well i did my own little test. i took a complete wav file form a audio cd. then converted that file into a 160 mp3 a 160 ogg and a 160 wma. now after running these different conversions i tried to keep all settings the same and control it the best i could. and i found that the ogg and the mp3 files sounded almost exactly the same, but the WMA file sounded the best! yea i hate to admit it but M$ won in my little test. all files were also about the same size. with that being said besides the obvious reason that mp3 and wma are copyrighted and have legal stuff behind them why should i switch to using ogg? Plus these two formats (not so much WMA) all more supported by hardware and software.

EDIT

i did the test a second time with different files and different settings and same results!!!!!

You may want to try adjusting volume levels. WMA boosts volume (IIRC) which makes most people perceive the format as more appealing - somewhat like boosted radio stations catching your ear as you surf the airwaves.

Cheers, Paul

If it's about audio comparison people here call it "test" if it's double-blind (ABX), in your case original vs. lossy compressed copy. Otherwise the results won't be reliable because of placebo-effect. Besides this you should give more details (after having performed the test in a double-blind way): What encoders/versions were used, what settings, what software for plaback/testing. If you successful ABX differences, give a detailed description of them and provide a lossless compressed short sample of the original file.

More about this you'll find in HA terms of service and in the FAQ.

I think the most important thing is to try out lots of different samples before concluding that one format sounds better than another.

About the volume difference: you should be able to use wavgain or mp3gain (if you re-encode to mp3 using something like --alt-preset insane) to figure out if the volumes are really the same or not. I wouldn't worry too much unless the volumes are out by a half dB or more.

Also of note: I've been told that wma may work differently from other cbr codecs at the beginning of a file. So it may sound better near the beginning but worse later on. To be safe, I might try chopping off the first 5 seconds of every sample before comparing.

ff123

We will give you a chance to explain yourself before we completely tear you to bits.
But........

1. If you did not include the original WAV sample in your comparison you have no Idea what quality is or how to judge it properly. And we will bitchslap you.

2. If you did not write or use a software for blind testing or have a friend give you samples to judge without knowing what encoder produced them. We "WILL" bitchslap you.

That covers basic general testing. Now on to encoder usage sepecifics for testing. WMA has the least developed VBR modes of any codec AFAIK. But as is typical with Microsoft they stick with outmoded technology and techniques so their CBR mode is one of the best developed.

1. If you encoded your Vorbis sample with the -b 160 switch. Not only will we bitchslap you. I think mortal wounding is called for. Vorbis is a VBR codec by nature if you tried to force it to be CBR. Well that's just wrong.

2. If you used Lame in a CBR manner and did not use one of the alt-presets at this range then we will bitchslap you as well. Mortal wounding is at our discression. The lame MP3 encoder is known for it's superior VBR implementation in refference to other MP3 encoders. If you wanted a better CBR MP3 encoder to test then you should have used Fhg if I am not mistaken.

As a total test of how good a codec ultimatly is. It is not unfair to compare CBR and VBR encodes. Each codec is designed to be used in a specific manner. To use them otherwise is to ensure they will fail. Also the more test samples the better your test will be. Each codec has it's strong and weak points.


Oh and since we are talking mid range bitrates why did you not test MPC?

well as you all pointed out ( and i admit it) that my test was not complete.
as for what i used to encode

mp3 - lame (LAME 3.90)

WMA - i just used WMA Workshop

Ogg - OggdropXPd

both mp3 and ogg VBR

i am not an audiofile ( so dont rip my ass ) , but i do not understand how the wav files matters. a wav file is a wav file no? and if the same file was used then it shouldnt matter? right?

also i do understand what you have said about volume difference with WMA codec and you were right part of the reason it sounded better was the volume.

as i have heard from many audiofiles that ogg vorbis is the superior codec so i will take that more than i trust my own novice test.

have any of you yourself ran some more complete tests? what did you find?

i understand that sometimes comparing some different codecs and different settings within them is like comparing apples and oranges, but i was just looking for a simple, "average listener" audio compression

The Search button is holy.
This is all i think should be said here.
You may be surprised how many times wma vs whatever has been discussed here.

the completely right answer

Also, if you make a claiim that it is better, you should say why, because you couldn't notice a difference from the original? Or because you could notice but the difference compared to other files is the slightest? Or what ever reason?
Using Search you'll be able to find all the information you need here.
If you'll do this first, it's very possible this could spare useless debates.
Search for things, and read rule #8 in the Hydrogen Audio Terms of Service.

1 Lossy encoders handle sound samples differently. Especially with VBR encoding. So I don't think a wav is just a wav as you put it, at least when it comes to lossy compression.

2 Right

3 Musepack has been stated as the superior codec "right now" as far as I know. Why didn't you try that? I doubt you've actually read that Ogg Vorbis is superior, but it is very good and I would love to see it progress...

Also you should have used Garf's new beta encoder for Ogg Vorbis.. I've heard it's more betterest at being good I hear..

The WAV file matters, as a comparison. The wma sounding better than the MP3, has no relevance if it was not compared to the original WAV.

No. Lossy audio compression is no DSP (e.g. equalizer).
Compressed file indistinguishable from original .wav -> compression is good.
Compressed file sounds different (e.g. "better") -> compression is bad.

So to find out what's good or bad compression you have to compare .wav vs. lossy compressed file and do this in a double blind way (e.g. ABX). Where to find more information about it has been said already.

I just wanna say I did my own little test and I copied an MP3 to audio tape. And guess what.. the sound quality improved.

prove me wrong..

And pigs flew above your head... right?

The only thing I can add to this thread is having been told ... such comparisons are "apples" to "oranges", if comparing the reported kbps. If a practical property of the file versus sound quality is ABX'd, wouldn't it be the resulting file size??

The isn't obvious until you think about it, so no slap here( ) : If you don't use the wav, how will you know what the sample is supposed to sound like.



Heh so you decided the result before you even ran the test. Why even bother then?