[FFmpeg-devel] [PATCH 09/11] aaccoder: add a new perceptual noise substitution implementation

[Claudio Freire]

On Tue, Jun 30, 2015 at 1:33 AM, Rostislav Pehlivanov
<atomnuker at gmail.com> wrote:
>
> Why the need to have "sce->pns_ener[w*16+g] = energy /
> sce->ics.group_len[w];", normalization? The energy value is log2'd during
> scalefactor indices calculation so it's hardly going to change much. Will do
> some tests to see if it smooths out the transition of normal bands -> PNS.

The change is still noticeable though during transients (making them
harder), when group_len can go all the way to 4. Since you're adding
energies, but the scaling factor is a scale, if you don't get the
average energy instead the sum, you overestimate the scaling factor.

Subtle, but noticeable on transientful signals.