Ive <ive### [at] lilysoftorg> wrote:
> Hmm, I'm not sure if I understand you correct. You mean using adaptive 2
> makes only sense when using *more* than 5 by 5, right?

Yes.  These are from my notes:

  Array Size   Maximum Useful Value   Break-Even Value
    2 x 2              none                  0
    3 x 3               0                    1
    5 x 5               1                    2
    9 x 9               2                    3
   17 x 17              3                    4
   33 x 33              4                    5

Accordingly, adaptive 2 has no effect on 5 by 5.

> Actually I never did a serious speed/quality trade-off compare regarding
> area_lights parameters and used more a "I guess this might work" and
> alternatively a "better more than less" approach. Probably not a good
> idea, so thanks for the head up.

I've done such comparisons, based on Alain's advice.  My general findings are:

  - Adaptive 0 tends to produce unrealistic effects, especially if
    an object (e.g., the lamp fixture!) is close to the light source.

  - Adaptive 1 is sufficient; I haven't had the need to go higher.

  - IIRC, 9 by 9 adaptive 0 was faster than 5 by 5 adaptive 1.  (I
    don't know if that test can be generalized.)  However, the poor
    quality of the adaptive 0 usually isn't worth the speed advantage.

  - Jitter helps.

  - 9 by 9 adaptive 1 is usually a good setting.

However, I haven't tested 9 by 9 adaptive 1 against 5 by 5 adaptive 2.

Post a reply to this message