[Vorbis] What's the value range of float samples?

[Nikos Chantziaras]

That only works for known files though, where you know beforehand that 
the scale factor is going to be 0.9.


On 10/01/13 18:17, Joost van Dongen wrote:
> We had audible problems with this in both Swords & Soldiers and
> Awesomenauts. As Paul suggested, we worked around it by simply
> multiplying all our samples to get them back in the -1,1 range. We
> multiply by something like 0.9 or 0.99, don't remember the number exactly.
>
>
> On 10 January 2013 17:10, Paul Martin <pm at nowster.me.uk
> <mailto:pm at nowster.me.uk>> wrote:
>
>     On Mon, Jan 07, 2013 at 11:50:20PM +0200, Nikos Chantziaras wrote:
>      > On 07/01/13 23:23, Paul Martin wrote:
>      > > On Mon, Jan 07, 2013 at 08:53:56PM +0200, Nikos Chantziaras wrote:
>      > >
>      > >> Apparently, there are Vorbis streams that use float samples
>     with values
>      > >> in the range of [-32768, 32767]?
>      > >
>      > > GIGO.
>      >
>      > So the answer is no?  Vorbis float values always have a nominal
>     range of
>      > [-1,1) and I can always clamp/clip to that?
>
>     If you feed in "normalized to full scale" values to any lossy encoder,
>     you are likely to get overshoots on decoding.  It's good to be aware
>     that this might happen and to do your volume scaling (playback gain)
>     in the floating point domain before you convert to integers.
>
>     You'll often notice this on YouTube videos where they clip audibly on
>     playback due to insufficient headroom.