[Audacity-devel] [Fwd: New Label Track Features]
A free multi-track audio editor and recorder
Brought to you by:
aosiniao
From: Shane M. <smu...@ob...> - 2004-07-15 20:02:54
|
This didn't go through when I sent it before because I sent it from a non-subscribing account. Sorry if a duplicate shows up. ... I really like the new label track features. They fits in really well with some new functionality I'm adding and should be ready within a week or so: a "transcription" toolbar. It will provide functions that help you identify and select individual words, and easily create labels from them. The "Add label at selection" menu option will be accessible via a button there, as well as other things that are useful for a small set of users but clutter to many others. I wanted to open up a discussion on what the format of the "export labels" file should look like. Currently, labels can have either a single point or a cover a region. The way this was made possible is really nice--many other software programs have two types of tracks for these two types of labels. And, labels can overlap with the current setup. But, only the start point gets saved (and loaded) from the text file. Currently, the text file looks like: 0.847528 word1 2.565805 word2 3.333325 4.223255 word3 ... Note that a label can be blank. I'm not sure if the labels have to be sorted in the text file. It would be nice to save _either_ the length of the word, or the start and end points of the word. Without being too concerned about backward compatibility for the moment, here are some options: (1) tagged format: START: "word" 0.898 END: "word" 0.992 START: "word2" 1.003 END: "word2" 1.432 TAG: "word3" 1.644 TAG: "" 1.833 This above could be modified to produce XML for greater buzzword- compliance. I don't think it is a good idea, because it hurts the ability to read into excel, MatLab, R or awk other data analyses programs. (2) start-end format "word1" 1.3223 1.533 "word2" 1.66 1.66 "region3" 1.8 3.0 "word4" 1.9 1.95 "" (3) start-length format "word1" 1.323 0.15 "word2" 1.53 0 "word3" 1.8 1.0 "" 2.1 .3 "word4" 2.2 .1 Note that by placing the text label at the end of the line, we might be able to get rid of the "", but I'm not sure if that works well--it might be impossible to save "", " ", and " " differently. I think any format change should include moving to a "" around the text, and maybe even separated with a visible symbol (commma) instead of tab- delimited. Any thoughts? Any from people who actually use the label track? How important is backward compatibility of the file format? FWIW, I favor something like #3, but I'm flexible. Stm... |