File: TODO

package info (click to toggle)
cicero 0.7.2-2
  • links: PTS, VCS
  • area: contrib
  • in suites: wheezy
  • size: 384 kB
  • sloc: python: 1,235; makefile: 25; sh: 6
file content (19 lines) | stat: -rw-r--r-- 813 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
-ALSA interface
-Remove dependency on sox for saypho and bulktalk.
-Filters: consider adding a dictionnary-based filter for full word matches.
-Filter non-printables properly.
    Préfiltrage: ttp.PhonetizeError: Char unmatched <«> 171
-Roman numbers filter.
-Volume and speed controls.
-Trace calls: disabling tracing just throws away debugging strings instead of
   printing them, but we still waste a lot of effort preparing those strings.
-Chunking in phonetizer is somewhat crude...
-MBROLA output estimation: improve heuristic, timeout to save us if
   we guess wrong.
-Use very long text as a reference to see impact of new rules.
-Make a regression example text file for numbers.
-Improve text preparation.
-Less crude config file.
-FIXME's...
-User dictionnary.
-Bridge with more standard speech APIs.