Comskip Support Forum

I'm reading though tuning.html and am looking for verification that I'm doing things right. Here are the first number rows of the short.csv file, sorted by brightness (low to high):

frame brightness scene_change logo uniform sound
1 3 0 0 0 107
25169 16 0 0 0 1387
25170 16 500 0 0 137
25171 16 495 0 0 815
25172 16 495 0 0 194
25173 16 490 0 0 556
25174 16 495 0 0 178
25175 16 500 0 0 130
25176 16 495 0 0 56
25177 16 500 0 0 39
25178 16 500 0 0 18
25179 16 500 0 0 10
25180 16 500 0 0 14
25181 16 500 0 0 7
25182 16 500 0 0 12
25183 16 500 0 0 5
25184 16 500 0 0 10
25185 16 500 0 0 4
25186 16 500 0 0 7
25187 16 500 0 0 3
25188 16 500 0 0 6
25189 16 500 0 0 3
25190 16 500 0 0 5
25191 16 500 0 0 3
25192 16 500 0 0 4
25193 16 500 0 0 3
25194 16 500 0 0 3
25195 16 500 0 0 3
25196 16 500 0 0 3


If I'm reading the tuning file, then the appropriate values for the following attributes would be:

non_uniformity=1
max_volume=12
max_avg_brightness=21
test_brightness=41
max_brightness=61

Is that right??? It seems like my max_volume and non_uniformity are wayy too low.

In digital broadcast you can have a very clean signal, no audio hiss so VERY low levels at silence.
Also non_uniformity can be low in digital broadcast.
I would set
max_avg_brightness=18

Experiment with some values

Thanks Erik. I'll tinker around. What is the lowest non_uniformity I should consider using? The help suggests that 250 is the lowest, but based on my data above, it seems like it could be set much lower.

I have recordings where 1 would even work. But its rather sensitive so no need to put that low.
For digital recordings 10 could be could