Rockbox.org home
Downloads
Release release
Dev builds dev builds
Extras extras
themes themes
Documentation
Manual manual
Wiki wiki
Device Status device status
Support
Forums forums
Mailing lists mailing lists
IRC IRC
Development
Bugs bugs
Patches patches
Dev Guide dev guide
Search



Donate

Rockbox Technical Forums


Login with username, password and session length
Home Help Search Staff List Login Register
News:

Rockbox Ports are now being developed for various digital audio players!

+  Rockbox Technical Forums
|-+  Installation / Removal
| |-+  Rockbox Utility
| | |-+  Rockbox and Voice file generation
« previous next »
  • Print
Pages: [1]

Author Topic: Rockbox and Voice file generation  (Read 2484 times)

Offline trumpton

  • Member
  • *
  • Posts: 4
Rockbox and Voice file generation
« on: September 11, 2010, 11:25:13 AM »
I've just generated voice files for my sansa clip using the latest rbutility, and have noticed that some of the spoken phrases are difficult to understand.

The cause of this is that the strings that are passed to the speech synthesis contain technical / joined words, for example:

  "Replaygain type"
  "Idle Poweroff"
  "Filesize"
  "Cuesheet"

It's not clear to me if the source of the text for the speech is also used for displaying on the screen, but simply hyphenating words goes one massive step towards improving clarity, e.g:

  "Idle Power-Off"

The absolute ultimate would be to put data into the file using something like ssml, so things like "A" would be pronounced correctly too.

Steve
Logged

Offline bluebrother

  • Developer
  • Member
  • *
  • Posts: 3421
  • creature
Re: Rockbox and Voice file generation
« Reply #1 on: September 11, 2010, 02:43:13 PM »
Quote from: trumpton on September 11, 2010, 11:25:13 AM
The cause of this is that the strings that are passed to the speech synthesis contain technical / joined words, for example:

This is a known issue. If you generate voice files from the Rockbox build system there are already such transformations applied (see tools/voice.pl). However, there are two problems with this:

1. The transformations needed vary between TTS systems. For example, I'm using the default voice provided with OS X 10.6 and "Idle Poweroff" sounds ok for me.
2. Rockbox Utility doesn't do any transformations at all right now. Optimally the transformation list would be some kind of text file that is shared between voice.pl and Rockbox Utility.
 
Quote
The absolute ultimate would be to put data into the file using something like ssml, so things like "A" would be pronounced correctly too.

I've never heard of ssml, can you tell me more about that?
Logged
Rockbox Utility development binaries (updated infrequently) · How to ask questions the smart way · We do not estimate timeframes.

Offline Lope

  • Member
  • *
  • Posts: 30
Re: Rockbox and Voice file generation
« Reply #2 on: October 06, 2010, 02:48:06 PM »
When I first installed RB I was noob and I told it to create voice files for my whole Sansa Clip+ Drive.
It made .talk files for all the rockbox files etc etc. Big mess.
I've since wiped everything out and started fresh and only created talk files in the music folder :)

A nice tweak to the utility would make it dis-allow misuse like that.
Logged

Offline Llorean

  • Member
  • *
  • Posts: 12931
Re: Rockbox and Voice file generation
« Reply #3 on: October 06, 2010, 03:20:36 PM »
How is that "misuse"? Why should someone be disallowed from having the .rockbox folder voiced? I don't go into it often, but I do occasionally and I imagine were I not sighted, under those circumstances I'd want the voice files there to be present.
Logged

Offline Lope

  • Member
  • *
  • Posts: 30
Re: Rockbox and Voice file generation
« Reply #4 on: October 12, 2010, 05:47:23 PM »
Oh. OK.
Logged

  • Print
Pages: [1]
« previous next »
+  Rockbox Technical Forums
|-+  Installation / Removal
| |-+  Rockbox Utility
| | |-+  Rockbox and Voice file generation
 

  • SMF 2.0.17 | SMF © 2019, Simple Machines
  • Rockbox Privacy Policy
  • XHTML
  • RSS
  • WAP2

Page created in 0.082 seconds with 15 queries.