Previous page (Acronyms) <------- Back to index ------->Next page (Overview)
First official release
Corrected manual, par.7.1 (“Stop after...” in place of “Shot duration”)
Progressive event count is zeroed at midnight and a new CSV statistic files is created
Improved readability in full report: the checkbox values are no more numeric 0/2 but strings NO/YES.
Full report files have now a naming convention.
Fixed .plt files to generate graph titles regularly formatted
The waterfall window can be shrunk enough to fit netbooks screens
The customizable logo set in Site Infos dialog is now showed on waterfall window too.
Fix: corrupted CSV files caused crash while stopping acquisition, during postprocessing. Fixed by skipping the malformed records.
HTML files are resources. No more need to install them.
IQ buffers have now tunable size, but default is still 100% that was the previous hardcoded value
Added notch filters, they can be added by right-clicking on ruler
Added some more checkboxes to report tab to allow generating reports with only plots included (no screenshots) and to restart acquisition automatically after report generation
Removed obsolete preference (autoclean)
Changed RTS file structure, that now is subdivided in sections. The sample RTS files have been updated to this new format.
At 00.00 of each day, the csv, plots and shots are stored in dedicate subdirectories, one for each day (Daily Archive) .
The shots/plots are subdivided in further categories/subdirectories: iperdense/underdense/fakes on a event lasting basis.
The file names contains, besides the configuration name and date, also the acquisition type (automatic, periodic, continuous) in place of the time.
Only one statistic CSV will be produced for each day for the active configuration.
In automatic acquisition, it is opened in append mode
In continuous and periodic acquisition, it is created new for each session, eventually truncating an already present one.
Changed the "Erase files before acquisition start" checkboxes: now the first one deletes only statistic CSV files, while the second regards all the other files (shots, plots and reports).
Some installation-dependent parameters previously kept in RTS files, like the windows geometries, the gnuplot path and the events count, have been moved to a local configuration file (The registry under Windows, .config files under Linux) in order to allow the use of RTS files produced by other stations without impacts on those local settings.
The "Data lasting" and the "Erase files before..." controls have been moved from Output tab to the Preferences tab.
Added in Output tab the "Join time" control. It specifies the minimum time distance between consecutive echoes to be considered as a single event. The default is 1s. The count restarts at each echo, in order to make possible to recognize long dashed patterns as a single overdense event.
The area covered by the events - meaning the area covered by all pixels having power higher than lower threshold and contiguous to the maximum peak - is added as new columns and valorized only at Fall front (in other rows is zero)
Statistic CSV: added 3 new columns:
Area echo interval: it's the area covered by the rectangle having height equal to event lasting and width equal to the detection interval. The ratio between the area covered by the echo and this area is useful to discriminate fake events.
Peaks count : counts all the peaks that crossed the higher threshold, including the maximum peak. It's another useful data for fake events discrimination.
LOS (Line Of Sight) speed of the event, calculated from its Doppler shift.
Added a control to reserve an amount of disk space to be kept free, in MB. If the available space falls below this limit, the acquisition stops spontaneously.
Thresholds controls now allow to set values with one decimal, for a finer thresholds tuning.
PPM error compensation now allows 2 decimals precision. The integer part is set in hardware, as in previous versions. The decimal part instead is used to apply a transparent offset to the waterfall.
The N filtering has been replaced by a sliding average of the last n scans, where n is the value set in spinbox.
Added an archive flattening button, that allows to specify a single directory where to copy the entire content of the hierarchical archive of the current configuration. Only the files together, without subdirectories. This is useful when building a video joining all the images.
Now it's possible to mix all the three kinds of plots in the same full report. The gnuplot data files generated in different days with different plot types can be mixed in the same full report. The previous versions instead plotted all the files using the format set when the report has been requested.
Automatic report generation: a checkbox has been added to set this feature. The acquisition stops at midnight UTC, the reports are generated, the files of the day are archived then the acquisition restarts automatically
Fixed the following bugs:
automatic report not working under linux
after an acquisition session longer than 1 day it is possible to generate a manual report (by pressing the button). This didn't work if Echoes were restarted before pressing the report button.
Fixed bug: the files in archive older than the expiration date weren't deleted.
Fixed bug: statistic CSV: when two consecutive peaks were joined, the area echo and area interval fields could assume negative values
Improved false positive filtering in automatic mode: besides event lasting, through some calculation now Echoes can discriminate most of false positives due to Graves carrier (still TODO: make editable the filter parameters) added two new preferences:
Enable event notification sound ("ping") and
Absolute thresholds (to apply the thresholds on S value instead of S-N)
The instantaneous S, N and S-N values are now showed at right-top corner of waterfall window. When taking a shot, the values showed are the peak ones, in order to match the instantaneous spectra graph at window's bottom.
There is a new of thresholds mode: besides absolute and differential (the default) the "automatic" mode (be careful about confusion with the word "automatic", already used for acquisition) has been added. In this mode the thresholds must be specified as percentage of variation of instantaneous S-N against the average S-N, resulting in mobile thresholds that can self-increment or self-decrement depending of variations of the background noise level N along the day, preventing many fake captures.
The mode can be selected in the Preferences tab. Obviously, each time this preference is changed, the thresholds values in Output tab must be adjusted.
The postprocess carriers filter, introduced in 0.25, can now be excluded from GUI by unchecking the apposite checkbox in Reports tab.
A new postprocess filter has been added, suited for lightings, artificial electrostatic discharges and other kinds of disturbances producing an horizontal pattern covering the entire waterfall's bandwidth. This filter can be excluded by unchecking the apposite checkbox in Reports tab.
In HTML report the color scale used in counting tables now matches the colorgramme scale and the zero value is rendered in black. The table orientation remains the same.
Fixed the calculation of interval areas and echo areas in CSV reports; the lightings filter is based on these values and a wrong calculation caused the filter to fail sometimes.
On the waterfall, the average S-N difference value, used as reference for automatic thresholds (1), is now displayed at the right-top corner (brown color)
Still on waterfall, at bottom side, near the event counter, the geographical coordinates of the station have been added, since it can be an useful info to keep in the screenshots. The coordinates must be set in Reports/SiteInfo dialog.
Added sample rate downsampler to perform FFT on a narrower band than the entire sample rate. The old "bandwidth" control in waterfall window is now called "zoom" because the real bandwidth, after the downsampling, is tuned in main window "FFT" tab.
Added support for device "UDP server" that is another Echoes instance whose IP address has been specified in "preferences" tab. An Echoes instance connected to an UDP server becomes an UDP client. The UDP client clones the settings in "device" tab directly from the server, while all the other controls can be changed and saved independently.
Reporting: the colors scale used in hourly count tables (Colorgramme lookalike) are no more absolute but relative to the minimum/maximum value included in the report: the black value means zero, red is the maximum value.
New command line switch -d to activate the dumb server mode: the Echoes instance launched with -d works as an UDP server forwarding samples to other Echoes listening instances without recording any data for itself.
In statistic CSV, three new columns has been added:
"Unix timestamp" to match the time labels used in plots,
"Average S" (average of S values) used by saturation fake filter
"Std.dev." standard deviation of S for future purposes
Added new FFT windows "Hann", "Flat top" and "Blackman Harris 7°"
Nicer representation of dBms scale numbers in side graphs (when compiled with Qt >= 5.12)
Fixed HzShift calculation. Now it's the difference between the frequency when an event starts and the central frequency of the waterfall.
Added sample rate 226000 Hz for dongles.
The "Start" button has been split in two buttons "Start" and "Stop" to avoid unwanted acquisition restarts
Added a new "Generate Screenshots" checkbox to enable/disable the screenshots production
Added a new waterfall label "TOO NOISY" that appears when the N value exceeds a threshold set in "preferences" tab when the label appears, Echoes continues the acquisition but stops capturing events until the noise falls below that limit.
In "preferences" tab is possible to choose an alternate "Black&white" waterfall palette, besides the classical "Color"
The I/Q samples buffer size and FFTW flags controls in "fft" tab have been removed. The first because obsoleted by the downsampler feature, the second because the FFTW dependency has been removed from the project and replaced with liquidsdr library, that implements both the downsampler and the FFT.
Removed the main window's controls locking when acquisition is active. Some changes are immediately applicable, while others become active after acquisition restart.
The ODT manual has been replaced by a Wiki, published on SF. A snapshot of the wiki is deployed with the software.
The translations (currently Italian only) won't be maintained anymore, unless volunteers come out.
The rts files values are now expressed only with integers and strings in order to make them easier to edit manually. The previous float values have been converted to integers by multiplying them by a power of 10, while QDates are now expressed as strings.
Renewed program icons and logo
Closed the following bugs found in 0.27:
The Error Compensation control in main window didn't work, due to a wrong slot signature.
Under Windows, the first time Echoes was launched after fresh install, when quitting the application (pressing Quit button) a segmentation fault occurred. This happened only once. The following sessions terminated regularly.
Due to a mistake in the device detection code, the NESDR dongle could not work (problem reported by Alec Alden). Now is fixed.
The TEST_*.rts files were completely broken and could work with test_patterns only. The current files instead work with my dongle and are tuned on GRAVES. Be careful: once loaded the file, you'll need to select manually in the device selector YOUR dongle before starting acquisition, because it will most likely be different from mine. Otherwise the test_patterns pseudo-device will be used.
The default main window's vertical size was too small and the content of acquisition mode box under Output tab was not visible (problem reported by Alec Alden). Now the main window is correctly sized to display all the controls and will shrink vertically to fit the desktop size.
Closed the following bugs found in 0.28:
Fixed problem with mismatch between configured dongle and plugged one. The test patterns were started by default, losing the settings loaded from rts file with -s option. Now a message box appears without starting acquisition, so it's possible to select the plugged dongle while maintaining the loaded settings
Fixed tune offset slider preset on waterfall window while loading a rts file.
Fixed console mode acquisition.
Fixed dump files cut above 20s coverage
Closed the following bugs found in 0.29:
Fixed bug: Removed condition on automatic shot triggering: the shot happens when the "Shot after..." timer expires, regardless the event status.
Fixed bug: In client/server configuration, in case of interaction with the server's GUI, the power history diagram of the client goes nuts.
Still in client/server configuration, the FFT resolution increment/decrement are now locked on client.
Fixed bug: the background noise N changes while resizing the waterfall window and the frequency zoom factor.
Fixed bug: the fractional part of Error Correction needed acquisition restart to be applied.
Moved the HDD available space controls from Output to Preferences tab to keep the main window within 600 pixels height.
New feature: the peak detection interval is no more stick at waterfall's center. It can be shifted left/rightwards through a new slider placed above the frequency scale (eccentricity slider).
New feature: keeping clicked a point on the waterfall, a tooltip will show its frequency and timestamp at the moment of clicking.
New feature: the peak detection interval boundaries are now showed on waterfall, above the relative blue arrows. They can be hidden by unchecking the related preference box.
Fixed bug: Changing the bandwidth while acquisition is running causes the immediate frequency ruler update, while the downsampling rate is updated at acquisition restart. This behavior is confusing and must be fixed.
Fixed bug: In manual shots, the instant spectra diagram doesn't appear
Fixed bug: A bug in subsampling code caused the average noise N to increase a lot when subsampling was switched on, with narrower band than sample rate.
Fixed bug: An error in stddev calculation caused the waterfall display loop to run 3..4 times slower than due.
Replaced underlying librtlsdr library with SoapySDR to support more SDR devices.
Increased scrolling speed.
Thresholds can be displayed in the historical graph.
Added a button in main window to open the online manual.
The frequency scale markers (center frequency, notches and peak detection interval boundaries) are now draggable with the mouse.
Added a popup window to show on the waterfall the frequency at mouse position.
Screenshots are now archived keeping the original file date and time.
The rts files have improved their human readability.
Added support for audio spectra (with zero Hz at left) that gets enabled when the SoapyAudio device is selected.
Two controls have been added in main window to calibrate the dBfs gain and offset in order to display power values as dBm.
Added in main window the button Save besides Save as.
The full report can now be required manually without having to stop the acquisition.
Added notification sound for Acquisition fault errors (SDR faulty or disconnected).
Previous page (Acronyms) <------- Back to index ------->Next page (Overview)