[Split from iTunes 4.5 MP3]



So noted. iTunes 4.5 still appears to have trouble with this sample.

This has left me wondering. The reason I chose to experiment with iTunes MP3 in the first place was because I had a hard time getting LAME 3.96 to sound transparent using the recommended settings of --preset standard or -V 1 etc. I couldn't put my finger on it, but something was "missing".

So, to try to explain how iTunes encoded MP3s sounded more "detailed" than what I'd been able to produce with LAME, I used Adobe Audition (trial) and did a Frequency Analysis on each of my encodes as follows:

iTunes 4.5 MP3 192kbps VBR


LAME 3.96 --preset standard


I was somewhat surprised that --preset standard clipped high frequencies like that. I did a bit more reading up on LAME switches and found "-k".

LAME 3.96 --preset standard -k


By using the -k switch to disable frequency filtering, the highs were no longer clipped and LAME finally sounds transparent to me. I realise that by adding the higher frequencies this may introduce artifacts at bitrates higher than had it not been added. As a result I've been using: -V 1 -k for my encodes ever since.

I hope this helps to solidify my earlier comments.
Thanks.

That shows nothing. ABX, then comment.
I really doubt -k did anything for you.

Most people can't hear that.
And it is intended to remove those frequencies.

Quoted from Garf:

A GRAPH DOES NOT TELL YOU ACCURATELY HOW THE MUSIC SOUNDS.

YOU CANNOT JUDGE A PSYCHOACOUSTIC CODEC BASED ON FREQUENCY ANALYSIS.