Full changelog:
openvibe-applications/skeleton-generator
* ovsgCBoxAlgorithmSkeletonGenerator.cpp : fixed a crash when adding several times the same codec (mantis #0000107)
openvibe-applications/acquisition-server:
* acquisition-server/interface-Emotiv-EPOC.ui : added a spinbutton for the headset user ID
* ovasCConfigurationEmotivEPOC : added a reference to store the headset user ID for the driver
* ovasCDriverEmotivEPOC.cpp : added the user ID reference to the config call. This patch allows use of multiple headsets on the same computer. (forum :
viewtopic.php?f=5&t=565)
+ added a patch made by Anton Andreev to enable the driver to handle the triggers from the device
* interface.ui : added a combobox to select the NaN replacement policy
* ovasCAcquisitionServer : Added the NaN value replacement mechanism (mantis #81). User can choose to replace NaN values by the last correct value (default behavior) or zeros - or disable it.
* ovasCAcquisitionServerGUI : added NaN replacement mechanism. The Preferences are now saved into the application local configuration file (Drift and NaN mechanisms, oversampling factor, etc.)
* ovasCAcquisitionServer : replaced the NaN detection routine with _isnan and added infinite value detection with !_finite
* ovasCAcquisitionServer : bugfix for Linux compatibility.
* ovasCAcquisitionServer : fixed a compilation warning with gcc (order of initialisation for the class members.
* fixed a bug with encoding in the TMSi driver
- reverted to revision 3240, removing the uncomplete patch to add the event channel to gtec driver
+ merged branch wip-jlegeny-mitsar -r3253:3254 into trunk resulting in following changes:
+ added files for the Mitsar driver made in Gipsa-lab
+ added a closed DLL which we have rights to distribute (approved by Mitsar company)
* merged branch wip-jlegeny-gtec-event-channel -r3256:3261 into trunk resulting in following changes:
+ added the first code to support gTec USBamp event channel
+ added code by Anton Andreev to enable the additional Event Channel for the gTec USBamp device
* added the actual constant for number of channels
* bugfix from Anton Andreev, fixed configuration
* merged branch wip-jlegeny-fieldtrip -r3259:3284 into trunk resulting in following changes:
This driver was developed as a joint effort of CEA and Gipsa-lab. The code of third-party fieldtrip buffer is licensed under the BSD license according to the COPYING file included in the project at :
https://code.google.com/searchframe#BFY ... er/COPYING
+ added files for the fieldtrip driver and modified the cmake file
- removed unnecessary logging from the fieldtrip driver
* fixed some warnings in the acquisition server
* changed the license included in the file extern.c according to the COPYING file superseding it at
https://code.google.com/searchframe#BFY ... er/COPYING the license of the fieldtrip buffer chosen is BSD, compatible with the OpenViBE LGPLv2 license
openvibe-applications/designer:
* the unstable boxes are read everytime
+ added a checkbox to show/hide unstable boxes
+ the searchbox now searches in descriptions as well
* patched the designer to not crash when opening non-existing files
* fixed a potential bug
* fixed the empty list of boxes when loading an inexisting scenario
openvibe-toolkit:
* fixed building of codecs with GCC4.7
openvibe-kernel:
* fixed building of the kernel with GCC4.7
scripts:
+ add dependancies to libXaw-devel in Fedora; disable CEGUI samples
* fix log write permission on fedora install with su- c 'yum..'
+ add tee to view password prompt in fedora su -c .... command
+ Add end message when finish. Needed for automatic build end detection
- delete make OpenViBE-documentation in build loop to fix "rule not found" in all modules exept doc
+ linux-install_dependencies : added libncurses5-dev for ubuntu 12.04
* bumped up the version of boost library to 1.49
+ added a dependency to pthreads on windows
* win32-generate-vc-proj.cmd : fixed the old "recall" tweak to a proper solution. Fixed the script to generate VS2010 projects.
- removed the dependency to pthreads in windows, moving to the community forge
* added a dependency of python-devel on Fedora in order to compile the Python box by default
* force OpenViBE-documentation as default rule in compilation
+ added python-dev dependency on ubuntu linux
+ re-added the phtreads dependency
* fixed the skeleton to take this dependency into account
openvibe-kernel-omk:
* openvibe.conf : added default value for token AcquisitionServer_NaNReplacementPolicy.
* fixed building of the kernel with GCC4.7
openvibe-scenarios:
+ added a tutorial for Sign Change Detector Box
+ added a tutorial for Cue Image Display Box
+ box-tutorials/matlab : added 3 new tutorials for the matlab scripting box (matlab functions and scenarios)
+ added scenarios for python box
openvibe-plugins/matlab:
* ovpCBoxAlgorithmMatlabFilter : moved the source code to the src root, as it's not in the "tool" category anymore.
+ ovpCBoxAlgorithmMatlabScripting : a general purpose box (support IO/settings types).
+ ovpCMatlabHelper : matlab helper class for the matlab scripting box. Does the interface between OV and Matlab Engine.
* change by Joan Fruitet - added warning for stimulations arriving too late
* consider stimulations arriving at the end time of a chunk
+ doc/Doc_BoxAlgorithm_MatlabScripting.dox-part : Matlab Scripting box documentation page.
* ovpCBoxAlgorithmMatlabScripting.cpp : fixed a bug when a fail opening of the matlab engine sometimes returned a non-null value as handler. Testing it at uninitialize could then produce a crash.
* ovpCBoxAlgorithmMatlabFilter.h : moved this box in the "Scripting" category
* ovpCBoxAlgorithmMatlabScripting.h : moved this box in the "Scripting" category
openvibe-plugins/python
* fixed wrong sample rate
* fixed crash after multiple loading of boxes with numpy/scipy dependencies
* fixed the un-clean unloading when initialization failed
* applied patches to move objects out of the main dictionnary
* fixed a bug with (de)initialization
+ added Laurent George to the list of contributors
* modified the box name
* ovpCBoxAlgorithmPython.h : moved this box in the "Scripting" category
* merged branch wip-python -r3196:3217 into trunk
openvibe-plugins/stimulation
* ovpCBoxAlgorithmLuaStimulator.h: moved this box in the "Scripting" category
* Doc_BoxAlgorithm_SoundPlayer.dox-part : minor fix
* ovpCBoxAlgorithmOpenALSoundPlayer.cpp : fixed an issue with stimulation output not sending empty sets, resulting in unusable stream afterward.
+ added a box detecting the sign of an input. Original box by Joan Fruitet
* fixed a compilation bug on windows in Sign Change Detector
+ added a log message when an undefined key is pressed
* ovpCBoxAlgorithmStimulationMultiplexer.h : added short and detailed description
* ovpCBoxAlgorithmLuaStimulator.h: moved this box in the "Scripting" category
openvibe-plugins/file-io
* electrode_set_standard_cartesian.txt : added AF3;AF4;FC5;FC6 for Emotiv EPOC compatibility
* fixed (bug #0000104) a bug with 0 length strings in gdfFileWriter (thanks to Anton Andreev)
* ovpCBoxAlgorithmSignalConcatenation.cpp : bugfix patch from Dieter Devlaminck. While waiting for all the stream headers, some chunks were lost (see forum post
viewtopic.php?f=10&t=411&p=3607#p3607)
openvibe-plugins/simple-visualisation:
*added patch from Joan Fruitet to fix stimulation dates to fit the actual visual stimulation
* set the Display Cue Image Box as stable
* ovpCBoxAlgorithmP300MagicCardVisualisation : changed 6th box setting name from "card filename" to "default background filename". Default value is openvibe logo (offscreen version). If this setting is empty, then the background image for a card is automatically "cardname-offscreen".
* doc/Doc_BoxAlgorithm_P300MagicCardVisualisation.dox-part : updated the documentation page to explain this behavior.
openvibe-plugins/python:
* ovpCBoxAlgorithmPython.h : changed default clock frequency from 60 to 64
openvibe-plugins/signal-processing:
* ovp_defines.h : added an identifier for a new input parameter for the SBE Algorithm.
* ovpCAlgorithmStimulationBasedEpoching : added an input parameter so the algorithm know the last chunk end time to process
* ovpCBoxAlgorithmStimulationBasedEpoching : bugfix for negative offset values (see bugtracker #082)
+ add first version of signal enveloppe box using Hilbert Transform. It depends on itpp::fft and itpp::ifft