Help - Search - Members - Calendar
Full Version: Listening to Lossless
Hydrogenaudio Forums > Hydrogenaudio Forum > Listening Tests
shokoshoko
I've just started listening to lossless audio and like some people, sometimes i have a hard time finding the differences between a flac and a mp3 file. I can tell the diffreneces between them with music that have more simple arrangements like acoustic, vocals or plain instrumental peices. I noticed that the lossless files sounds more open, is more layered ( a better soundstage i guess) and when listening to vocals, it sounds more brighter and more dynamic.

But it starts get harder to tell when i listen to music like pop, rock, music that are etc... Are there any ways or special characteristics about sounds from lossless files that help you guys determine the differences?
odigg
QUOTE (shokoshoko @ Apr 23 2009, 22:33) *
But it starts get harder to tell when i listen to music like pop, rock, music that are etc... Are there any ways or special characteristics about sounds from lossless files that help you guys determine the differences?


I think you should do a blind ABX test to figure out if you can really hear the differences or you just think you can. A lot of people think they can hear the differences between high rate MP3s and Lossless, but find that they cannot accurately identify which is which in a blind test.

You can do a blind test using a music player called Foobar 2000. http://www.foobar2000.org/ When you install it you will have an option of installing an ABX comparator.
Axon
For high bitrate comparisons you're going to be looking at lowpass effects, subtle changes in soundstage, pre-echo with percussion instruments, and quiet noises alongside relatively loud noises which exist in one version or the other. All of which are extremely difficult (and perhaps impossible) to tell at high bitrates.

Acoustic music is often times the easiest music for MP3 to encode, and rock can often be the hardest. So I would strongly encourage you to follow odigg's advice and run an ABX test to confirm your own sighted listening experiences.
greynol
QUOTE (shokoshoko @ Apr 23 2009, 19:33) *
I noticed that the lossless files sounds more open, is more layered ( a better soundstage i guess) and when listening to vocals, it sounds more brighter and more dynamic.

These are the types of descriptions that people typically use to describe differences they're only imagining.

http://www.hydrogenaudio.org/forums/index....showtopic=16295


shokoshoko
I did what odigg said and did a ABX test and my result was:

foo_abx 1.3.1 report
foobar2000 v0.9.6.3
2009/04/22 16:26:58

File A: C:\Users\Andrew\Music\Zune\??? (Lossless)\?? (Flac)\??.FLAC
File B: C:\Users\Andrew\Music\Zune\???\??\??.mp3

16:26:58 : Test started.
16:27:56 : 01/01 50.0%
16:28:16 : 01/02 75.0%
16:29:43 : 02/03 50.0%
16:31:05 : 03/04 31.3%
16:32:08 : 04/05 18.8%
16:33:02 : 05/06 10.9%
16:33:05 : Test finished.

----------
Total: 5/6 (10.9%)

So i guess i can tell it apart >.<
odyssey
Are both encodes from the same source?

Did you replay-gain them?
shakey_snake
QUOTE (shokoshoko @ Apr 24 2009, 10:07) *
So i guess i can tell it apart >.<
Your results are inconclusive. There's a 10.6% chance you were randomly guessing.

Generally, p-values need to be less than 5% (1/20) to be statistically interesting. Please re-read

Also note: you need to run at least 7 trials to get a very significant result (i.e. 7/7 p=0.8%)
Tahnru
QUOTE (shokoshoko @ Apr 24 2009, 09:07) *
----------
Total: 5/6 (10.9%)

So i guess i can tell it apart >.<


It's awesome that you're conducting ABX testing to check out your hypothesis. Please allow me to add my two cents.

To produce a much more conclusive test, follow these steps within the testing module:

1. Set a fixed number of trials before you begin testing. Finish the test - don't stop halfway. I would set my minimum at 10 trials, my max at 16. So, perhaps use 12 as your default.

2. Turn off Foobar's View-the-results-as-I-Go option. It gives you no useful information, and only serves to encourage people to skew their results by throwing out results they don't like.

(by the way, I HATE this option. The abuse of it leads to more badly conducted ABX tests than almost anything else)
cpchan
QUOTE (odyssey @ Apr 24 2009, 09:26) *
Are both encodes from the same source?

Did you replay-gain them?


Also, which mp3 encoder are your using? What quality are you encoding to (what switches are you using on the command line or GUI frontend if you are using a cli based encoder)?
odigg
QUOTE (shokoshoko @ Apr 24 2009, 10:07) *
So i guess i can tell it apart >.<


That's a good result. There are people than can ABX 320 kbps MP3s. But when people show results like yours it naturally raises the skepticism of people on this board because there are a large number of people (me included) who cannot do it.

To eliminate any other reasons for being able to ABX, I suggest you do the following.

Among your lossless files, find an album or 2 you suspect you can ABX easily. Then, using a current version of an MP3 encoder (lame is a good and popular one), convert the lossless files to MP3. When you convert to lossless, don't try to use any fancy switches (setting the filter cutoffs, fiddling with joint stereo) etc. Just use -v0 (variable rate quality 0) or cbr 320 (constant bit rate 320 kbps) or whatever setting you feel is best for you to ABX with.

I suspect you know how to convert lossless to lossy, but if you don't you can use Foobar to do it.

Then when you have these lossless and lossy files, try to ABX them. Follow the guidelines people have already stated for how many trials you should do.

If you can still ABX, you've got some great ears. Feel free to brag to all your friends and tell them how pathetic their hearing is.
Bodhi
I can tell the difference between a flac file and a mp3 file as well!!!
Axon
Bodhi deleted his first comment but it's more than a little valid... that is a bizarre timestamp.
Bodhi
QUOTE (Axon @ Apr 24 2009, 20:33) *
Bodhi deleted his first comment but it's more than a little valid... that is a bizarre timestamp.

rolleyes.gif


He didn't say what bitrate he used as well (did he?).

I can tell the difference between flac and mp3 V7! cool.gif
pdq
QUOTE (Bodhi @ Apr 24 2009, 14:41) *
I can tell the difference between flac and mp3 V7! cool.gif

I won't even ask for ABX results for that comparison. smile.gif
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2009 Invision Power Services, Inc.