Outlier/filter on the characteristic or feature side?

[[Th...]]

Do you apply the outlier/filter on the characteristic or feature side?

[[Cl...]]

I choose characteristic side. This allows more flexibility if you have the same feature in multiple characteristics.

[[Ma...]]

Can be done both ways, but i preffer using it on feature. You can always name them as you like, so you can know by name how it is evaluated. And from my opinion it is much more readable. And still you can use another settings on characteristic.

If you use it on characteristics, then a rookie can do misunderstanding in results.

[[Ja...]]

I'd like to resurrect this topic. I've always filtered in the characteristic in the past. A new co-programmer says he's had good luck with filtering in the feature and evaluating in the characteristic. So I tried that in the last program I wrote and found whatever I put in the feature automatically populates the characteristic. That's great when you have multiple characteristics coming off of the same feature because it saves a little time. But when it does that, how do I know that the data isn't getting filtered twice, or is it?

[[Ri...]]

I filter/outlier inside the Feature that way it applies to any theoretical feature by default as well (intersections, etc.)

I adjust inside the Characteristic if needed.

The newer way to program is directly from the Characteristics tab.

Do whatever you feel comfortable doing.

[[Ja...]]

I like how it applies what's in the feature to the characteristic, but I don't know if the data is getting filtered twice (once in the feature and again in the characteristic). Is there a way to find out?

[[To...]]

Please sign in to view this quote.

It is not processing twice.

[[To...]]

A benefit of applying filter/outliers on the feature side is the base alignment features can be processed, if desired.

[[Ja...]]

Please sign in to view this quote.

I would have to agree with you but I wasn't sure. I'm sure Zeiss would've addressed this a long time ago. But then that brings up another question - Which one is doing the filtering - the feature or the characteristic? I'm starting to think the feature doesn't do any filtering, it simply pre-populates the characteristics and the actual filtering happens there, with the exception of curves. Feasible?

[[To...]]

Please sign in to view this quote.

When you apply Filter/Outliers on the feature, the data in the feature is affected by those settings. If F/O are applied on the characteristic, the results could, and will likely, be different when compared to the feature as F/O will not be applied.

When I took my 2nd Basic Calypso class in 2008, the instructor introduced us to F/O settings via Save/Load defaults, which applies F/O on the characteristics. The instructor said, "Even if the settings are not optimal, the results would still be better than no F/O settings at all."

Once the cookbook came around, we got more information about the ideal settings, which was good. The cookbook even said "it is better to apply F/O at the feature level because it would make for more stable features." However, whether you apply them at the Feature or the Characteristic, the results are the same. When I use F/O, I prefer to use them on the features so when I am reviewing a feature, I am seeing processed data.

I look at features as a "bucket of points". Whenever I open a feature or a characteristic, I am pulling in data from the bucket of points and applying whatever evaluations and F/O are active in that feature or characteristic.

When you "Recall Features Points" to create a new feature, the F/O settings on the features does not affect the new feature. It is just the raw data and you still need to process with F/O.