Hacker News Reader

Comment by gruez | Hacker News Reader

AFAIK spotify premium uses 320kb/s vorbis for audio compression. I understand why you'd want lossless for archival purposes, but does it make a difference for playback? ie. can you tell the difference between 320kb/s vorbis and lossless in a ABX test?

View on HN

I used to keep a local library of FLAC/ALAC lossless but in recent years have switched over to 256kbps AAC Apple Music for the stuff they have.

My audio setup isn't the best in the world, just a JDS Labs Element II and Sennheiser HD6XX, but for the stuff I listen to I have a hard time distinguishing between lossless and 256k AAC. I've ABX tested various bitrates of MP3s and can pick those out, but AAC eliminates the telltale artifacts that MP3s bring.

Might help that Apple streams "mastered for iTunes" copies for a number of albums which mandates less compression prior to encoding but I think it mostly boils down to AAC being better suited for music than MP3.

Usually it's not just the bitrate, but also the compression in mastering process that's different (see https://en.wikipedia.org/wiki/Dynamic_range_compression for context). The "HiFi" versions are less compressed and on my highend audio system the difference between Tidal HiFi masters and Spotify is pretty stark and obviously noticable.

(The difference obviously has nothing to do with compression format.)

I hope Spotify will match the quality because Spotify Connect is way more reliable than the Cast broken mess.

Exactly, it's literally a different sound file - not just format/delivery/etc.

Where is the evidence that Tidal is actually using masters that exclude DRC for their HiFi mixes?

I can't find anything that suggests this is actually true.

The only thing I can find indicates that their HiFi versions are based on MQA-supplied 24-bit masters, but that doesn't tell you whether or not DRC was applied to the source mix that MQA used.

And given that an uncompressed mix would require remastering by the artist/studio (DRC is not just a byproduct of conversion to a 16-bit format), you'll forgive me if I'm incredibly skeptical of this claim.

Does it make a difference? It certainly sounds different (on my very midrange Audioengine speakers).

I certainly can't consistently identify which one is lossless in a blind test, but there is an audible difference between the same song in Vorbis and FLAC.

I don't know how useful this distinction is though.

http://abx.digitalfeed.net/

Do you really hear a difference? My setup is probably really bad and I probably really suck but even 96 bit lame was indistinguishable from lossless to me, except for one part in one song.

Wow. I really can’t tell the difference at all. Like, I didn’t bother completing the test because it all sounded exactly the same. To the point where I doubted it was working and checked out the “Validate This Test” link to confirm it worked.

The differences are due to psychoacoustics -https://en.wikipedia.org/wiki/Psychoacoustics . Our perception changes depending on situation, time, you knowing what is played, etc. To say that you hear a difference you should succeed in a blinded test and without any delay between listening. And succeed not only once.

Not every passage of every song will produce a meaningful difference. The areas where a 24bit vs. 16bit file differ most are in subtle dynamic changes. For example, there was a specific ~ 2-second transition in a recording of Steve Reich where I could definitely tell the difference between a 96k/24bit file and Apple Music.

So I ran a half-baked experiment here. Had a friend queue up a song (15 in total) in "HD" and not, play 30 seconds of one, pause for a few seconds, then play the same 30 seconds again, randomly choosing between the same quality (placebo-is) and the other.

Then I guessed "same or different" quality. Got about 85% right overall.

I'm sure it wasn't perfectly blind, but seems to indicate to me that it does in fact sound different. Now, I cannot say for sure it's not sound cards or other things in the output path causing differences as well.

Now if you asked me to pick which was HD and which was 320, I did worse than random. So I am not claiming that HD sounds "better", just different.

Almost certainly not, but the audiophile community is a profitable one... :)

On AirPods, no. With audio set ups that cost few grand, yes.

The difference will not be audible to most people because of the equipment and environment. Besides, audio has been mastered for 128kbps MP3, which most people prefer. So unless they are going to change the mastering, this is double worthless.

But I’m not in the habit of telling people how to spend their money. If they enjoy the music more because they paid more for it, no objections here. I’ve got plenty of wasteful hobbies myself.

Reply on news.ycombinator.com