Posted by fred-bloggs on 11/16/07 11:36

southend.unitedfc.5.lukebosman@spamgourmet.com (Luke Bosman) wrote in
news:1h8nj3q.g7v8jv1cys0aN%southend.unitedfc.5.lukebosman@spamgourmet.com
:

> fred-bloggs <fred-bloggs@hahahotmail.com> wrote:
>
>> southend.unitedfc.5.lukebosman@spamgourmet.com (Luke Bosman) wrote in
>> news:1h8n71s.gj7j7pncxbx5N%
>> southend.unitedfc.5.lukebosman@spamgourmet.com:
>>
>> > fred-bloggs <fred-bloggs@hahahotmail.com> wrote:
>
>> >> In a set of public double-blind listening tests in 2004, the Lame
>> >> 3.96 mp3 encoder was rated *equal* to Itunes AAC 4.2 at 128 kbps
>> >> CBR. http://www.rjamorim.com/test/multiformat128/results.html
>> >
>> > Lame looks to be slightly inferior (4.18 vs. 4.26), although sample
>> > sizes of 20 or so are hardly statistically valid.
>
>> You are quoting the mean.
>
> Indeed.
>

The confidence intervals overlap. Saying that 4.26 is better than 4.18 is
to ignore the statistical analysis.

> I am aware of the author's conclusion. However, as I have said sample
> sizes of 20 are hardly statistically valid.

Who AM I to believe? Someone who has done the work and subjected it to
public review or someone who says *I know* without any substantiation?

> Interestingly, for many
> tracks iTunes AAC is considered to be superior.
>

Again, the confidence intervals overalap, the difference is not
significant.


--
fred

[Back to original message]