hi guys, i need to find some sort of tests between mp3s of different bitrates and CD audio. my basic aim is to have someones reseach to "disprove" fraunhofers claim that 128kb/s mp3 is cd quality. i have looked in the listening test section but it seems to be all mp3 vs X+Y, if anyone could point me in the direction of a general mp3 quality test at various bitrates preferably with only one encoder that would be great.

in summary, tests of different bitrate mp3s encoded with lame (or whatever) against cd/uncompressed audio

thanks
dan

ps sorry for asking such a simple question, i jsut need to refer to some research in my dissetation.

I believe C'T did such a test now quite a while ago. Their conclusion was that 128k mp3 was not enough to fool most listeners whereas 256k was.

thanks for the reply, thats basically my proposal, but i need a link to some research to support my claim...

i just cant find anything that simple, all the listening tests i can find compare mp3 to aac or something else or compare encoders.

If you look at any test on Hydrogen Audio e.g. this one:
http://www.hydrogenaudio.org/forums/index....showtopic=13464

...you'll see that the codecs are given a grade between 1 and 5. It gets a 5 if it sounds the same as the CD, and a lower number if it doesn't.

In some tests (check for yourself) people are required to ABX between the original (CD quality) and coded version. They can only give the codec a score lower than 5 if they prove they can hear a difference.

So, any codec scoring lower than 5 has been proven to sound different from the CD.

You'll see that lame mp3 @ 128kbps scores 4.29 overall in the test I've just linked to, showing that many people could reliably detect the difference between the original track and the mp3. Full results are linked to at the top of that thread.

Hope this helps.

Cheers,
David.

5 means indistinguishable from the original. However it seems that a 4.0 or higher score is accepted as 'CD quality' in some testing procedures, see comments from Ivan regarding listening tests and NeroDigital webpage for example.

It's a problem of definitions, too.

Here is the EBU definition of "indistinguishable"

http://audiocoding.com/wiki/index.php?page=indistinguishable

However the definition of "CD quality" is pretty much whatever the person or company using it wants it to mean.

ff123

guys, what can i say. thank you kindly for the help you have saved me a lot of time and trouble and unecessary work. thats great...

be warned i may be back...

download this wav file and encode it in mp3 128kbit/s. that pretty much closes the "CD quality" case.

Lol, I was just experimenting with iTunes AAC and fatboy yesterday (I happened to be ripping that Fatboy Slim CD anyway). Even at CBR 320 it is easily distinguishable from the original.

Heh, I never knew Fatboy.wav was a clip from "Kalifornia"!
It seems MP3 has trouble encoding these low frequency
sawtooth-like sounds. I guess it's related to sfb21, or possibly
block-switching.

In my experience gpsycho(gogo) does a lot better with these types
of sounds at lower bitrates, whereas nspsytune chokes badly.

maybe im being very stupid... but are there any tests for 192kbps+ mp3s i can only find 128kbps

bear with me, im a n00b

thanks

Can you really use listening tests for a research project? While it may be a generality that the listening test prove quality, it is still a subjective result.

sorry... the mp3 quality thing is not the basis of the reseach.

the title is
"What effect has the Internet had on the way consumers buy music"
or something along those lines

basically in the introduction to the project i want to talk about why MP3s are good and why they are bad, and one of the most important aspects is quality obviously.

to show the subjectiveness of MP3 quality i have quoted fraunhofer and a couple of other sources who rate 128kb/s as CD quality, i want something for the other side of the coin. to be honest i dont care if its subjective or not, its a minor part of the report but allows me to demonstrate that i have done some research and comparisons!

ps im not really a slacker!

pps i need some MP3 experts to interview about the whole MP3 ethics debate etc. so any volunteers get in touch

go on, help me out....

point me in the direction of some higher bitrate tests, i.e 128-320kbps

you know you all wanna help the n00b.
..
.

Here are links to the article that Garf referred to:

English translation (not machine translated!):

http://www.geocities.com/altbinariessounds...al/mp3test.html

And the original article in German:

http://www.heise.de/ct/00/06/092/

ff123

thank you very much for the link ff123 that is pretty useful.

thanks to everyone who gave advice on this thread