The linked files at the bottom of this post are short samples taken from Shakira's "Don't Bother" track off her first edition Oral Fixation Vol2 CD. One is what came from the CD, the other is the result after careful de-clipping with Geniesys DeClipper 2.5 DX plugin followed by dynamic range expansion using the Dominion VST plugin, and mild very low end bass boost to compensate for Dominion's bass loss effect. I leave it to you the tell which is which (using an editor is cheating!).

Both samples have been MP3Gain adjusted and now show 90.8 on whatever scale MP3Gain uses and calls dBs (can someone please tell me dB's relative to what?).

I have misgivings about presenting short samples like this as on PC speakers and/or to untrained ears they may well sound the same. Plus one of the big issues about max volume CD's is that they fatigue the listener after a few tracks - not after 30sec!

FYI, pictures below are the complete track before processing (top) and after (bottom), left channel wave file screenshot using Audacity. This track showed interesting asymmetry after dynamic expansion.





Click to view attachment
Click to view attachment


Cheers,
rysw

I can hear the distortion in sample Y and the lowend doesn't sound as loud as it does in X. I like the sound of X much more than I do Y.. So is Y the original and X the processed one? If so, I am really impressed.
And just in case:

These samples relate to the discussion in this thread...

http://www.hydrogenaudio.org/forums/index....showtopic=55855

...better to posts thoughts / opinions / results there.

Cheers,
David.

For David I've had a play with the two samples from http://www.hydrogenaudio.org/forums/index....showtopic=26691

as requested in thread http://www.hydrogenaudio.org/forums/index....showtopic=55855

and attach the results.

I'm not that happy with the results. Refer my comments in the second thread above.

I vote for X > Y

2Bdecided, the Replaygain developer, has mentioned this on a previous thread, and there's more on www.replaygain.org under Contents/Setting the Reference Level.

Interesting - I'm not sure I have that Shakira track, I only have 'Laundry Service' with 257ms lost in 96171 clips (click for details) - my 7th worst measured album.

Can you give my DeClip program a go on your track and see what you think?

If you type declip -l input.wav output.wav it lowers the amplitude the same amount as if you were declipping it (declip input.wav output.wav) so you can do an A/B comparism with the same level. the quieter one will probably sound better anyway as the DAC has an easier time but you should still notice a difference, confirmed by a look at the waveforms:



Please note this program is posted in the Uploads section here at HydrogenAudio. BTW I listen to ALL my CDs declipped first now via this Free Declipping iTunes Server wot I hacked my declip algorithm into