>
> Hi!
>
> New test piece: "The Ecstasy Of Gold" by Ennio Morricone; <10sec; (very
> complex part!)
>
> The Encoders: lame 3.70; lame 3.81BETA; FHG 3.1DEMO; SCMPX 1.51; (constant
> bitrate; 128kb/sec; highest possible quality;)
>
> Personal rating (sound quality only):
>
> 1st place: The original sample.
> 2nd place: FHG MP3ENC 3.1 Demo (-qual 9)
> This is the one and only encoder I found, which does not make any
> additional noise.
> However the bells sound a little bit different. (not a LAME problem)
> 3rd place: LAME 3.81 BETA (-h) (no real difference to v3.83)
> Best quality! But there are some high frequency metallic
> vibrations. (at ~6kHz?)
> 4th place: LAME 3.70 (-h)
> More high frequency metallic vibrations. Sounds like low cost
> trumpets are used.
> (resonance between the metal parts of the trumpets)
> These vibrations even occur, if the trumpets are not playing.
> 5th place: SCMPX 1.51
> This one even makes additional noise.
>
> This was very exhausting- I hope it is of any use!
> 5 files=2.4MB; Any place to upload/mail?
>
I just put the original .wav on the LAME website, in the
samples section.
This kind of information is always *very* usefull. LAME is at the
point where most stuff requires listening tests, and you found
out how tedious that can get!
I tend to agree: LAME (128kbs CBR) still lags behind FhG for
some samples, particularly very complicated pieces like this
one and "ftb_samp.wav". However, most 128kbs samples I now
have a hard time distinguishing LAME from FhG, and the
% where LAME does better than FhG is growing :-)
Mark
--
MP3 ENCODER mailing list ( http://geek.rcc.se/mp3encoder/ )
