This is the bug/patch tracker for Rockbox. Click here for more information.
Quick links: Bugs · Patches · Rockbox frontpage
FS#12106 - Rockbox Utility V1.2.10 doesn't let me create voice files
Attached to Project:
Rockbox
Opened by Kevan (KevanGC) - Tuesday, 10 May 2011, 01:05 GMT+2
Last edited by Dominik Riebeling (bluebrother) - Monday, 06 February 2012, 20:02 GMT+2
Opened by Kevan (KevanGC) - Tuesday, 10 May 2011, 01:05 GMT+2
Last edited by Dominik Riebeling (bluebrother) - Monday, 06 February 2012, 20:02 GMT+2
|
DetailsHi,
I've been trying to make a voicefile for my Sansa Fuze V2. I used the SAPI5 engine with a NeoSpeech voice. It's a big voice but it sounds great. In Rockbox utility 1.2.8, it would do it half the time, and sometimes it would give an error. I guess it's because the voice is so big, it takes up lots of memory. However, in the new 1.2.10 utility, i'm completely unsuccessful in creating voicefiles with this voice. It always gives an error. It says encoding of "C:/Users/KevanGC/AppData/Local/Temp/rbvoice//random file name.wav"" failed. It gives a random filename, I guess because it's trying to encode the .wav files, taking too much memory, and getting stuck on a random file. Rockbox utility version 1.2.8 did the same thing, but only half the time with the settings set exactly the same. The new V1.2.10 does it every time I try, but I'm able to create voice files on it just fine if I use a different voice. The NeoSpeech voice does work correctly. I'll attach one of the logs it gave me so you can get more info. I'm just wondering, since the old utility worked half the time, and the new one doesn't work at all, does the new one use more memory? Thanks 
Rockbox Log.txt
(223.3 KiB)
|
This task depends upon
Closed by Dominik Riebeling (bluebrother)
Monday, 06 February 2012, 20:02 GMT+2
Reason for closing: Fixed
Additional comments about closing: Multithreading support for voicefile generation has been removed, will be re-added later.
Monday, 06 February 2012, 20:02 GMT+2
Reason for closing: Fixed
Additional comments about closing: Multithreading support for voicefile generation has been removed, will be re-added later.
FS#11994-- it was supposed to be fixed, but it seems that disabling running the TTS engine with more than one thread (which is the change why it is running slower) didn't help. Interestingly said change didn't make the problem occur anymore for me. Can you reproduce the problem with a different voice (especially the one shipped with Windows, MS Sam / MS Anna)?I can reproduce the problem with Loquendo as well. I think you may want to go back to the old way of doing it, the old one at least works sometimes.
I think it might improve things to have the utility pause slightly after the creation of each .wav file.