BAS - SpeechRecorder release notes
-
- Windows installer vulnerability (WiX toolset CVE-2024-29188 vulnerability)
- Update to OpenJDK 17.0.10+7
- Session remains in active state if an error occurs when starting playback
- Speaker XML database file contains empty informedConsents element
- Update to OpenJDK 17.0.8.1+1
- Improved security with XML file handling
-
- Bug fix ID0118 "The update manager does not find the matching installation package for a certain system architecture."
- Added installation package for macOS on Apple Silicon processors M1,M2. This versioon runs without Rosetta2 emulation.
- Update dependency libraries.
-
- Security updates of Apache Batik (SVG prompting support) and the Java runtime.
-
- Activated combined plain text and audio presenter.
-
- New prompt phase 'prerecordingonly' which shows the prompt only during the pre recording phase (traffic light: red/yellow).
- Update Java runtime to 17.0.2_8
-
- Added missing DLL file in Windows installation.
- Use Java 17 for Windows installation.
-
- Added border start stop signal.
-
- Bug fix ID0117 "GUI problems with exceptions with scripts that contain at least one media prompt with a source URL whose target file does not exist."
-
- Bug fix ID0116 "Audio source exception using beep with default beep audio file"
-
- Added menu item to export recording script.
-
- Mitigate bug ID0114 "Incorrect appearance with macOS dark theme"
-
- Manual / quick start update.
-
- Removed unused attribute in script DTD
-
- Formatted text prompts (without requirement for external SVG,RTF or HTML prompt media files)
- Separate prompt presenter to show (animated) GIFs
- Open project from command line.
- Extended workspace projects panel context menu
- Alphabetically ordered project list
-
- ID0112 2020-02-12 "HTML prompts do not work on macOS. Prompt display appears empty"
- Use Java AdoptOpenJDK LTS version 11.0.6_10
-
- ID0111 2019-11-20 "The TextGrid generated by SpeechRecorder lacks a space between `intervals` and `[0]`. It should be `intervals [0]:` not `intervals[0]:`."
- Back to Java AdoptOpenJDK LTS version 11. (11.0.5_10)
-
- ID0109 2019-07-08 "Export to EMU-DB: The level definitions do not configure the corresponding required attribute definition with the same name."
- ID0110 2019-07-08 "Export to EMU-DB: Directory chooser duplicates last part of path when actually the empty content of the directory was selected."
-
- ID0106 2019-07-02 "Wrong link in project references editor into manual."
- ID0107 2019-07-02 "Button to continue in autorecording mode was not localized."
-
- ID0103 2019-06-16 "Reset to defaults in prompt
configuration UI resets to auto prompt play to configuration
default (true) instead of to configuration for new projects
(false)"
- ID0104 2019-06-27 "The 'Order' 'randomized'
should not be displayed and selectable in the script UI editor."
- ID0105 2019-06-27 "EMU-DB export: The annotations of a
bundle are not shown in EMU-webapp hierarchy if a level is missing
in its annotation, e.g if no annotation template exists."
-
- Fixed: Script editor drops media source URL of second media
item if the script is not immediately saved.
- Fixed: Random crashes on start of videos on Windows.
- Fixed: Seamless recording together with "Advance to next
recording" does not work.
- Linux version uses same packaged JRE as for Windows and
macOS (AdoptOpenJDK JRE 12.0.1+12).
-
- Support audio device names containing non-ASCII characters
on Mac OS X. (Not tested yet)
-
- Application updater uses encrypted https instead of http to
fetch list of available updates.
-
- Fixed: Mac OS X 10.14 Mojave: Recordings have no sound
level. SpeechRecorder application did not request microphone usage.
(Fix not tested yet.)
-
- Fixed: Script editor did not work for prompt items with
multiple media items in arbitrary order.
- Fixed: Script editor always used resource directory of the
first project for which the script was edited.
-
- Recording items can be grouped.
- Project export as EMU Speech Database Management System
(EMU-SDMS) database.
- Project export optionally without session data (speaker
database and recordings).
- Recording script prompt instructions and comment can be
imported from text table.
- Audio playback selection bar.
- Navigation through items is possible during playback now.
Playback will stop.
- Technical note: Windows and Mac OS X version run on OpenJDK
11 now
- Technical note: Refactoring of internal structure
- Technical note: Windows MSI installer has changed
- Technical note: Apache Batik SVG library update to version
1.10 (Mac and Windows, Linux uses version of OS)
-
- WARNING! Removed multi-channel capable audio interface on
Windows!
-
- Fixed: ID0085 Recording session info file is not written to
session directory in seamless and capture session scope mode
- ID0086 Wrong information in recording session info file
regarding recording sequences
- ID0087 Duplicate display of ID in speaker data view
- Audio controller is stopped when the script is edited or
the speaker database is displayed.
-
- Removed obsolete menu item in settings menu.
-
- Fixed: CoreAudio adapter is now working with multi-stream
devices.
-
- Fixed: ID0082 Level meter shows wrong levels if big endian
encoded output wave files
- Fixed: ID0083 Audio format not supported using Mac OS X
CoreAudio and high end audio interfaces
-
- Fixed: ID0081 Level meter does not show levels for advanced
audio formats (multiple channels, high resolution)
- Fixed: ID0080 Plain text prompt presenter ignores right to
left text direction (e.g. for Arabic text) for the words.
- Fixed: ID0079 Mac OS X CoreAudio audio interface adapter:
Only supports sample sizes given as physical sample size of audio
device. (Often 24-bit for USB audio interfaces)
- Fixed: ID0078 Mac OS X CoreAudio audio interface adapter:
Is sometimes not able to set the project sample rate
(asynchronously)
-
- Fixed: ID0077 CoreAudio audio interface adapter does not
work in full duplex mode
-
- Fixed: ID0074 Auto annotation using only the TextGrid file
persistence causes a display error
- Fixed: ID0075 Image prompts are rendered with low quality
if images have low resolution
- Fixed: ID0076 Auto annotation writes empty Praat TextGrid
files which cannot be handled by Praat
- Fixed: ID0071 Media prompt errors leave SpeechRecorder in
an undefined state. Experimenter can only quit the application.
- Fixed: ID0070 and ID0072 Some error messages appeared on
the speaker monitor instead of experimenter monitor.
-
- Added auto annotation persistence for MAUS processing
(template text file without filename suffix).
-
- Fixed: Prompt instructions cannot be completely deleted in
script editor UI.
- Improved video prompt presenter (Windows only).
- Improved playback for some audio devices (stop during
drain).
- Recording script prompt file can be selected manually.
-
- Fixed: Missing DLL redistributable.
-
- Added project configuration option to enable start and stop
of media prompts while recording.
-
- Video prompts (Windows only) can be started and stopped by
mouse click.
-
- Fixed: ID0054 Wrong annotation file names in non overwrite
(versioning) mode
- Fixed: ID0056 Recording with limited duration does not stop
in capture scope session mode when media prompt (audio,video) is
shown.
- Limited support for video prompts on Windows
-
- Fixed: ID0050 On Mac OS X the speaker window is not
automatically displayed on a connected external display.
- Fixed: ID0051 Recording script volume control for audio
prompts has no effect.
- Fixed: ID0052 No sound output for audio prompts or beep on
Linux using PulseAudio.
- Fixed: ID0053 Help window appears on external connected
screen.
- Adjustable gain for prompt beep audio file.
- Selectable prompt beep audio file.
-
- Fixed: ID0047 After project configuration change
SpeechRecorder always loads the speaker with the highest ID and not
the speaker loaded before.
- Fixed: ID0048 Recording script is not randomized when
speaker has changed.
-
- Fixed: High CPU Load when capturing from some Linux
devices. (Bug ID 0045)
- Fixed: Recording script cannot be saved. (Linux/Mac OS X
only)(Bug ID 0046)
-
- Fixed: If a project is immediately loaded after renaming in
the workspace view, it tries to load the script by the old script
file path and fails. (Bug ID0043)
- Fixed: Could not rename projects with blanks or other
special characters in the name. (Bug ID0044)
-
- Projects can be configured to hold the capture device open
during the recording session. This is the default for new projects
beginning with this version.
- New XML based file format for speakers database.
- Speaker database viewer shows speaker/session ID (first
column).
- Text table export capability for recording script and
speaker database
- Prompt text can be written to annotation file. Supported
formats are EMU JSON, Praat TextGrid and plain text.
- Fixed: Project configuration panel always shows project
name of initially opened project. (Bug ID0040)
- Fixed: Unchecked configuration property 'Reset peak on
start of recording' does not work. The level meter peak hold is
always reset. (Bug ID0041)
- Fixed: Peak hold reset button does not work. (Bug ID0042)
-
- Fixed: Capture using channel routing did not request the
line with the adapted format. Error messages "Invalid format" may
occur.(Bug ID0038)
-
- If last section is of type autorecording recording progress
is not stopped at the end of the script. (Affects only previous
2.12.x versions) (Bug ID0032)
-
- Changed calculation of acceptable minimum recording length:
Taking audio interface line activation delay into account (250ms
default, 1250ms on Mac OS X).
-
- Fixed: Mac OS X: Speaker database viewer or script editor
have odd GUI behavior (different colors, selected buttons do not
show their selected state) when opened for second time (Bug ID
0031)
-
- Fixed: Speaker index counter is not reset on project close
(Bug ID 0028)
- Fixed: Backups files were not updated if they exist (Bug ID
0029)
- Fixed: Integrated HTML help search function returned wrong
(outdated) results (Bug ID 0030)
- Improved level meter
- Improved speaker database UI
- Script editor: Multiple prompt items can be selected and
their equal properties can be edited
- Improved sonagram: Added controls for DFT size,window
size,max. frequency,dynamic range and preemphasis
- New sonagram defaults: Max. frequency: 5kHz, dyn. range:
70dB, preemphasis: 6dB/octave
- Added pedestrian lights as alternative start stop signals
- Workaround for likely bug in some multichannel USB drivers
which causes occasionally empty recordings: Experimenter is
prompted if recording is much shorter then expected
- Fixed potential bug which could cause shortened recordings
-
- Fixed: In Audio device chooser expert view the add button
is sometimes not enabled. (Bug ID 0026)
- Fixed: File handle of project configuration was not closed
after save.
-
- Fixed: Update manager cannot handle different OS versions
-
- Fixed: Import of already existing media file to resources
will delete file (empty,zero length)! (Bug ID 0022)
- Fixed: Wrong display of Khmer characters on Windows 7 (Bug
ID 0021)
- Fixed: Script writer does not generate pretty (indented)
XML. (Bug ID 0018)
- Fixed: Wrong effective recording mode display in progress
viewer e.g. autorecording, if a default project setting is used.
(Bug ID 0017)
- Fixed: Speechrecorder tries to parse hidden project files.
Project cannot be opened. (Bug ID 0015)
- Fixed: Speechrecorder can no longer be operated if user
sets non number values to number input fields like postrecording in
script editor and commits with OK (Bug ID 0014)
- Plain prompt texts are displayed with anti-aiasing
- For prompt, instructions and description font alternative
font families can be configured. (related to bug ID 0021)
- Some project configuration panels can be reset or reset to
defaults.
- Input channels of soundcard can be arbitrary routed to
channels of recording file. Playback output channels can be shifted
for multichannel soundcards
- Postrecording phase (yellow traffic light) can be forced
-
- Applied workaround: SpeechRecorder opens too many audio
file handles (Bug ID0012)
- Fixed: Audio prompter did not close audio prompt files (Bug
ID0013)
-
- Fixed: Build of 2.8.0 for Mac OS X was broken.
(Recording,project preferences and table import did not work)
- "About" splash screen does not appear on startup anymore.
-
- Added itemcode generator to script editor.
- Audio device for prompt and beep playback is configurable.
- Behaviour of audio prompts is more conforming to 2.2.x
versions.
- Improved handling of speaker display frame or window.
-
- Fixed: Dual monitor mode does not work properly. Speaker
addressed window does not appear or maximize. (Bug ID 0006)
- Fixed: Special characters in project path are not
supported. (Bug ID 0005)
- Fixed: Incorrect display of instruction numbering with
prompt phase set to recording. (Bug ID 0004)
-
- Fixed: Non recording items do not work.
-
- Fixed: Scripts with empty metadata key or value could not
be saved.
- Fixed: Performance issues with GUI script editor.
-
- Update notification for the installation packages.
- Integrated Standard and AJS versions into one installation
package
- Element 'nonrecording' is implemented.
- The attribute 'beep' is implemented.
- The signal display is able to display multichannel 24/32bit
recordings.
- Recording script editor.
- Improved control over audio prompts.
-
- Fixed in installation package build 03: DSJavaSound library
update to 0.2.1: COM library was not correctly initialized, caused
truncated to 31 chars capture names. (Standard JavaSound still has
this problem)
- Fixed: Speaker data editor did not accept the value of an
active cell editor.
- Up to 4000ms audio line buffer size with AJS (Alternative
JavaSound).
- New Speechrecorder icons!
- Speaker data editor accepts editing of all columns even if
the speaker code was not entered.
-
- Fixed: Capture engine does not try to read from audio line
if line is stopped.
- Fixed: Logfile handlers were not closed on when a project
was closed.
- Fixed: Reset peak (hold) on start of recording option in
project configuration has no effect.
-
- Added image MIME type PNG (image/png).
- Fixed: Error message "Cannot associate element ..." with
turkish (and other) locales.
-
- Numbering of instructions is switchable.
- Fixed: "null" strings will not be inserted for empty fields
in speaker database
- Fixed: Column entries after skipped column entries moved
over to the skipped position
- Fixed: Transport stop button label was not always correctly
set for manual mode
-
- Fixed: Filehandle of the script was not closed. (Windows
users could not edit the script, while Speechrecorder was running)
- Visibility of action buttons is configurable now.
- Progress viewer table scrolls automatically to current
item.
-
- Fixed: Speakerdisplay settings of project configuration are
not ignored anymore.
- Fixed: Section mode, prerecording and postrecording delay
settings of project configuration are not ignored anymore.
- Fixed: Timelogging of prerecording,recording,postrecording
was broken.
- Fixed: Export on Windows with wrong file separator ("\"
backslash).
- Fixed (hopefully): Some audio events caused deadlock
hangups.
- Speaker database is saved if changed to avoid dataloss on
hangup.
- Increased audio line buffer sizes.
- Improved reliability.
-
- Formatted text (HTML,RTF) is aligned in the middle same as
plain text.
- It is possible now to write HTML and RTF prompts directly
to the recording script in a CDATA section.
- Added project export feature
- Fixed: Removed ALSA workaround in capture engine.
-
- Fixed: Switched off keyboard control on speaker/project
editing
- Added audio prompting feature
- Added library version infos
-
- Fixed: hangup using keyboard control
- Session ID is visible in Speaker window
- Added Java Runtime and script file path info
-
- Added keyboard control capability
-
- Java 5.0 required
- Single JAR file splitted into three JAR libraries
- New audio controller interface
- Improved level display
- Running with incremental garbage collector
- Logging and timelogging configurable