sdrangel

Following changes in dsdcc v1.8.7 DMR basic privacy has been implemented in DSD decoder.

Fixed codec setting when using REST API. Fixes #526

• PlutoSDR: fixed handling of open device status. Fixes issue #510
• Prevent use of not existent devices when fiddling with user device parameters. Fixes issue #510
• Main GUI app: moved app creation after setting attributes. Fixes #522
• AudioDeviceManager: make sure a negative or null sample rate is never returned

Code reorganization to isolate the spectrum analysis code (SpectrumVis class) from GUI opening the way to remote spectrum functionality already available on the v5 branch.

Also fixed following bugs:

• FFTFactory: add mutex for thread safety. Implements issue #514
• SpectrumVis: release previous FFT engine allocation from FFT factory only if allocated once. Fixes issue #515

• ATV mod: Ubuntu 20.04 OpenCV code compatibility. Fixes issue #511
• Fixed incorrect frequency display when transverter is engaged. Fixes issue #509

• DATV demod: applied some recent changes in leansdr
• Fixed some warnings from Mac ports compilation

• LimeSDR: revised labels and tooltips for antenna select. Fixes issue #506
• LimeSDR: GUI: flip SR and Hw/Sw decimation or interpolation displays
• XTRX: GUI: flip SR and Hw/Sw decimation or interpolation displays
• XTRX: REST API: fixed wrong setting object
• XTRX: fixed sample rate setting
• XTRX: reduce block size to 4096 bytes

• AM demod: fixed initialization of squelch counter. Fixes issue #502
• Make OpenCV optional in main cmake file. Fixes #503
• Fixed some warnings from Mac ports compilation
• Compile with c++11 comaptible version of CM256cc

Implemented right click on the record button of relevant sample source device GUIs to open a file dialog to choose the file to record I/Q stream to. Implements #483

• BladeRF2: REST API: implemented actions that were missing

• BFM demod: added initialization of message to GUI address. Fixes #488
• REST API: removed useless SWGDevicesetApi files. Fixes #491
• Mac OS: Removed cm256cc dependency and therefore remote plugins. Fixes #492

This applies to Linux only. On Windows the KISS FFT is implemented in place of FFTW thus no difference is expected.

With the FFT factory the FFTW plan structures are preserved during the lifetime of the application. Allocating these plans take time. Depending on the FFT size even on powerful machines this can take several seconds so they are valuable resources not to be wasted. This factory system allows the re-cycling of plans and saves time where a plan of the same size has already been allocated and is not used anymore.

In addition to this the possibility to use a so called "wisdom file". The wisdom file is built offline using the fftwf-wisdom command on the same machine where SDRangel is run. This command runs the lengthy process of testing the hardware to choose the right code to perform FFTs and saves the result to the file thus speeding up the allocation of plans. A process that takes several seconds will take only milliseconds to execute when using this file.

A typical fftwf-wisdom command that can produce an efficient wisdom file in the context of SDRangel is: fftwf-wisdom -v -n -o ~/fftw-wisdom f128 f256 f512 f1024 f2048 f4096 f8192 b128 b256 b512 b1024 b2048 b4096 b8192 To use this file when invoking SDRangel you use its -w command line option: sdrangel -w ~/fftw-wisdom

This is a significant addition to the REST API with the possibility to perform actions on devices and channels. Hence the middle (minor) version is bumped by one at 4.14.0. Now we have 3 families of APIs for devices and channels:

• Settings API (GET, POST, PUT, PATCH) to manipulate devices and channels persistent parameters
• Report API (GET) to fetch information from devices and channels
• Actions API (POST) which is the new one to trigger actions. For now these actions are:
  • Start or stop recording on recoding capable devices
  • Start or pause playback on the FileSource channel plugin

• LimeRFE USB: added cellular band 3 and improved API documentation
• FileSource: fixed various issues with settings with GUI and REST API. Fixes issue #481

Implementation of a Continuous Integration and Continuous Deployment system using Appveyor. This CI/CD now produces the artifacts found in releases. The system is not yet completely operational and artifacts had to be copied manually from Appveyor. For this release these are the artifacts produced in this run: https://ci.appveyor.com/project/f4exb/sdrangel/builds/31157356

For more details on artifacts and their contents please read the quick start in the Wiki. This does not change much for Windows but the Debian package for Linux is now part of a .tar.gz archive containing other elements. Note that the artifact for Linux should have been sdrangel-22-master.tar.gz however at the time dev and master were coinciding with the 4.13.2 tag. Automation is not yet fully functional as mentioned earlier.

• XTRX output: added missing file in cmake file

• FileSource: fixed various issues with settings with GUI and REST API. Fixes issue #481
• LimeRFE USB: fixed attenuator and AM/FM notch settings
• Removed librtlsdr in the list of dependencies in Debian install

• FFT: Added Kaiser window with alpha factor of pi

The LimeRFE or Lime RF Front End is a power amplifier and LNA board designed to augment the capabilities of the LimeSDR in order to build an operational radio solution. The usage is not limited to LimeSDR any Rx or Tx device can be connected to it. The LimeRFE can be controlled directly via its USB port independently of a LimeSDR device. This interface allows exactly that from the SDRangel GUI. Note that this interface is available only in Linux. This implements issue #463

To open the LimeRFE USB dialog open the Preferences sub-menu from the top bar and click on the LimeRFE item. This item is available only when the code is compiled with LimeSuite v20.01.0 or with the binary packages.

More information on LimeRFE is available from Crowd Supply

• BFM demod: fixed crash when adding the demod while the device is running. This fixes issue #469
• Audio devices manager: fixed crash when changing audio device and there is no other channel plugin with the default device selected. In fact this lets the default device run once started until program is quit. This apparently fixes this issue (Qt audio related).

This release does not bring any new functionality nor fixes but is the first release after the split into master and v5 branches. It removes the existing MIMO plugin. There are also some minor changes in the REST API.

This release concerns only Debian builds. The dependency to distribution packaged librtlsdr has been removed in favor of a locally packaged librtlsdr compiled from source (part so called "external libraries") on a definite version that fits with the rest of the code. This should solve problem with Debian Buster build as reported in #455 and will serve in the future if a special version of RTL-SDR library is needed.

Removed the auto discovery at program startup for two reasons:

• It is a pain for anybody not having a Pluto in their network to wait ~5s at each startup for nothing
• It is not working when more than one Pluto is in the network as reported in #445

Instead you will have to specify the Pluto(s) IP address(es) in the user arguments dialog by creating a non discoverable device for the Pluto(s) and associate a uri=ip:x.x.x.x argument to it. The user arguments dialog is accessed from the main window menu via Preferences > Devcies > User arguments.

• MainCore: fixed sink device set creation. This fixes issues when creating a new sink device set in the server variant,

The classes and their organization were reworked extensively to match the already existing (since v4.12.0) Tx side. This qill hopefully make the code easier to understand. While this concerns mostly developers it was also the occasion to make further refactoring in the following plugins that may have a significant functional impact hopefully fot the best:

• ChannelAnalyzer
• ATVDemodulator
• RemoteSink and accordingly the RemoteOutput device plugin

• Tx plugins: further refactoring and fixes
• Scope: added DOA projections and updated ChannelAnalzyer
• Upgraded codec2 library to v0.9.2
• NFM demod: fixed audio FIFO fill
• Various changes to base classes for MIMO

This is an extensive refactoring of the Tx part i.e. plugins/samplesink, plugins/channeltx and related common classes around two main points:

• redesign of the sample source FIFO (SampleSourceFifo)
• reorganization of the code with better separation of admin, baseband generation and source generation roles split in 3 different classes. See developer's notes for details: https://github.com/f4exb/sdrangel/wiki/Developers-notes. This implements #438 and partly #440

The main consequence for the average user comes from the new sample FIFO that overcomes some issues with the previous design. Thus if you are using SDRangel for transmit it is strongly advised to upgrade to this version.

This large refactoring implies a jump of minor release to 4.12.0

• Corrected network manager process memory leaks. The reply object and request buffer were not disposed properly upon reply. This memory leak was problematic if the REST API was used on the same SDRangel instance over a long period.
• Fixes for Mac OS build in leansdr code (contributed)

• New device sink plugin for Tx testing: https://github.com/f4exb/sdrangel/tree/master/plugins/samplesink/testsink
• Channel Analyzer: set trace length to FFT length when auto-correlation by IFFT is used

• Tx plugins: corrected FIFO begin iterator vs number of samples read
• GLSpectrum: set default window to Hanning
• More preparatory works for MIMO channels support