Activity
From 15.02.2016 to 15.03.2016
10.03.2016
- 12:05 Task #3796: Add more info about alternative candidates
- Is it possible to show word boundaries (to differentiate between word internal and cross-word diphones) in a similar ...
09.03.2016
- 16:21 Task #3813 (Closed): Submit a paper for INTERSPEECH2016 (D)
- *Title:* Voting Detector: A Combination of Anomaly Detectors to Reveal Annotation Errors in TTS Corpora
*Publicati... - 15:52 Task #3684 (Closed): Redmine issues
- 15:39 Task #3811 (New): Experiment with statistical outlier detection and removal
- 15:36 Task #3810 (Closed): Submit a paper for TSD2016 (D)
- *Title:* Correction of prosodic phrases in large speech corpora
*Result type:* D
(Abstract to be sent by March ... - 15:21 Task #3809 (Closed): Submit a paper for INTERSPEECH 2016
- 15:20 Task #3808: Submit a paper for TSD2016 (D)
- Abstract sent.
- 15:19 Task #3808 (Assigned): Submit a paper for TSD2016 (D)
- 15:14 Task #3808 (Closed): Submit a paper for TSD2016 (D)
- *Title:* On the Influence of the Number of Anomalous and Normal Examples in Anomaly-Based Annotation Errors Detection...
- 15:10 Task #3796: Add more info about alternative candidates
- concatCost targetCost nextConcatCost
prosodemeType
leftCntx - rghtCntx
begWeight | midWeight | endWeight
sentence - 14:30 Task #3796 (Feedback): Add more info about alternative candidates
- Thanks,
please write here what do the numbers mean. I guess that the 4th number stands for a prosodeme but what abou... - 14:02 Task #3796 (Resolved): Add more info about alternative candidates
- Done,
more info added
word the unit came from - that would require some work; so for now this information is no... - 09:04 Task #3796 (Closed): Add more info about alternative candidates
- When listing the alternative candidates, only three numbers (btw. what do the numbers mean?) and source utterance are...
- 15:08 Task #3757 (Closed): Simultaneous training of single detectors within a voting detector framework
- 15:07 Defect #3767 (Closed): Cannot jump to source units for some voices
- 14:54 Task #3807 (Closed): [á] sounds like [a] in "Ahoj, jak se máš?"
- *Voice:* Jan
*Description:* @[á]@ sounds like @[a]@ in _Ahoj, jak se m<notextile></notextile>*á*<notextile></notex... - 14:47 Task #3806 (Closed): "mlýneček" sounds like "mlýmeček"
- *Voice:* Jan, Iva (maybe more)
*Description:* Word _mlýneček_ sounds like _mlý<notextile></notextile>*m*<notextile... - 13:22 Task #3780 (Resolved): Testing placement and fixing of senzors
- *The third experiment*
This time I glued all seven sensors. Firstly, I placed those on tongue (TD->TB->TT), then on ... - 11:36 Task #3797 (New): Artifact catalogue
- _(Parent task for cataloguing of artifacts)_
08.03.2016
- 23:36 Defect #3767 (Resolved): Cannot jump to source units for some voices
- done, all voices were imported
Moreover, the editor was enhanced with new features
01.03.2016
- 14:40 Task #3780: Testing placement and fixing of senzors
- *The first experiment*
I tested three senzors attached to a tongue, namely to tongue tip (TT), tongue body (TB) and ... - 13:43 Task #3780 (Closed): Testing placement and fixing of senzors
- 13:38 Task #3697 (Assigned): Electromagnetic articulography (EMA) based research
- 13:31 Task #3763 (Assigned): Synthesize utterances with phonetically incorrect phoneme realizations
26.02.2016
25.02.2016
- 14:42 Task #3773 (Closed): Submit a paper to journal Measurement Science Review
- *Title:* GMM-based Speaker Age and Gender Classification in Czech and Slovak
*Abstract:*
The paper describes an ... - 14:39 Task #3772 (Closed): Submit a paper for TSP 2016
- *Title:* Comparison of one and two-level architecture of the GMM-based speaker age classifier
Paper submitted, wai... - 14:36 Task #3734 (Closed): Evaluate analysis of unexpected results of listening test on voiced/unvoiced "ř"
- 14:23 Task #3734 (Resolved): Evaluate analysis of unexpected results of listening test on voiced/unvoiced "ř"
- Analysis shown that the unexpected behavior occurred mostly due to the use of other units which were chosen to minimi...
- 13:52 Task #3771 (Closed): Create a paper for TSD 2016
- Create a paper for TSD 2016 describing experiments with the evaluation of TTS by GMM-based speaker gender and age cla...
- 13:49 Task #3770 (Closed): Experiment with gender and age classification for synthetic speech error predictions and evaluation
- _Parent task for research in the field of alternative approaches to synthetic speech evaluation (with the aim to repl...
- 11:43 Defect #3768 (Resolved): Alternative candidates are not available
- Fixed
24.02.2016
- 21:41 Task #3746 (Closed): Enable to mark artefacts in WebProkus
- 21:40 Defect #3768 (Closed): Alternative candidates are not available
- Alternative candidates are not available - the panel to show the candidates opens but the message "probíhá výpočet" i...
- 21:35 Defect #3767 (Closed): Cannot jump to source units for some voices
- For some voices in WebProkus (e.g. voice Iva or Katerina but probably also for other ones), source units (or whole ut...
- 16:22 Task #3761 (Resolved): Create script for conversion ASF to SNT
- Script for this task is placed in SVN repository @ARTIC_UTILS/trunk/hmm_synth/LabLight/asf2snt.py@.
Created SNTs are...
23.02.2016
- 17:07 Task #3763 (Closed): Synthesize utterances with phonetically incorrect phoneme realizations
- Synthesize utterances from #3747, #3748, #3749, #3750, #3751 with phonetically incorrect phoneme realizations.
- 15:07 Task #3751 (Closed): List of words containing voiced / unvoiced "ch"
- 15:04 Task #3750 (Closed): List of words containing voiced / unvoiced "ř"
- 15:04 Task #3749 (Closed): List of words containing alveolar "n"/ velar "N"
- 15:03 Task #3748 (Closed): List of words containing syllabic / non-syllabic "l"
- 15:02 Task #3747 (Closed): List of words containing syllabic / non-syllabic "r"
- 12:26 Task #3754 (Closed): WebProkus feature request: Export TextGrid
- 01:04 Task #3754 (Resolved): WebProkus feature request: Export TextGrid
- fixed
- 09:52 Task #3761 (Closed): Create script for conversion ASF to SNT
- * Create script for conversion ASF to SNT.
* Create new SNT files from extended ASFs (#3709) and compare default and...
22.02.2016
- 21:10 Task #3750 (Resolved): List of words containing voiced / unvoiced "ř"
- Please synthesize the following phrases with the conflicting voicing information in Ř.
hořící keř
zuřivá bouřka
... - 21:01 Task #3751 (Resolved): List of words containing voiced / unvoiced "ch"
- Please synthesize the following sentences with the voicing status of the velar fricative ([x] vs. [ɣ]) in conflict.
... - 20:59 Task #3749 (Resolved): List of words containing alveolar "n"/ velar "N"
- Please synthesize the following phrases in which the place of articulation of the nasal - alveolar vs. velar - will b...
- 20:54 Task #3748 (Resolved): List of words containing syllabic / non-syllabic "l"
- Please synthesize the following phrases in which the syllabicity of L will be in conflict with reality.
Kripl vypl... - 20:52 Task #3747 (Resolved): List of words containing syllabic / non-syllabic "r"
- Please synthesize the following phrases in which the syllabicity of R will be in conflict with reality.
Kapr to na... - 18:59 Task #3754 (Feedback): WebProkus feature request: Export TextGrid
- Bug: improper format. Labels have to be in double quotes, e.g.
0
0.215875
"$R"
0.2159375
0.2995625
"Re"
Th... - 14:14 Task #3754 (Resolved): WebProkus feature request: Export TextGrid
- Done, please try it
(it is in the "File" menu) - 17:45 Task #3746 (Resolved): Enable to mark artefacts in WebProkus
- Done.
You can mark artefact in prokus (m key). You can set score (-1 bad, +1 good) and comment.
You can view st... - 14:14 Task #3753 (Closed): Add "Kateřina" voice to WebProkus
- 14:14 Task #3753 (Resolved): Add "Kateřina" voice to WebProkus
- Done.
- 10:00 Task #3699 (Resolved): Compute features (MFCC, LPC, LPCenv and FFTpow)
- Distances computed. Starting to experiment with one-class-classifier.
21.02.2016
- 17:16 Task #3757 (Closed): Simultaneous training of single detectors within a voting detector framework
- Revise voting detector paradigm - train single detectors simultaneously within a single development framework.
- 17:14 Task #3752 (Closed): Influence of train/test splits on detection accuracy
19.02.2016
- 18:43 Task #3751 (Assigned): List of words containing voiced / unvoiced "ch"
- 13:10 Task #3751 (Closed): List of words containing voiced / unvoiced "ch"
- Send us a list of words containing voiced / unvoiced "ch".
Include both phenomena in the list, e.g.:
_abych byl... - 18:43 Task #3750 (Assigned): List of words containing voiced / unvoiced "ř"
- 13:07 Task #3750 (Closed): List of words containing voiced / unvoiced "ř"
- Send us a list of words containing voiced / unvoiced "ř".
Include both phenomena in the list, e.g.:
_bratři_
_... - 18:42 Task #3749 (Assigned): List of words containing alveolar "n"/ velar "N"
- 13:05 Task #3749 (Closed): List of words containing alveolar "n"/ velar "N"
- Send us a list of words containing alveolar "n"/ velar "N".
Include both phenomena in the list, e.g.:
_banka_
... - 18:42 Task #3748 (Assigned): List of words containing syllabic / non-syllabic "l"
- 13:01 Task #3748 (Closed): List of words containing syllabic / non-syllabic "l"
- Send us a list of words containing syllabic / non-syllabic "l".
Include both phenomena in the list, e.g.:
_vlko... - 18:42 Task #3747 (Assigned): List of words containing syllabic / non-syllabic "r"
- 12:58 Task #3747 (Closed): List of words containing syllabic / non-syllabic "r"
- Send us a list of words containing syllabic / non-syllabic "r".
Include both phenomena in the list, e.g.:
_svet... - 14:14 Task #3754 (Closed): WebProkus feature request: Export TextGrid
- Generated TextGrid in Praat Short-text format should contain boundaries and labels of diphones.
A sample Matlab sc... - 13:59 Task #3753 (Closed): Add "Kateřina" voice to WebProkus
- Add "Kateřina" voice to WebProkus.
- 13:14 Task #3752 (Closed): Influence of train/test splits on detection accuracy
- Make multiple train/tests splits and evaluate results.
- 12:27 Task #3746 (Closed): Enable to mark artefacts in WebProkus
- Typical usage:
# Use WebProkus to synthesize a sentence
# Mark segment(s) in which an artefact occurs - i.e. two ne...
Also available in: Atom