Archive for category: Release Notes

This release brings together several new features that we wanted to do for a long time, many of which depend on each other. This includes various changes to the underlying architecture of the software, which are not directly visible to the user, but which were necessary to make the new features possible.
One of those architectural changes is the whole file pitch cache. In previous versions of VoceVista, the pitch track was derived from the visible spectrogram. This means its accuracy was dependent on the current time zoom level, and this would give different results for different zoom levels. The new pitch track is maintained independently of the visible spectrogram, so it always has the same accuracy regardless of the zoom level. This allows the mean pitch and also the vibrato tracks to be calculated with more accuracy and robustness. It also allows to export the pitch track of the entire recording to a spreadsheet file, and it will allow to easily compare pitch tracks of different recordings in future versions of VoceVista.

Another exciting new feature are the Smoothed Curves. This allows averaging of the spectrum over the frequency axis, which reveals the underlying spectral envelope. This is included in VoceVista Video 5.7.0 as an experimental new feature, and it will probably evolve into an official formant tracker in future versions.

Some new features for VoceVista Video Pro are a greatly revised Analysis View that provides numerical output of many metrics, along with the new Export Data feature that allows exporting many items to a spreadsheet file that can be opened in Excel, LibreOffice, and other programs that can read .xlsx files. This goes together with the new ability to analyze a set of checked Markers simultaneously for comparison.

The UI for the Long-Term Average Spectrum has been changed, so that the LTAS is now separate from the current selection (although this can be set to the previous behaviour as explained below). This allows to work with time selections without the LTAS getting in the way when it is not needed. Similarly, the cursor crosshair has been separated from setting the current spectrum, so that these two features can be used independently.

And as usual, this new version comes with a long list of bug fixes and improvements for usability, stability, and performance.

Also, we are working hard to update the user manual for all these new changes, and to create new video tutorials. As this will take some time, we decided to release this version 5.7.0 without the new manual, and strive to release the new revised manual in the coming weeks.

New Features

Whole file pitch cache

• The pitch line is now more stable when zooming the displayed time range in and out, because it is calculated at a fixed time resolution, instead of being calculated at the current time resolution of the spectrogram.
• This allows the mean pitch and vibrato analysis to be calculated independently of the visible range of the spectrogram.
• The pitch time resolution (by default 100 points per second) can be specified on the Analyzer View / Vibrato settings page.
• When the view is zoomed-out very far, the pitch track will be drawn as points instead of continuous lines.

Cursor and Mouse Hover

• The cursor crosshair visibility is now separate from spectrum updating. So it is possible to enable the cursor crosshair, but uncheck the Enable hover spectrum option.
• The options for this can be set on the Pointer & Selection settings.
• The grey cursor lines have been reviewed and fixed to be more consistent across all views.
• The context menu entries for the cursor text display on the short-term view have been consolidated into a single entry.
• Only visible curves are now traced on the short-term view.
• The text of the toggle snapping button has been changed into Toggle Cursor and Slider Snapping, as it applies to both.

Customizable Toolbars

• The toolbars can now be fully customized through the menu Tools / Customize toolbars.
• The toolbars can be detached from the main window, and used as floating toolbars, for example to have them on a separate screen.

Settings

• The options/preferences window has been renamed into Settings.
• New tool icon for the settings window.
• Analyzer Settings: express FFT and pitch time resolutions as fractions of 1 second instead of steps/s.

Smoothed Curves

• Allow display of smoothed versions of the current spectrum, the Long-Term Average Spectrum, and the average spectra of Markers that are checked for comparison. This is different from the Long-Term Average spectrum, which applies an averaging over time. The Smoothed Curves, in contrast, apply averaging over the Frequency axis, which can be used to reveal the underlying spectral envelope. These two calculations can be combined by applying the smoothing to the average spectra.
• The Smoothed Curves can be enabled in the context menu of the Short-term View (by right-clicking on the Spectrum), and on the Analyzer View / Acoustic settings page.
• The radius of the averaging can be specified on the Analyzer View / Acoustic settings page.

Vowel Chart

• The Vowel Chart can now zoomed (with the mouse wheel), and panned (with Ctrl+Click+Drag).

Sound Generator

• new settings page, for example to specify the sampling rate of the generated sound
• improved support for profiles

Long-Term Average Spectrum

• The Long-Term Average Spectrum (LTAS) is no longer synced automatically with the selection by default. The workflow is now to make a selection, and then click on Markers / Set average spectrum to selection.
• The old behaviour, where the LTAS is automatically set when the selection changes, can be achieved by enabling the option Markers / Sync average spectrum with selection in the Markers menu, or on the Pointer & Selection settings page.
• The LTAS overlap can new be specified on the Analyzer Settings page.

Analysis View

• The workflow for using the Analysis View has been revised.
• The layout and contents of the output display have been revised and improved.
• Improved sorting options for Spectral Peaks analysis.
• The Analysis View can now analyze the average spectra of Markers that are selected for comparison as separate items.
• New option to analyze only content of selected Markers (and not of the current spectrum).
• The refresh rate of the Analysis View can be set at Advanced Settings / Display / Analysis View updates/second.
• The Analysis View control panel is now collapsible to save screen space.

File boundaries

• When zooming to fit the current selection, the viewing rectangle will now be centered around the selection, which makes the workflow more intuitive when working with very short selections.
• Separate recorded segments or generated sounds have their spectrogram and spectrum cleanly separated without bleeding into each other, for more accurate analysis.
• The spectral content at the end of the files is now displayed when not recording. This improves usability of short files and very high frequency resolutions.

Vibrato Analysis

• This feature has been revised for greater accuracy and robustness. Due to the new whole file pitch cache, the accuracy of the vibrato analysis is now independent of the current time zoom level. And the algorithm for calculating vibrato rate and extent has been improved.
• It also has a new button to select the largest time segment that encloses the current cursor position, so that such segments can easily be selected.
• The current vibrato cycle that encloses the cursor can optionally be highlighted.
• New metrics for Vibrato Jitter and Vibrato Shimmer, to calculate the measure of variability/consistency for vibrato rate and extent.

Data Export

• New command File / Export Data to open the Export Data dialog, which allows to export many items to a spreadsheet file that can be opened in Excel, Libreoffice Calc, and other programs that can read .xslx files.

Exportable items:

• Long-term:
  • Fundamental Pitch
  • Mean pitch
  • Vibrato
  • the corresponding excerpt of the audio content can be saved into a separate audio file.
• Short-term or aggregate:
  • Analysis View Result
  • Spectrum
  • Long-Term Average Spectrum
  • Average Spectra of selected Markers
  • Marker List

Cache Management

• The management of temporary and cached files has been completely rewritten to be more robust and scalable.
• The cache folder can now be manually cleared and inspected through commands on Advanced Settings / Cache Folder.

Licensing

• Multi-user editions now support floating licenses (in addition to node-locked licenses).
• Administrative license activation for all accounts has been reviewed and improved.

Miscellaneous

• Piano can set active slider.
• The text on the File / Loaded file info window can now be selected and copied.
• improve the built-in benchmark tool to measure system performance.
• display more helpful error messages in various cases, for example when the app doesn’t have permissions to access the microphone.

Fixes

• Fix typo on German translation that would impede export of note slider layouts
• Toolbar icon for Options/Preferences has been revised
• Fix: Opening a file changes its “last changed” date.
• Fix: Sound generator uses incorrect sampling rate in some cases.
• Fix: Recording file might use incorrect sampling rate in some cases.
• Fix possible crash when stopping playback.
• Fix possible crash when reloading file from previous session (which could cause multiple crashes in a row).
• Fix display of zoomed-in Waveform and Timeline on ultra-wide screens.
• Fix exporting current sound generator profile to a file.
• Fix display of divider frame between spectrogram and spectrum.
• Fix Analysis View not being refreshed or cleared in some cases.
• In Analysis View, display unavailable values as “n/a” instead of “0”.
• Improve drawing of highlighted selection background and borders in various views.
• fix comparing Markers from files with different sampling rates.
• fix possible crash in Image Export when the spectrogram time scale in turned off.
• fix display of an invalid path for new unsaved files on the Loaded File Info window.
• the notation system can be set to Scientific even when the main app language is set to German.
• fix distorted lines when drawing pitch, or spectrum outline, with certain window ratios.
• fix line width settings for various cursor lines not being observed.
• fix color of LTAS line not being observed.
• fix display of a wait cursor during various long-running operations, especially when long files are edited.

Hotfix Build 5.7.1

Based on the initial user feedback for build 5.7.0, we have fixed two issues:

• Spectrogram sometimes disappears when starting recording
• The lines in the EGG History and Short-term EGG view are drawn incorrectly

VoceVista Video 5.6.2 is a service release to improve stability, performance, and fix various bugs.

New Features

• Input Level Meter: When the app starts, the input level meter will always be off. It now has a second instance on the Recording Settings dialog, and the level meter will start when the recording settings are opened. It can also be started with a new entry “Restart Level Meter” in its context menu. Improve reliability and detection of current settings when level meter starts.
• Multiple user accounts on one machine: The license can be activated by an admin for all user accounts that use the same computer. That was previously only possible on Windows. Now it is also possible on macOS.
• Syncronize fundamental frequency of resonance matrix with focus slider:
  This allows to easily switch the fundamental frequency used in the vowel chart by having multiple overtone sliders. The vowel chart will always use the fundamental frequency of the slider that has the focus.
• Command to clear entire file:
  With Edit / Delete all the entire file can be deleted without first making a selection.
• ASIO support:
  The recording settings page now has a checkbox to enable ASIO. By default, ASIO is turned off, because it may cause stability issues on some systems with outdated drivers. So now it can be more easily enabled and used.
• Solfege Note Names: In addition to Scientific and Helmholtz notation, the note names now also support the Solfege system with Do Re Mi Fa Sol.
• Normalize Dialog: Add ‘N’ as default hotkey. Change “Current maximum” into read-only field. Add more options for target level and add loudest and quietest labels to list entries. Rearrange some entries in the edit menu to make normalization easier to reach.
• Alt + mouseWheel over the spectrogram will now zoom the frequency scale.
• Doubleclick on a Marker will now select the marker, and set the selection to the time range of the marker, but not change the time zoom level of the spectrogram.

Fixes

• Fix possible freeze when connecting USB devices
• Improve detection of compatible device settings when recording starts
• Fix second timeline not visible when windows are tiled vertically
• Piano keyboard graphics don’t immediately show change of Concert Pitch
• Sinewave playback of sliders produces fake harmonics when fundamental moves
• when there is nothing to delete, delete icon should be inactive
• Spectrogram time scale doesn’t update when moved with the mouse
• Change hotkey for selecting from cursor pos to click pos from Alt+Click to Shift+Click
• Fix spectrum still visible after sound is deleted
• Fix Backend R not recording 32 bit audio
• Fix reliability of downloading sound font for midi playback
• Fix updater on macOS downloading dmg twice and improve error handling. Also support installation outside the Applications folder.
• Fix windows uninstaller not revoking license on uninstall
• Don’t set cursor position when doing Ctrl-Click to select a slider label
• Fix Windows installer to prevent the taskbar icon being corrupt on a reinstall on Windows 11.
• When using multiple views on a file, show the instance number of a tab in the tab list in the Windows menu.
• Remove mention of DirectSound in the advanced settings on macOS.
• Fix playback stopping when moving markers or sliders during playback.
• Fix tabs not being refreshed after using File / Save to save a new file.
• Fix Alt key not being usable in marker editor to enter special characters.
• Fix drawing of marker boundaries for markers on spectrogram and timeline.

New Features

This release focuses mainly on stability, performance, and usability. The most significant new feature is the possibility to set an upper frequency limit on the pitch detection, which makes the pitch detection much more useful for working with speech. Also, there are new presets for the Analyzer Settings to toggle between working with singing, speech, and instruments.
Another important new feature is crash recovery. This will restore all unsaved recordings after a power failure, system crash, or other unplanned program exit.
On Windows, the default audio driver is now WASAPI, which offers much smoother scrolling during recording than DirectSound.

• recovery of last recording after a crash or power failure
• set highest detected pitch for better analysis of speech
• profile support for Analyzer Settings, with presets for singing, speech, and instruments
• Options/Preferences dialog is now modeless, so the app can be used while the options windows is open
• Preferences dialog layout has been overhauled for macOS to be more compact and use less screen space
• on Windows, WASAPI is now the default recording system for reduced latency and smoother scrolling during recording
• Sound Generator can generate click tracks
• Sound Generator can generate custom melodies
• profile manager shows list of existing categories, for better organization of long profile lists
• detect Windows Sleep mode to stop recording when laptop gues to sleep to prevent power drain
• improved diagnostic output in log file
• Manage License dialog links to license page to allow resetting the activation counter
• update Windows installer to use latest dependencies

Fixes

• show mean pitch analysis even when vibrato is not enabled
• enable spectrum when pitch analysis is enabled
• fix LTAS line color for colormaps with white background
• improve saving dock window layout (Vowel Chart, File/Marker List, Analysis View)
• fix context menus getting stuck
• fix crash when using OpenGL software renderer for better compatibility on old systems
• improved detection of changing USB devices
• improved support for High-DPI screens
• fix flicker on macOS when preferences dialog is opened
• fix crash when entering extreme frequency ranges
• fix crash when app window is too small
• fix potential crash when using filters with multiple views

New Features:

Pitch Detection Sensitivity

The Analyzer Settings page now has two entries to determine the sensitivity of the pitch detection: Minimum Clarity and Minimum Intensity. The intensity is a threshold in dB. If the incoming audio is quieter than this, it is rejected, even if it has a valid pitch. The clarity threshold is a measure of confidence that the detected pitch is valid, regardless of the volume of the incoming audio. Together, these settings can filter out background noise when processing speech, and will work together to keep audio that contains a fundamental pitch that is loud and clear enough.

Draw unfiltered spectrum

When using filters, the original, unfiltered spectrum is drawn as an outline:

 

Automatically install updates

New versions can now be installed with the click of one button. This should make it more convenient to keep the software up to date.

Bugs fixed:

• Fix the line thickness of the zoomed waveform.
• Image export dialog might be included in the exported image under some conditions.
• Spectrum filtering could be stuck with showing the wrong content.
• “Apply filters permanently” did not correctly use filters with custom strength values.
• Loading and re-saving mp3 files did not work.
• Textual cursor info on spectrum had the wrong position when “note name” was not shown.