also, lossy compression – MP3 and AAC – takes into account psychoacoustics, which removes or reduces information for sounds we cannot hear.
Now, while many may argue (and indeed Craig does above) whether they can really hear the difference between CD-quality and high-resolution audio — which delivers up to 9,216 kbps of information — the difference between CD-quality, lossless tracks and compressed, 320kbps files, really should be clear to anyone with a decent pair of headphones or a good music system at home.
Actually, most people can’t hear the difference. And, again, using the incorrect argument that more bits makes music sound better is simply wrong. While high-resolution audio can offer bit rates as high as 9,216 (actually, much more, but who’s counting), most high-res files at 24 bits, 96 kHz, come in at around 2,000 – 3,000 kbps (but, again, it depends on the volume and density of the music). And much of those bits are wasted, because there is audio in frequencies that you can only hear if you have magical bat ears.
Want to know more about high-resolution music? this episode of The Next Track podcast tells you all about it.
Musicians, producers, engineers and more, all currently work hard to deliver a piece of music to your eardrums. If you’re listening to a 320kbps stream, then in all honesty, you’re not hearing the full picture. A lossless, CD-quality file is much closer to “what the artist intended”.
Balderdash. Music is mastered for its delivery format, so a correctly mastered album designed for downloads and streaming will be mastered so it sounds best at that bit rate. And it’s never anything like “what the artist intended;” that’s just marketing. I recall the oft-cited anecdote about how the Rolling Stones, after they finished mastering Exile on Main Street, drove around Los Angeles in a car to hear how it sounded where most people would listen to it.
Learn about mastering on this episode of The Next Track podcast.
While picture technology has streaked ahead with HD, 4K and HDR, and surround sound has seen a host of advances from DTS to Dolby Atmos, music quality simply hasn’t kept pace.
Bunkum. The eye can clearly see the difference between HD and 4K, if you’re close enough to a TV. Beyond a certain distance, you simply cannot see the difference. With audio, it’s the same. If you have a very expensive stereo system, you may hear the difference between MP3s at 320 kbps and lossless, but it’s unlikely. If you’re listening outdoors, on the go, there is absolutely no point in having better quality. You’re just pissing away money.
But go on, if you think it’s better, then give them your money.
Source: The tech debate: Is it a big deal that Spotify’s going hi-fi? | Stuff
“Also, lossy compression – MP3 and AAC – takes into account psychoacoustics, which removes or reduces information for sounds we cannot hear.”
It depends on the program material. Complex material — such as full orchestra — is more-likely to show a loss in quality.
“Music is mastered for its delivery format…”
How is this done? If it’s possible, I doubt it’s a straightforward process.
“…so a correctly mastered album designed for downloads and streaming will be mastered so it sounds best at that bit rate.”
What do you mean by “sounds best”? Does that mean that lower compression will make the material sound //poorer// in some way?
A correctly mastered album will be mastered in such a way as to optimize the sound in the format for which it is mastered. The best example of this is Apple’s iTunes Plus format. Apple provides tools so record labels can test their recordings before and after conversion to that format, and tweak them so they sound as good as possible in that format.
That’s a largely tautological “explanation”. If a recording is “ideal” for one compression ratio, does that mean lower compression will degrade the sound in some way? That reasonable question needs an answer.
“Also, lossy compression – MP3 and AAC – takes into account psychoacoustics, which removes or reduces information for sounds we cannot hear.”
It depends on the program material. Complex material — such as full orchestra — is more-likely to show a loss in quality.
“Music is mastered for its delivery format…”
How is this done? If it’s possible, I doubt it’s a straightforward process.
“…so a correctly mastered album designed for downloads and streaming will be mastered so it sounds best at that bit rate.”
What do you mean by “sounds best”? Does that mean that lower compression will make the material sound //poorer// in some way?
A correctly mastered album will be mastered in such a way as to optimize the sound in the format for which it is mastered. The best example of this is Apple’s iTunes Plus format. Apple provides tools so record labels can test their recordings before and after conversion to that format, and tweak them so they sound as good as possible in that format.
That’s a largely tautological “explanation”. If a recording is “ideal” for one compression ratio, does that mean lower compression will degrade the sound in some way? That reasonable question needs an answer.
When I finally ponied up and bought a good set of earphones (Etymotics) that fit my ears, I was surprised at how good compressed audio files can sound. All this time I just wrote it off to the compression. Now I realize it was the cheap ear phones I always bought (Apple’s earbuds never fit my ears).
A year or so ago, NPR had an audio test on their website accompanying an article on some audio service and this whole debate of compressed vs uncompressed. I scored 50%. Ah well. So much for thinking I was a super hearer.
Joe
When I finally ponied up and bought a good set of earphones (Etymotics) that fit my ears, I was surprised at how good compressed audio files can sound. All this time I just wrote it off to the compression. Now I realize it was the cheap ear phones I always bought (Apple’s earbuds never fit my ears).
A year or so ago, NPR had an audio test on their website accompanying an article on some audio service and this whole debate of compressed vs uncompressed. I scored 50%. Ah well. So much for thinking I was a super hearer.
Joe
“the difference between CD-quality, lossless tracks and compressed, 320kbps files, really should be clear to anyone with a decent pair of headphones or a good music system at home.”
Maybe in theory, yes, everyone should hear it. In practice, most people (even hifi guys) don’t.
Did you ever do a proper double blind test?
Most people fail: https://www.heise.de/ct/artikel/Kreuzverhoertest-287592.html
Do your own test: http://abx.digitalfeed.net/
And here is why: https://people.xiph.org/~xiphmont/demo/neil-young.html
Extra: double blind test of expensive hifi stuff: http://www.hifi-forum.de/viewthread-18-1857.html#3 (tl;dr: most people, even those with golden ears, don’t hear any difference between 500$ and 10000$ amplifiers)
In my opinion, most people who bought a 1000$ Hi-Res Media Player and 2000$ headphones WANT to hear a difference, in order to justify their purchase.
Double-blind testing is a protocol. It is not “science”, and is of questionable utility. Nor is uncontrolled casual listening, especially with rapid switching. I prefer to listen for extended periods. If that doesn’t reveal differences, then there probably aren’t any significant ones.
There are amplifiers so bad you don’t have to compare them with anything else to hear how bad they are. The Crown K-series amplifiers are a good example, The estimable Gerald Stanley * designed them, and — forgive me — botched it. A former employer, who later worked at Crown for many years, told me that had he still been there, the K amps would never have gotten out the door.
I quit subjective reviewing in 1992, when my own listening revealed that “confirmation bias” //did not// apply (to me, anyway). More about that if you want to hear it.
PS: I generally respect “Monty”‘s discussion of digital principles. I also respect his recognition of the general superiority of Ambisonics for surround (having been the chief US cheerleader for that technology at one time). The “problem” with Ambisonics is that it actually works — which means it has to be set up correctly. This is difficult for most consumers, and in some cases, it doesn’t work at all!
* Stanley is credited with designing the first really good solid-state amplifier, the Crown DC-300A, He also (I believe) designed the PowerLine amps, which could achieve an almost 90dB input/output null with music.
“the difference between CD-quality, lossless tracks and compressed, 320kbps files, really should be clear to anyone with a decent pair of headphones or a good music system at home.”
Maybe in theory, yes, everyone should hear it. In practice, most people (even hifi guys) don’t.
Did you ever do a proper double blind test?
Most people fail: https://www.heise.de/ct/artikel/Kreuzverhoertest-287592.html
Do your own test: http://abx.digitalfeed.net/
And here is why: https://people.xiph.org/~xiphmont/demo/neil-young.html
Extra: double blind test of expensive hifi stuff: http://www.hifi-forum.de/viewthread-18-1857.html#3 (tl;dr: most people, even those with golden ears, don’t hear any difference between 500$ and 10000$ amplifiers)
In my opinion, most people who bought a 1000$ Hi-Res Media Player and 2000$ headphones WANT to hear a difference, in order to justify their purchase.
Double-blind testing is a protocol. It is not “science”, and is of questionable utility. Nor is uncontrolled casual listening, especially with rapid switching. I prefer to listen for extended periods. If that doesn’t reveal differences, then there probably aren’t any significant ones.
There are amplifiers so bad you don’t have to compare them with anything else to hear how bad they are. The Crown K-series amplifiers are a good example, The estimable Gerald Stanley * designed them, and — forgive me — botched it. A former employer, who later worked at Crown for many years, told me that had he still been there, the K amps would never have gotten out the door.
I quit subjective reviewing in 1992, when my own listening revealed that “confirmation bias” //did not// apply (to me, anyway). More about that if you want to hear it.
PS: I generally respect “Monty”‘s discussion of digital principles. I also respect his recognition of the general superiority of Ambisonics for surround (having been the chief US cheerleader for that technology at one time). The “problem” with Ambisonics is that it actually works — which means it has to be set up correctly. This is difficult for most consumers, and in some cases, it doesn’t work at all!
* Stanley is credited with designing the first really good solid-state amplifier, the Crown DC-300A, He also (I believe) designed the PowerLine amps, which could achieve an almost 90dB input/output null with music.