ABX ogg vs mp3

Here's what someone has to say about that:

http://www.chemphys.lu.se/Homepages/h_nienhuys/p/oggmp3test/

This person seems to think you can go a lot lower than ogg -q6 to
achieve transparancy, which I find strange. I use ogg -q5 (160 kbs
nominal bitrate) for my files, and the difference between that and wav
files is clearly audible. I picked q5 because it seemed a good
compromise between compression and quality. Apparently listening tests
show few people can hear the difference between q7 and lossless encoding.

Here's some tests (a little old) which suggest your results might
depend a lot on what kind of music you are encoding; it's a low (64)
bitrate test from 2002:

http://ff123.net/64test/results.html

>Here's what someone has to say about that:
>
> http://www.chemphys.lu.se/Homepages/h_nienhuys/p/oggmp3test/

Yes, I've seen that. My question, though, was if Aaron could
identify the crescendo in a blind test, in this particular case.

>This person seems to think you can go a lot lower than ogg -q6 to
>achieve transparancy, which I find strange. I use ogg -q5 (160 kbs
>nominal bitrate) for my files

Your average looks more like 130Kbps from what I have here.
But then, your stuff is "classical", whereas the nominal bitrate
is probably estimated with at least some electric music in
the mix.

-Carl

On Friday 01 October 2004 03:54 pm, Carl Lumma wrote:
> >Here's what someone has to say about that:
> >
> > http://www.chemphys.lu.se/Homepages/h_nienhuys/p/oggmp3test/
>
> Yes, I've seen that. My question, though, was if Aaron could
> identify the crescendo in a blind test, in this particular case.

At least on my mp3 player during playback, it's *blatantly* obvious that there
is a large increase in volume, so I would pass this test 100% of the time.

Aaron Krister Johnson
http://www.akjmusic.com
http://www.dividebypi.com

>> > Here's what someone has to say about that:
>> >
>> > http://www.chemphys.lu.se/Homepages/h_nienhuys/p/oggmp3test/
>>
>> Yes, I've seen that. My question, though, was if Aaron could
>> identify the crescendo in a blind test, in this particular case.
>
>At least on my mp3 player during playback, it's *blatantly* obvious
>that there is a large increase in volume, so I would pass this test
>100% of the time.

Easy to say, harder to do.

-Carl

On Saturday 02 October 2004 12:13 am, Carl Lumma wrote:
> >> > Here's what someone has to say about that:
> >> >
> >> > http://www.chemphys.lu.se/Homepages/h_nienhuys/p/oggmp3test/
> >>
> >> Yes, I've seen that. My question, though, was if Aaron could
> >> identify the crescendo in a blind test, in this particular case.
> >
> >At least on my mp3 player during playback, it's *blatantly* obvious
> >that there is a large increase in volume, so I would pass this test
> >100% of the time.
>
> Easy to say, harder to do.

BTW, I plan to do just this with a simple randomize script which will run a
AB-X test by using a command line player. When I get round to it, I'll let
you know the results.

I might have a harder time with other audio examples that *don't* have this
audible artifact of a crescendo, but in a way, it's cheating, that's how
obvious it is on my player.

Aaron Krister Johnson
http://www.akjmusic.com
http://www.dividebypi.com

On Saturday 02 October 2004 12:13 am, Carl Lumma wrote:
> >> > Here's what someone has to say about that:
> >> >
> >> > http://www.chemphys.lu.se/Homepages/h_nienhuys/p/oggmp3test/
> >>
> >> Yes, I've seen that. My question, though, was if Aaron could
> >> identify the crescendo in a blind test, in this particular case.
> >
> >At least on my mp3 player during playback, it's *blatantly* obvious
> >that there is a large increase in volume, so I would pass this test
> >100% of the time.
>
> Easy to say, harder to do.

You pick the live TV station and time, I'll be there ! ;)

Aaron Krister Johnson
http://www.akjmusic.com
http://www.dividebypi.com