Ignore more samples that are near the edge of the block. The reason is that the filtering tends to cause these samples to go above the window more than the samples near the middle. If these samples are included in the unwindowed peak estimation, the peak can be overestimated. Because the block is windowed again before overlapping, overshoots near the edge of the block are not very important. 0.1 is the value from the version originally contributed to calf.
Showing the changes to libavfilter (1 of 1 file). Show all
Unified · Split
| @@ -87,7 +87,7 @@ static void generate_hann_window(float *window, float *inv_window, int size) · 86 unmodified lines | |||
| 87 | 87 | ||
| 88 | window[i] = value; | 88 | window[i] = value; |
| 89 | // 1/window to calculate unwindowed peak. | 89 | // 1/window to calculate unwindowed peak. |
| 90 | inv_window[i] = value > 0.01f ? 1.f / value : 0.f; | 90 | inv_window[i] = value > 0.1f ? 1.f / value : 0.f; |
| 91 | } | 91 | } |
| 92 | } | 92 | } |
| 93 | 93 | ||