[Earwax]

Well, I've picked up my project again, this time comparing ogg vorbis to flac.

I choose 2 more albums, Guide My Feet and Circle of Drums and used dbpoweramp to create ogg lossy compressed files.

I started with -q6 which is variable bit rate around 192kbps. I was hearing different things with different headphones so I settled on using HD580s for the comparison. Which also means I have to use an amp.

I heard no "artifacts" with Q6, they might or might not be there, I haven't the patience to compare bit by bit. What I do hear is the sense of air and space in the flac compared to the ogg. Fairly easy to distinguish. [Side note - One thing I noticed, the Ogg tracks have a pleasant smoothness to them (reminds me of ATRAC compression) which makes them more pleasant compared to MP3 at a similar bitrate.]

I moved on to Q7, about 224kbps, again there's more "3D-ness" to the flac recordings. I went 6 for 6 in correctly identifying the lossy from lossless.
I just load the Clip with Q8. It will probably be tonight at the soonest before I make any listening trials.

 

[upstateguy]

FWIW, I have 2 Clips and a Fuze. I have always felt that both the Clips and the Fuze sound better with 320MP3s than they do with FLAC level 5. As a control, compare a WAV rip to 320MP3 and FLAC.

YMMV

USG

 

[Earwax]

WAVs don't have ID3 tags do they? So they would look different on the display.

If I decompressed both ogg and flac back to wav, I'd be stuck without any way to ID one from the other except the file name and that would be visible on the display.

If the flac and lossy files weren't converging in sound as the lossy bit rate gets higher, then I'd be more concerned about finding a way to practically implement this.

 

[iriverdude]

wav has no tags. But I think if you leave a .cue file in the directory your player should pick out details and display them. Also when converting wav to another format, tagging will be correct, rather than just saying track 01

 

[upstateguy]

Quote:

Originally Posted by Earwax /img/forum/go_quote.gif WAVs don't have ID3 tags do they? So they would look different on the display. If I decompressed both ogg and flac back to wav, I'd be stuck without any way to ID one from the other except the file name and that would be visible on the display. If the flac and lossy files weren't converging in sound as the lossy bit rate gets higher, then I'd be more concerned about finding a way to practically implement this.

I was suggesting a fresh WAV rip rather than a decompressed FLAC or OGG.

Just compare the same track in 320mp3, WAV and FLAC and see what you think.

Then if you really want to be depressed, compare what you just heard to the 320mp3, WAV and FLAC on your HD.

I also remember that there was a small volume difference, with the FLAC not playing as loud as the WAV or the 320mp3.

Let me know what your findings are.

USG

 

[ilney]

Quote:

Originally Posted by b0dhi /img/forum/go_quote.gif I think this is an excellent experiment, but it needs some more statistical power to be conclusive. Here's what I would do if I had 2 players: 1. (Get somebody else to do this step) Load the music into the players. Since you seem to be able to tell V3 easily, I suggest using V2 VBR as the lossy format to start with. Wrap the players in socks and shuffle them in a closed box. Label one of the players A and the other one B. Note down which player has the lossless. 2. Listen to music on them for 2 weeks. At the end of the 2 weeks, write down which player sounds better. If you more often instinctively reached for one player over the other, write that down. 3. Repeat from step #2. Over a few months you should get a reliable set of data.

Well, you know what player has loseless beforehand. This isn't fair.

Here's what I did:

I loaded 10 tracks in my DAP, 4 of which were the same songs in wmv and v0 lame (2 lame, 2 wmv). 6 others were different songs. The playlist looked something like:

Song A, v0, B, wmv, C, D, wmv, E, v0, F.

Twice a day, I listened in shuffle mode. Whenever one of those 4 played, I listened until 2 seconds toward the end and turned off shuffle mode so that I could know if it was v0 or wmv.

I failed miserably. And I doubt anyone would succeed. Remember: only twice a day.

 

[Earwax]

Quote:

Originally Posted by upstateguy /img/forum/go_quote.gif I was suggesting a fresh WAV rip rather than a decompressed FLAC or OGG. Just compare the same track in 320mp3, WAV and FLAC and see what you think. Then if you really want to be depressed, compare what you just heard to the 320mp3, WAV and FLAC on your HD. I also remember that there was a small volume difference, with the FLAC not playing as loud as the WAV or the 320mp3. Let me know what your findings are. USG

The reason I was thinking about decompressing a mp3 or ogg to wave would be to eliminate any possible playback idiosyncrasy. Whatever was lost during inital compression is still lost but the player will treat both tracks the same.

My PC sound card is not very good, I'm not too interested in testing with it.

 

[Earwax]

Quote:

Originally Posted by ilney /img/forum/go_quote.gif Well, you know what player has loseless beforehand. This isn't fair..

If they are shuffled fairly and look identical then you don't know which is which.

Quote:

Originally Posted by ilney /img/forum/go_quote.gif Here's what I did: I loaded 10 tracks in my DAP, 4 of which were the same songs in wmv and v0 lame (2 lame, 2 wmv). 6 others were different songs. The playlist looked something like: Song A, v0, B, wmv, C, D, wmv, E, v0, F. Twice a day, I listened in shuffle mode. Whenever one of those 4 played, I listened until 2 seconds toward the end and turned off shuffle mode so that I could know if it was v0 or wmv. I failed miserably. And I doubt anyone would succeed. Remember: only twice a day.

Without I reference, I couldn't tell.

 

[Earwax]

Quote:

Originally Posted by upstateguy /img/forum/go_quote.gif Just compare the same track in 320mp3, WAV and FLAC and see what you think.

I tried one song in flac, wav, ogg Q10, and lame V0. I do not hear any volume difference. Nor any other difference really (so far anyway)

 

[ilney]

But that's the point. I, however, can tell if someone gives me something below 192 kbps, w/o reference (unless the original recording blows really hard, happened once)

 

[Punnisher]

I'm a bit late in the thread, but it's an interesting test.

Most tests I hear of draw conclusions from mere minutes of testing. Taking this the extra step and truly experiencing the music over a long period might yield some surprising results.

It's almost like comparing an SR60 to a 325i with only one song played on each. Conclusions can't be accurately drawn because you haven't had ample time to learn every facet of the sound.

Try it and see if you can know (not guess) which is which. Rinse and repeat.

 

[Ferrariman]

Most people can't ABX a V2 MP3 from a FLAC. Some people say they can hear the difference, and they either have sensitive ears or aren't willing to admit they can't tell the difference, and claim the FLAC sounds better for whatever reason.

 

[Berlioz]

So, its impossible to tell the difference between 320kbps and FLAC? Hmm....

I swear I've done it before, but it could just all be in my head. Well, I'm gonna test this too I think. Once it's all set up, I'll get back with the results.

 

[ilney]

Try this:

Load a symphony with 4 movements in lossless. Ask one of your friends to convert 3,2,1 or none of them into V0 lame. Test in 1 week.

If there is a difference, you should easily tell which of the 4 is the V0 or none is. This is ,in theory, fairer and easier than ABX test. Why is it fairer? Because there are too many uncertainties that inhibit guesses. Why is it easier? Because if there is a difference, it would immediately create discrepancy.

 

[krmathis]

Quote:

Originally Posted by Berlioz /img/forum/go_quote.gif So, its impossible to tell the difference between 320kbps and FLAC? Hmm....

Not at all!
It all depends on the audio file, the encoder, playback gear, and not to forget your ears...