[Vorbis] oggdec.exe not using album gain stdout option

[Chris Harrington (Personal)]

Ulrich Windl wrote:
> On 24 Oct 2006 at 18:26, Chris Harrington (Personal) wrote:
>> I believe that is the case; "radio" replaygain is for a radio station 
>> that wants all tracks to be equal level. Audiophile replaygain is 
>> obvious in contrast.
> 
> Hi,
> 
> Hmm...: If replaygain is something like the "mastzer volume", I see no problem 
> with _applying_ the gain for audiophile, but I see the problem that the gain might 
> be _computed_ incorrectly (e.g. making a silent track of a CD very loud).
> 
> (The problem is that old CDs have their peak as low as -10dB, while recent one are 
> up to 0dB)

The computation is not at decode; it's a separate step with a different 
application. So the computation is irrelevant with regards to oggdec.

In any case, with the audiophile mode the entire CD is analyzed for 
normalization, not just single tracks. Take Tool's Undertow; it has 60 
tracks of silence. In audiophile mode, those tracks are only normalized 
when the other 9 tracks are taken into consideration. Thus, the silent 
tracks are only gained slightly.

In contrast, in radio mode, individual tracks are gained in the way 
you're referring to. Silent tracks, if not "perfect" silence, will 
contain audible noise. However, when one uses the radio mode, typically 
they will be removing silent or near-silent tracks from their playlists.
Perfectly silent tracks will have no audible noise, since your player 
cannot multiply zero by anything to achieve a value other than zero.

-Chris