|
From: Jonathan D. <jo...@ta...> - 2010-06-29 01:04:52
|
On 28 Jun, Reece Dunn <ms...@go...> wrote: > So you don't need to run espeak-phoneme-data if you run espeakedit? The espeak-phoneme-data is only needed to convert espeak-data which has been compiled on a little-endian processor into espeak-data for a big-endian processor. If you compile the data using espeakedit on a big-endian procrssor then you get big-endian data. > Q: Do you need to run espeakedit to construct the espeak-data, or is > it sufficient to use what is in the svn repository? The svn repository only includes the source data, not the compiled data. To get compiled data (the files in espeak-data), you can either run espeakedit and do Compile->Compile phoneme data from its menu (or do espeakedit --compile from the command-line). Or alternatively you can download the it from the eSpeak website. Note that if you using a development version of eSpeak (eg. from SVN) then you must use the corresponding version of espeak-data (i.e. not the Ubuntu version). The simplest method is to download both the espeak program code and its compiled data from: http://espeak.sf.net/test/latest.html Then unpack the ZIP archive, go into the /src directory and do: make sudo make install The fact that there were files missing from the SVN repository shows that nobody has got the files from there recently! > $ espeak -v en hello > env2 = 1 > env2 = 105 > Segmentation fault That's probably an error in the file espeak-data/intonations (perhaps an incompatible version). This file is compiled from its source data at: phsource/intonation by using espeakedit with the menu option: Compile -> Compile intonation data Currently this file is not compiled with espeakedit --compile or Compile -> Compile phoneme data Probably I should include it. > I am then copying the built voices to ./espeak-data-local and running > the platform's espeak-phoneme-data on it. You should only run espeak-phoneme-data to convert little-endian espeak-data to big-endian espeak-data. If your processor is little-endian, then don't run the espeak-phoneme-data program. > > Does eSpeak pick up your data from $HOME/espeak-data which has been > > compiled or converted for big-endian? > From an installed version that has been through `espeakedit --compile` > and `espeak-phoneme-data` (installed to /usr/share/espeak-data). eSpeak picks up espeak-data from the top level of your HOME directory (if it exists there), otherwise from /usr/share/espeak-data |