[Trevor Meier]

Fair enough… the problem is that the notes field requires stopped playback and a lot of clicks. Keywords are fast and can be edited while playback continues – great for logging IMO… when really this could be remedied by a true subclipping feature and/or the ability to search the text of keywords.

[Trevor Meier]

Works well… except, try this: select a range that overlaps those two favourites. Press F. Now look at your notes field. Gonzo.

[Trevor Meier]

I agree it’s nice to be able to differentiate between keyword metadata and generic text search. The problem, really, is that there is no sub-clipping in addition to keywording. While I agree that keywording is an amazing facility, it falls short as a scheme to uniquely identify clips. A single clip name is insufficient for identifying long clips that contain more than one subject.

Right now, keywording is our only tool for name-based subclipping, and so in this context the inability to search within the text of keywords becomes problematic. Using keywords to create subclips will result in a large number of unique keywords. In this situation an editor needs a facility to search for text within the keywords.

Let’s use interviews from your screenshot as an example. To break up the long interview, you range-keyword individual sections with a name that uniquely identifies it. So let’s say, in the first interview, the jockey says “The horse was running very fast.” A second interview with the trainer has him state “The jockey had his horse running, they were going fast.”

Using keywords to subclip, the range from the first interview would be keyworded something like [jockey] and [horse running very fast]. The range from the second interview would be tagged [trainer] and [jockey had horse running fast].

Then, in my events browser, I could have a smart collection (or search) for the text-within-keyword “running” and get the appropriate clips.

However, this doesn’t work in the current system. I would either have to manually check each keyword in the search dialog (untenable when there are a large number of unique keywords, as there will be when using keywords to subclip). Or, I would have to keyword topically and lose my ability to uniquely identify clips by their keyword (such as keywording each with [horse, fast, running])

[Trevor Meier]

Yes, what I mean is that text search doesn’t search the text within keywords.

As in your example, doing a text-only search for “Horse” will not find clips that have the keyword “Horses”.