The Audiophile Challenge

It's simple. I have uploaded a one-minute sample of "Dance V" by Philip Glass. Although the entire file is encoded in FLAC, it randomly switches between two different samples: A FLAC, and a 128kbps MP3. Both are exactly the same apart from the quality, and your job is to figure out exactly when it switches between the FLAC and the MP3.

If you really can't tell when it switches, congratulations! You're free! You can encode all your albums in 128kbps and have peace of mind. But if you notice the switches... well, maybe there's a good reason for FLAC encoding after all...

RULES:

Place ALL your guesses in spoiler tags

Ears ONLY (no computer analysis!)

Have fun!

Audiophile Challenge.flac 4.51 MB · 103 downloads

that's really hard:

You know, ears aren't necessarily the limiting factor if one is listening to some shitty earbuds plugged into the front of your desktop.

Also, you're not really free until you stick icepicks in both earholes and sever all bonds of aural maya (Krishna sez this somewhere).

Dance V kicks ass, just wanted to say that 8-)

lol

this test fails so much

Rhombix is pissed because he has always encoded his music in 128kbps.

oh this is interesting, never made this test with myself:

well?

  On 3/8/2011 at 9:11 PM, eugene said:

that's really hard:

 

  Reveal hidden contents

~0:08 switch from flac to mp3, 0:14 back to flac, 0:21 to mp3, 0:27 back to flac, 0:38 to mp3, 0:52 to flac

 

 

Wow, that's actually pretty close. You have a rough idea, for sure.

  On 3/8/2011 at 10:05 PM, Enter a new display name said:

Rhombix is pissed because he has always encoded his music in 128kbps.

I did it for 1 or 2 years when I first started collecting music, but I slowly started uploading my CDs in higher bitrates until now, where I usually use MP3 V2. I still have some music in 128kbps, but that's OK. I'll put up with 192kbps or higher when obtaining new music.

edit: you posted while I was writing

  On 3/8/2011 at 10:31 PM, plstik said:

oh this is interesting, never made this test with myself:

 

  Reveal hidden contents

flac until ~10s, flac again around 31s until 49s.

 

well?

Nice idea, but that track you posted is one of the worst possible examples you could have used for discerning compressed from uncompressed audio. It's a nice piece of music but it's absolutely shit for this kind of test. I'm definitely in the pro-mp3 camp, and I agree that it's much more difficult to tell the difference than people make it out to be, but there are very specific ways in which mp3s tend to show their faults: things like big splashy decaying cymbal hits, big layered percussive sounds, and ambiences with lots of sharp dynamic changes and complex harmonic content. This track is like the perfect antithesis of all of that. It's like trying to show off the quality of an HD television by playing old Disney cartoons.

yeah this is definitely a weird piece to choose for this test. I remember I was able to do it with some 192kbps vs wav test at some point, but it had a lot of tons of high range and low range, unlike this. Not that I really care. I listen to things in 128 if that's what's available, and it's 100% listenable for me. I rip/download/buy everything in flac though, because I can spare the space, and if I have to convert or re-render or whatever the case may be, I do not want to deal with generational loss if I can help it. Plus some people are snobby about DJs playing mp3s, so I try to avoid it when I can.

  On 3/9/2011 at 7:53 AM, MrSparkle666 said:

Nice idea, but that track you posted is one of the worst possible examples you could have used for discerning compressed from uncompressed audio. It's a nice piece of music but it's absolutely shit for this kind of test. I'm definitely in the pro-mp3 camp, and I agree that it's much more difficult to tell the difference than people make it out to be, but there are very specific ways in which mp3s tend to show their faults: things like big splashy decaying cymbal hits, big layered percussive sounds, and ambiences with lots of sharp dynamic changes and complex harmonic content. This track is like the perfect antithesis of all of that. It's like trying to show off the quality of an HD television by playing old Disney cartoons.

EXACTLY

  On 3/8/2011 at 10:33 PM, Rhombix said:

  On 3/8/2011 at 10:31 PM, plstik said:

oh this is interesting, never made this test with myself:

 

  Reveal hidden contents

flac until ~10s, flac again around 31s until 49s.

 

well?

 

at least I got the first part somewhat right :sup:

*enocdes himself to 128kb*

wank wank wank wank

I'd say something like :

  On 3/9/2011 at 7:53 AM, MrSparkle666 said:

Nice idea, but that track you posted is one of the worst possible examples you could have used for discerning compressed from uncompressed audio. It's a nice piece of music but it's absolutely shit for this kind of test. I'm definitely in the pro-mp3 camp, and I agree that it's much more difficult to tell the difference than people make it out to be, but there are very specific ways in which mp3s tend to show their faults: things like big splashy decaying cymbal hits, big layered percussive sounds, and ambiences with lots of sharp dynamic changes and complex harmonic content. This track is like the perfect antithesis of all of that. It's like trying to show off the quality of an HD television by playing old Disney cartoons.

whenever I play one of my high resolution dvd-audios I'm amazed at how great the cymbals sound.

PS : I did the test with some beyerdynamic DT 990 headphones

  On 3/8/2011 at 10:33 PM, Rhombix said:

I did it for 1 or 2 years when I first started collecting music, but I slowly started uploading my CDs in higher bitrates until now, where I usually use MP3 V2. I still have some music in 128kbps, but that's OK. I'll put up with 192kbps or higher when obtaining new music.

VBR makes a surprising difference

also, i think something like ae would have made for a better test

  On 3/10/2011 at 3:24 AM, xxx said:

Can't tell anything wit this but, in real-life file sharing, huge differences can be detected. I slsk'd "Hail To The Thief" in 2003 and then bought the CD. It was like I was listening to a completely different album. Same with "Temple Of Transparent Balls". In the underground, you can make anything say 192, 320, whatever but there is definitely something a lot deeper happening. Like the old melamine-in-the-milk trick that the Chinese pulled off.

There are a lot of just plain bad quality mp3s floating around out there. I don't know what encoders and settings people are using to get such shit results, but it happens a lot. If everyone just used LAME with VBR at settings above 192kbps, everything would be fine. It's poorly encoded mp3s that give the format a bad name. Using LAME at bit rates above 192kbps, you would be hard pressed to tell the difference between uncompressed and compressed audio. Double blind tests have been conducted many times. Even audiophiles usually have tough time telling the difference in a blind test.

The real culprit in this day and age is bit depth. 24bit audio is so vastly superior to 16bit it's like night and day, yet everything is still typically downsampled and released in 16 bit even though there is no fucking reason. It's just stupid. That's why DVD audio on movies sounds so much better than cds. It's the bit depth. We really need a shift in standards.

i can tell the diff between a flac vinyl-rip and a vinyl !

256kb encoding using iTunes is the shit

i can tell the difference between good and bad music

what about wrong music

I've created another version of this challenge, this time using 'Cichli' by Autechre. Same rules still apply, but HOLY SHIT did this surprise me. Just listen to this!

Audiophile Challenge 2.flacFetching info...

thanks. interesting.