Skip to main content
Topic: Download.com does an internal listening test (Read 6938 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Download.com does an internal listening test

Reply #1
Quote
Seems to be a blind listening test done with high end audoi gear in an audio studio:

http://www.download.com/Audio-Codec-Shoot-...20-5146271.html

Results that Microsoft might be very happy with.
[a href="index.php?act=findpost&pid=303774"][{POST_SNAPBACK}][/a]

The fact that 128 encodes were very often considered to better than 192 ones makes me very sceptical of this test. WMA 128 better than Ogg 192? Come on...  I am not especially sensitive concerning compression artefacts, but the metallic ringing of a typical WMA@128 really hurts my ears...
Proverb for Paranoids: "If they can get you asking the wrong questions, they don't have to worry about answers."
-T. Pynchon (Gravity's Rainbow)

Download.com does an internal listening test

Reply #2
they should also have used LAME instead of iTunes to encode their MP3s

...and of course an up to date version of Vorbis, instead of "Xiph.Org libVorbis I 20030909" ^^
Friends don't let friends use lossy codecs.  (char0n)

Download.com does an internal listening test

Reply #3
That the higher bitrate clips do worse seems to indicate that this is a test to find the  codec "which sounds best to you". Not most true to the original.

Bring in another crowd an the results could be the complete opposite.

Download.com does an internal listening test

Reply #4
Quote
The fact that 128 encodes were very often considered to better than 192 ones makes me very sceptical of this test. WMA 128 better than Ogg 192? Come on... tongue.gif I am not especially sensitive concerning compression artefacts, but the metallic ringing of a typical WMA@128 really hurts my ears.


This is the stuff that pisses me off right here. I say we all petition and send them a nasty e-mail or something pathetic  . Who did they hire a Johnny Nothing off the street to do this? It says the test is "non-scientific" how would you expect anyone to take you seriously?
budding I.T professional

Download.com does an internal listening test

Reply #5
Quote
... I say we all petition and send them a nasty e-mail or something pathetic  . Who did they hire a Johnny Nothing off the street to do this? It says the test is "non-scientific" how would you expect anyone to take you seriously?
[a href="index.php?act=findpost&pid=303813"][{POST_SNAPBACK}][/a]


Got my vote!
Ogg Vorbis for music and speech [q-2.0 - q6.0]
FLAC for recordings to be edited
Speex for speech

Download.com does an internal listening test

Reply #6
Wonderful. You work your butt off sometimes (at least it feels that way) just to try an help folks you encounter around the net in understanding what lossy compression is really all about and why >proper< testing is important when deciding quality. There is already tons of misinformation floating around out there as it is, and I would say greater than 60% of the netizens out there have it completely wrong thanks to that fact. Then a totally irresponsible article like this one comes out and sets the world back even more, negating any progress we've clawed our way through. Personally I think this has nothing to do with which is better, or sounds more pleasant either, but a shady attempt and pusing more people to use Microsoft "product". It's feels like being pushed back into the dark ages. Oh well (heavy sigh) at least the good people of HA.org won't be fooled... right?

Edit: Fixed bad typo some might find offensive lol.

Download.com does an internal listening test

Reply #7
Weird that in "Cherub Rock," Smashing Pumpkins, WMA@128kbps scored higher than WMA@192kbps.

And I would have thought OGG@Q6 would be transparent in most cases, much better than WMA@128kbps too.

Download.com does an internal listening test

Reply #8
This thread has already mentioned that wierd listening test.

Download.com does an internal listening test

Reply #9
Quote
That the higher bitrate clips do worse seems to indicate that this is a test to find the  codec "which sounds best to you". Not most true to the original.

Bring in another crowd an the results could be the complete opposite.
[{POST_SNAPBACK}][/a]

Apparently, no:

Quote
He also periodically played the WAV file for reference anytime a juror requested.

[a href="http://www.download.com/Audio-Codec-Shoot-Out/1200-20-5146328.html]http://www.download.com/Audio-Codec-Shoot-...20-5146328.html[/url]

Download.com does an internal listening test

Reply #10
If they couldn't differentiate between 128 and 192, then obviously 128 wins.

Download.com does an internal listening test

Reply #11
There was no mathematical analysis.
They probably did not hear any artifact in any file, and just thew random marks... Which explains why all codecs get nearly the same mark at all bitrates after summing up the results.

128 kbps got a better mark than 192 kbps 8 times, and 192 a better one than 128 12 times.

Download.com does an internal listening test

Reply #12
It probabbly goes to show for anything > 128 Kbps 'normal' people are rubbish at listening tests.

Download.com does an internal listening test

Reply #13
Quote
It probabbly goes to show for anything > 128 Kbps 'normal' people are rubbish at listening tests.
[a href="index.php?act=findpost&pid=303965"][{POST_SNAPBACK}][/a]


They were no "normal people", they were selected as "folks with audio expertise and high standards ", from Download.com and MP3.com crews.

Download.com does an internal listening test

Reply #14
For a test like this I'd really want some further descriptive statistics than just the average scores.

Also, I wonder how they enforced independence of listeners (who musnn't be allowed to influence each other) .

At least they say right up front that it wasn't scientific.

 
SimplePortal 1.0.0 RC1 © 2008-2020