If you want to store as many songs on a portable device as possible, but also want to enjoy the confidence of future proof lossless format, why not give WavPack’s hybrid mode a shot? It allows you to encode a song into two files: one .wv file at any bitrate you set, and one .wvc correction file containing all the rest of bitrate data. When both files are present, it plays as a lossless format. If only the wv file is present, it plays as that bitrate. The only downside is that you need a portable device which support WavPack
I’m just going to be that guy and say that CDs are not exactly the highest standard themselves - vinyl is. It’s sort of difficult to tell the difference between vinyl and a raw vinyl rip, but it’s generally easy to tell the difference between either of those and a CD.
Not that that’s very practical, or that it matters much if you’re not using quality headphones (which I don’t do myself on a day-to-day basis).
I listened to all of them intently. I enjoyed all of them and wouldn’t notice the quality disparity in any. I’ll continue to live in blissful audio ignorance. I hope I never gain the competence necessary to be able to rank these in order.
As a normal person, I just use Spotify. Problem solved.
128kbps is definitely not OK, but after 192kbps, I couldn’t tell the difference really. And c’mon, just listen to the damn music.
For what it’s worth, I encode all my music into THREE formats (thanks to Max which makes it trivial to do on Mac OS X): FLAC, Apple Lossless and 320kbps MP3. Apple Lossless is in addition to FLAC because unfortunately some FLAC files from years ago are no longer readable with the newest FLAC codec. Storage? I have plenty.
I generally consider myself to have pretty good hearing - I am a musician and sell recording gear for a living. I listen through a nice recording interface (well, reasonably nice) and $500 (each) reference monitors. I could only really hear a difference when I popped on the $200 headphones. I am a little surprised, though I suppose it depends on the song. Most songs I can easily tell if it’s 128, but if it’s 256 or better there’s not much difference. Uncompressed just seems clearer, but I can’t really place the difference there. All the in betweens are a wash for me. To those who say it is silly to use a song with crappy drum machines - there is a difference between lofi analog and lofi digital. We are talking about digital compression artifacts here and the drum machines don’t have those.
I note that at least one person in the comments doesn’t understand what lossless means, he said that flac sounds “almost as good as” uncompressed.
To that person, think of flac as a zip for music, data goes in, exact same data (bit-for-bit identical, hence lossless) comes out.
So it doesn’t sounds “almost as good”, it sounds exactly the same.
I enjoyed this test a little too much.
It seems that too many years of listening to loud rock has dulled my hearing ability. I used to be able to tell the difference easily, by listening to high-pitched stuff (such as cymbal crashes) but no longer. They all sound the same to me. ;( Right, where did those cassettes go to…
I look forward to taking the test being an audiophile person, at least to my friends and family. I enjoy good sound, that is all.
I have a few reservations to this blog post however and those are one, do we really need to be so effecient when it comes to space these days? and two, am I the only one who actually likes browsing my CD collection and enjoy taking the physical thing out of the box and inserting it into the machine? I love the experience of really listening to an album!
It’s interesting to see what some believe are the answers. This was a test for distinguishing compression, not overall sound taste. I recall an older test that was similar to this and many people actually rated the highly compressed files as sounding perceivably better than lossless. I’ve trained myself to recognise the artefacts created by compression and when I listened on some really bad Logitech computer speakers (I’m too lazy to get out my good phones) I could still definitely tell the difference. I analysed the tracks (only after voting of course, I just couldn’t wait for the results!) and found that I was correct except I had the 2 worst reversed. It will be interesting to see the success rate from the poll.
A few points…
a) if you’re listening to these samples through your browser
b) if you’re listening to these samples via the Windows WDM mixer
c) don’t have your output sound device locked to the same output bitrate of the audio samples e.g. via ASIO
then you are not listening to bit-perfect audio, you are listening to interpolated (dithered) audio and all your audiophile arguments go out the window.
C,G,B,L,F 
One more vote for “they all sound the same”. Or close enough to the same for me not to care, at any rate. Maybe if it was a song I liked…!
Jeff cheated by using a poorly mastered recording. Notice the clicks in the beggining? If he used something from ECM, I would INSTANTLY tell the difference.
Hacker News had an article about this not so long ago:
http://people.xiph.org/~xiphmont/demo/neil-young.html
The short of it is you don’t need more that 44.1KHz/16 bit, uncompressed. Compression - well, that depends - lossy or loss-less, and what encoder.
Once question I’d ask it what is the objective? Good music, or perfect reproduction? ATRAC compressions was lossy, and two thirds of people could tell the difference - but two thirds of them preferred the ATRAC compressed version. I was very surprised by that when our lecturer told us that.
Difficult. I can detect differences between them. I remember I conducted a test between two files ack when choosing the bit rate to use for my library. I was comparing kasbians club foot between 192 and 128, and all of a sudden, click- I could hear the waterynessif the cymbals,
It is a lot more difficult comparing between 5 samples especially on iPad ( with earphones of course). I think they are in descending order- I think the tell tale sign is the cymbal crescendo around 1:12. It’s damn difficult- there is a difference- I could tell between two directly, maybe three, but five is just confusing me- I can’t hear enough slight difference to rank them!
FLAC or Lossless WMA for storage, and CBR0 MP3 for playback.
AAC is not useful because most car stereos can’t play it back (except if you have the stereo controlling an iPod/iPhone in head mode in which case the external device does the actual playback) and many portable devices can only handle AAC in LC (Low Complexity) mode.
It’s unfortunate that most d/l vendors don’t offer a lossless version, but I must applaud Hyperion Records, Qobuz store and a lot of artists on Bandcamp for giving FLAC/WAV options.
I think that the second point is the one that everyone. Is missing. While it’s nice to talk about how you are not using that much more space with something like flac, I have limited days on my phones data on my phone (2 gb per month currently) . Given the the cell networks current pricing scheme and business practice schemes (ie they’re never going to change that cash cow) every byte I stream over my phone now becomes.precious. so as media companies move "to the cloud " (as I can’t imagine them doing anything else to stay on business) that means that I won’t actually be able to have the luxury of local storage. Whether you like it or not, that’s where we’re all headed. That makes this a valid argument.
Jeff, why not use Ogg Vorbis instead of MP3 for lossy compression of your music collection? It’s a free, non proprietary standard, (so better for archival purposes) and modern Vorbis encoders offer higher quality at lower bitrates than MP3. With the AoTuV fork of the encoder, you would probably be fine encoding your collection at 128 or even 96 kbps.
Pauli said: MP3 = LOSS at any bitrate. 320 kbps is great for earbuds - that’s it. I will not waste time comparing cds to mp3. There is none. Buy vinyl and a decent table if you want to hear the recorded track in it’s purest form. Until you jarheads start MAKING digital music (and some crap I hear these days sounds like 2 notes over and over… 010101010) you will always need to convert to analog to hear it.
Uh huh. Purest form. Right.
Well, except for the noise, and who cares about wear?
Tell you what, how about we get a CD and Vinyl of the same thing, decently mastered*, and compare output waveforms, or do some ABX testing (with “fake record noise” added to the CD so it’s as dirty as a record is in the real world).
Vinyl is not pure - it’s just distortion you’re used to and prefer.
Then you can posture to me about vinyl after you pass the ABX test.
Problem is, audiophiles have historically been very bad at that:
http://www.bostonaudiosociety.org/bas_speaker/abx_testing2.htm
http://www.hydrogenaudio.org/forums/index.php?act=ST&f=21&t=7953
(* In fairness, some things these days are mastered with so much compression that they just suck, and the one indirect advantage vinyl has is that people mastering for vinyl never ever do that, but that’s irrelevant to the format itself.)
@Miles: Why create a lossy archive in OGG when almost every time you have to play it on another device it would have to be transcoded? It’s even less useful than AAC in that respect.