Item — HN Reader

I know this is controversial but personally, Spotify Premium is already hifi. I simply cannot hear the difference between 320kbit/s and higher bitrates. I have high-end audio equipment at home, a decent DAC, etc...once I go beyond Spotify Free's 160kbps I truly can't hear the difference.

I'm not calling anyone a liar, I'm sure some folks can. But I think it's worth considering how important "hifi" bitrates are to you beyond a bigger number.

9 comments

I'm no audiophile, but even when downloading the highest quality files and playing them locally Spotify is still noticeably worse quality than the default AAC files that Apple Music downloads for me on my purchased music. Spotify just sounds like there is no depth to any of the sound, where I can immediately feel the difference with Apple Music - especially with percussion heavy music.

I don't doubt that you feel the difference between the two - but that doesn't mean there is a difference.

Without a double-blind test, it's just conjecture. Audio is notoriously subjective, and existing research shows that 320kbps audio is indistinguishable from lossless.

And to be clear, this isn't an accusation that you're necessarily imagining differences in sound. It could even be something as simple as one source being very slightly quieter than the other (due to the player software rather than differences in the audio encoding).

A good test I typically use is this ABX test:

https://abx.digitalfeed.net/

> I'm no audiophile, but even when downloading the highest quality files and playing them locally Spotify is still noticeably worse quality than the default AAC files that Apple Music downloads for me on my purchased music.

That's simply impossible, the Spotify client does not change your local files, it simply plays through your OS audio stack. You are noticing the difference in loudness, you can't know if the files you downloaded have been mastered the same way than the ones Apple Music is playing.

I just tried with some of my own music, played the exactly same file as a WAV rendered from my DAW and then a very high bitrate export through the Spotify client as a local file, there's absolutely no difference.

Did you confirm this with blind testing?

Often it’s also just loudness/dynamic range compression.

Yeah you'd need to level match a couple of tracks, then blind test between them. Or else turn off normalisation.

Agreed. I release my own music on everything except Spotify as they do something horrible to them (and they're pretty awful people too).

Audio compression is based on psychoacoustic methods. Perhaps it's not your ears but different wiring in your brain that makes you capable of telling the difference.

That may be the case. I believe if you've listened to a track long enough, some of the minor details that lossy encoding remove are remembered and if you're listening closely enough, you notice their absence. My most recent experience was a particular moment in a song. I was listening to the lossy encoded version and it sounded weird to me. I grabbed the original file and that's when I noticed the difference. I believe someone else would have not noticed, but I've been listening to the lossless version and perhaps my brain has a vivid memory of that section. I don't notice when I'm playing with cheaper headphones and speakers.

Totally anecdotal, but I got slightly better speakers and could instantly tell the difference between a Spotify stream and Bandcamp-sourced FLACs or mp3s

The rumor I've heard is that often artists upload tracks to Spotify with a lossy codec, and then Spotify re-encodes as ogg for transport imposing a second loss on the encoding.

I don't know if it's psychosomatic or not, but with a decent amplifier and speakers the difference was actually pretty shocking. I am not an audiophile and I was not trying to do a comparison -- I merely bought an album on Bandcamp and listened to it and thought "huh, my Spotify streams sound kind of muddy compared to this"

Take that with a grain of salt, I guess

Similarly, on Apple Music with some tracks I can hear a difference when switching between AAC and lossless modes. In fact the reason I know this is because one day when listening to music, one particular track sounded unusually bad and switching on lossless fixed it.

Genre likely has some bearing on whether a person can hear the difference or not, with some genres being generally being more compressed for example which will reduce the gap between formats. It would also make sense that each codec has genres it’s more or less suited for.

I think that technically you can make lossy sound really good - almost indistinguishable.

I also think that they may fail badly at it - I've heard lossy formats that sound like crap, and lossless that sound similarly bad. There's a lot more that goes into it beyond just "what is the setting on the compressor".

I'm honestly glad to hear you say that. I have some lossless audio I bought off Bandcamp, so I'll set it up again and do another test. I want lossless music to matter to me, but so far it hasn't. It's also worth linking to this famous NPR quiz comparing MP3 bitrates [0]

https://www.npr.org/sections/therecord/2015/06/02/411473508/...

That NPR quiz has a confusing interface.

You can play the clips by pressing the play buttons next to them, which works as expected. If you stop a clip it resets so next time you click its play button it again starts from the beginning.

You can also click within the progress bar of a playing clip and it will jump to that spot, which is handy for repeating a section somewhere in the middle.

You can also click in the progress area of a clip when no clip is playing and it will start playing at the appropriate offset, but not necessarily the clip whose progress bar you clicked. It actually plays the last clip whose play button you pressed. It does not start the progress bar for that clip (or the clip you clicked in) so you won't see that it is not playing the clip you expected. You'll only notice it if the last play button you clicked was for a different song.

This completely messed up the quiz for me. I'd press the play button on each clip of a set and listen all the way through. Say I thought I heard some percussion 3/4 of the way through that seemed off on one of them. I'd then play just that part by clicking in the progress area of the three clip about 3/4 of the way in, thinking that is letting me hear that section of each clip. Instead it is just giving me that section of the third clip each time.

320Kb/s MP3 is fine for most things, and so is spotify quality in most cases. But there's also a bunch of bad tracks that I feel like is probably bad re-encodes, and sometimes bad source material. And yes, whether you can hear crunchiness depends on how you listen - you can't tell the difference over a car stereo when driving on a highway - and if you listen over bluetooth you're already hearing a bad re-encode, at least until BLE audio and LC3 is the norm - but IEMs in a home or quiet office is a different story and then it can get quite... distracting.

A launch of a lossless feature might also imply either a change of workflow to import tracks from artists, or at least just a re-upload of the source material. That I find exciting, and is something I would consider paying for.

If anyone thinks they can tell the difference, I encourage them to take this abx test comparing lossy/lossless streaming. I would be shocked if you pass:

https://abx.digitalfeed.net/

Almost no one can tell a difference with today's encoders (unless the track is particularly degenerate for the encoder) but I think there is an edge case: listening over Bluetooth.

When the already encoded audio is then encoded again into AptX or SBC or AAC, it starts to introduce audible artifacting - nothing most people would immediately notice but audible artifacting nonetheless.

I don't believe a DAC is going to do anything for you, with Spotify. I agree with you there... but saying Spotify is already hifi? Absolutely not.

My DAC plugged into Spotify does nothing for me, and I don't expect it to. Plug that DAC into Apple Music or Tidal, and put on a good pair of cans? Completely different world.

> Spotify Premium is already hifi

Hard disagree.

I do agree that for the majority of people, the 320kbit/s is absolutely enough for most people (arguably 160kbps is).

But it's not "hifi". And I don't mean "With extremely high end headphones and a dac, a trained ear, and holding your tongue at the right angle you can just barely make out a difference" - it's very noticable even with mediocre equipment. It's a rather lossy 320kbit/s too.

That's why a "hifi" tier makes sense, they don't currently serve that market at all.

I don't care about lossless, but I find the Atmos mixes are usually exceptional. I don't know if it's that they are producing a more "audiophile" mix, or if it's the format, or something else entirely, but it's night and day for me.

Often (but not always) they fall less victim to the “loudness wars” compared to the same track mixed for stereo.

I don't think that's true - the loudness wars haven't really been a thing for a few years now, the streaming services all "normalise" the loudness which makes pushing it in the master pointless.

There are very noticeable differences on some recordings. I think that’s probably a question of whether the Atmos version is a straight re-encoding or whether they took the time to do some other editing as well, which can make these comparisons really confusing since people will swear they’re talking about the same thing.

As I understand it, atmos / spatial recordings are a full re-master when it's done to older recordings so yeah, there's definitely going to be a big difference when compared to masters from years or decades ago.

Definitely - in addition to higher bit rates and depths I’m sure some people are tempted to use denoising or more advanced filtering, too.

I came here to exactly say this. I would not pay more per month to have a "better" bitrate.