msycamore said:

AntcuFaalb said:

Now the question is: What's better, a median-of-three or a TooT?

I've heard that TooT is very effective but I have never been able to try it out for myself. The median script of g-force was very effective at getting rid of the herringbone pattern on my Technidisc-pressing of Star Wars but maybe it's worth a comparison to see which proves to be the most effective method? I recall that Moth3r was pointing out that a multi cap average proved to be more effective to get rid of dotcrawl but a median was more effective at getting rid of dropouts etc.

TooT and MedianOfThree both get rid of anomalies that exist in only one of the three sources; e.g., dropouts and rot.

The difference is: (The rot here is obviously 57.)

TooT(57, 127, 137) = 132

MedianOfThree(57, 127, 137) = 127

Since TooT averages the two most similar pixels of the three (effectively dropping the third) it seems to be the best-of-both-worlds.

However, g-force once made a pretty convincing case to me that TooT will result in more noise than MedianOfThree. I don't remember his argument and I can't think of a good one at the moment.