VoxForge
Speaker Characteristics:
Gender: male;
Age range: adult;
Pronunciation dialect: General American English.
Recording Information:
Microphone make: Sennheiser PC 131;
Microphone type: noise canceling headset;
Audio card make: Andrea USB adapter;
Audio card type: USB;
Audio Recording Software: Audacity 1.2.6;
O/S: Windows XP Professional.
File Info:
File type: FLAC;
Sampling rate: 48kHz;
Sample rate format: 16bit;
Number of channels: 1;
Audio Processing: no
--- (Edited on 11/25/2007 7:04 pm [GMT-0600] by ralfherzog) ---
Notice: many prompts in "English Speech Files" were adapted from the prompt files contained in the CMU_ARCTIC speech synthesis database, which were in turn derived from out-of-copyright texts from Project Gutenberg, by the FestVox project at the Language Technologies Institute at Carnegie Mellon University. |
Hi Ralph,
You've been busy! Thanks for the last bunch of submissions.
The Acoustic Model training script had a little trouble with one of your entries:
en22-69 She wants full immunity from prosecution.
Which I removed from the version in the VF Speech Corpus. It's strange ... it sounds OK to me, but caused the training process some problems.
Here is the link to the version in the Corpus:
ralfherzog-20071126-en22.tgz 28-Nov-2007 15:28 12.7M
Ken
--- (Edited on 11/28/2007 9:08 pm [GMT-0500] by kmaclean) ---
Notice: many prompts in "English Speech Files" were adapted from the prompt files contained in the CMU_ARCTIC speech synthesis database, which were in turn derived from out-of-copyright texts from Project Gutenberg, by the FestVox project at the Language Technologies Institute at Carnegie Mellon University. |
--- (Edited on 11/29/2007 12:03 am [GMT-0600] by ralfherzog) ---
--- (Edited on 11/29/2007 12:07 am [GMT-0600] by ralfherzog) ---
Notice: many prompts in "English Speech Files" were adapted from the prompt files contained in the CMU_ARCTIC speech synthesis database, which were in turn derived from out-of-copyright texts from Project Gutenberg, by the FestVox project at the Language Technologies Institute at Carnegie Mellon University. |
I personally think that you shouldn't let the fact that there is no one to process your speech files stop you from submitting more files. The more files (German) we have the more interesting it becomes for someone to help us process them (make a dictionary etc.)
Perhaps you can persuade someone from the Simon Project to help us out?
Maybe you should also write a small text on "Spracherkennung", Linux, open source and VoxForge. That way we might get more German visitors. You can publish something like that on the dev pages.
I have done that as well to attract more Dutch visitors. So far without success, however it is now possible to find VoxForge using some Dutch search terms.
--- (Edited on 11/30/2007 11:56 am [GMT-0600] by Robin) ---
Notice: many prompts in "English Speech Files" were adapted from the prompt files contained in the CMU_ARCTIC speech synthesis database, which were in turn derived from out-of-copyright texts from Project Gutenberg, by the FestVox project at the Language Technologies Institute at Carnegie Mellon University. |
Hi Robin,
Thanks for your answer.
Maybe you are able to help me out? You seem to be able to understand how to build a dictionary for the Dutch language. And you had submitted some prompts in the German language. So this could be some kind of teamwork: I submit a lot of prompts in the German language, and you could help me to process them.
Greetings, Ralf
--- (Edited on 11/30/2007 4:45 pm [GMT-0600] by ralfherzog) ---
Notice: many prompts in "English Speech Files" were adapted from the prompt files contained in the CMU_ARCTIC speech synthesis database, which were in turn derived from out-of-copyright texts from Project Gutenberg, by the FestVox project at the Language Technologies Institute at Carnegie Mellon University. |
Hi Ralf,
Timo has started some work on the German phoneset on the VoxForgeDev wiki, and was looking for feedback. This would be an excellent place to start in creating a German Pronunciation dictionary.
Creating a (GPL) German Phoneset is an important step to creating a (GPL) German Pronunciation dictionary. Essentially, what you can do is use this GPL Germane Phoneset in a Text-To-Speech Engine (like eSpeak, or Festival) to output the phonemes for a word (rather than speaking them), so you can use the TTS engine to give you a good first pass pronunciation dictionary for a list of words. This is similar to the approach I used to find the pronunciations for out-of-vocabulary words for the LibriVox text I segmented.
I believe I sent you a password a while back to for the German Trac site on VoxForge ... if not, let me know, and I can send you one (I still have not got around to fixing mod_security on the Web server to allow updates without signing-on and keep comment spammers out...).
Ken
--- (Edited on 11/30/2007 9:22 pm [GMT-0500] by kmaclean) ---
Notice: many prompts in "English Speech Files" were adapted from the prompt files contained in the CMU_ARCTIC speech synthesis database, which were in turn derived from out-of-copyright texts from Project Gutenberg, by the FestVox project at the Language Technologies Institute at Carnegie Mellon University. |