QUOTE (PJay-Z @ Oct 19 2008, 23:04)

I can't hear any difference between them
Please give me some special audio files. ...
I understand your question as being about an original track encoded to mp3 on one hand and to FLAC on the other hand.
Encoding an mp3 file to a FLAC file of course wouldn't produce any audio content difference as FLAC is lossless.
Encode for instance
eig which is one of the outstanding problem samples for mp3 (extreme pre-echo problem). harpsichord music like in
harp40_1 also needs a very high quality setting.
Your experience shows that mp3 @ 128 kbps is more or less perfect usually. These samples show that there are exceptions wich require a higher mp3 quality setting like -V0.
In the end it's all up to how to handle these exceptions. Many people don't care about them cause they're rare and at least the most extreme cases seem to be existing only with electronic and harpsichord music.
These people can be happy with mp3 @ 128 kbps.
Those who care more about these exceptions use an encoder setting which yield an average bitrate of ~170 kbps or more.
At an average bitrate of ~240 kbps and when using a good encoder (use for instance Lame 3.98 -V0) most of the exceptions become inaudible or nearly inaudible to most people.
An even higher safety margin can be obtained by using a setting like ABR 280 or CBR 320 though this is expected to be only a theoretical advantage.
Anyway because of these exceptions there is a justification for going lossless like with FLAC for those who are extremely cautious.
Midway between lossless and highly-efficient lossy codecs like mp3 lies the lossyWAV-FLAC-combination.
This is effectively a lossy codec, which at the expense of a relatively high bitrate compared to mp3 (typically 470 kbps) offers a quality which can hardly be expected to have an issue. No problem sample is known for this approach. The relatively high bitrate can be accepted easily by many users with respect to today's storage capacities.