Version 4.2.0

• Add support for building with libshout v2.4.6, see #382 and #422
• Add error checking for lowpass <= highpass, see #399 and #412
• Remove limit on count of mixer inputs (thanks @cdknox), see #408
• Add dated_subdirectories config option for output files (thanks, @marcin-osowski), see #413

Version 4.1.1

Changes in this release:

• Fix build issues when using VideoCore GPU, see #378

Version 4.1.0

• Changes in this release:
  • Add channel_squelch_level to stats file, see #332
  • Support "default" values in lists for squelch_snr_threshold and notch_q,
    see #334
  • Set cmake ENABLE_EXPORTS property, see #339

What's Changed

• add signal and noise as dBFS to output prom file by @charlie-foxtrot in #355
• update unstable with main (4.0.3) by @charlie-foxtrot in #357
• Fix Mac CI and update URLs by @charlie-foxtrot in #358
• merge MacOS CI fixes to unstable by @charlie-foxtrot in #359
• Reschedule daily build and add vscode devcontainer by @charlie-foxtrot in #364
• add gtest and unit test stub by @charlie-foxtrot in #365
• Pull appart classes by @charlie-foxtrot in #367
• add ampfactor as a per-channel config parameter by @charlie-foxtrot in #369
• run more build combinations in CI by @charlie-foxtrot in #372
• drop error messages when setting to defaults by @charlie-foxtrot in #371
• Add CTCSS support, second attempt by @charlie-foxtrot in #368
• updates to release v4.1.0 by @charlie-foxtrot in #373

Full Changelog: v4.0.3...v4.1.0

Version 4.0.3

Changes in this release:

• Add channel_squelch_level to stats file, see #332
• Support "default" values in lists for squelch_snr_threshold and notch_q, see #334
• Set cmake ENABLE_EXPORTS property, see #339

Other items to note:

• Repo maintainer has changed, see #342
• Repo URL has moved to https://github.com/charlie-foxtrot/RTLSDR-Airband
• Default branch / Top of Tree has been renamed to main

Version 4.0.2

• Added a new PLATFORM value default (which, as the name says, is the new
  default). It results in a portable binary without any architecture-specific
  optimizations. This also allows the program to be built with compilers that
  do not support -march=native option (notably Clang on Apple M1) (#303).

Version 4.0.1

• Fixed compilation error on RaspberryPi OS 11 (Bullseye)

Version 4.0.0

• RTLSDR-Airband is now built with CMake. Refer to the wiki for updated
  compilation instructions.
• When compiling the program, a new PLATFORM value native can now be
  specified. It enables -march=native -mtune=native compilation options. This
  causes the compiler to apply the most appropriate optimizations for the
  hardware on which the app is being built (thx @charlie-foxtrot).
• BACKWARDS-INCOMPATIBLE CHANGE: Signal level and noise level estimates
  displayed in the textual waterfalls are now expressed in dBFS (decibels
  related to the full scale of the analog-to-digital converter). The main
  benefit of the new approach is that these values do not depend on the
  fft_size value(thx @charlie-foxtrot).
• BACKWARDS-INCOMPATIBLE CHANGE: Improved squelch algorithm with new
  configuration parameters. squelch keyword has been replaced with
  squelch_threshold which takes an absolute signal value in dBFS as an
  argument. Alternatively, a minimum signal-to-noise ratio (in dB) that should
  trigger the squelch might be configured using squelch_snr_threshold option
  (thx @charlie-foxtrot).
• BACKWARDS-INCOMPATIBLE CHANGE: include_freq config option for file outputs
  now causes the frequency to be appended after the timestamp rather than
  before it. This feature now works correctly in scan mode, when
  split_on_transmission feature is enabled. (thx @charlie-foxtrot).
• BACKWARDS-INCOMPATIBLE CHANGE: sample format in files produced by rawfile
  outputs has been changed from CS16 to CF32. File name suffix is now .cf32.
• Improved squelch indicator in the textual waterfalls. In addition to the *
  character indicating that the squelch is open, there is also a ~ character
  indicating that the channel has a signal that is being suppressed because it
  is outside the band of the channel filter (thx @charlie-foxtrot).
• New output type udp_stream for sending uncompressed audio to another host
  via UDP/IP (thx @charlie-foxtrot).
• Added multiple_output_threads global option. When set to true, a separate
  output thread is spawned for each device (thx @charlie-foxtrot).
• Modulation in scan mode is now configurable per channel (thx
  @charlie-foxtrot).
• SoapySDR errors like TIMEOUT or OVERFLOW are no longer treated as fatal. They
  often appear intermittently, especially when the CPU usage is high. There is
  no point in failing the input in this case.
• Added .tmp suffix to the names of the output files currently being written
  to. The suffix is removed when the file is closed. External applications that
  consume recorded files can now figure out which files are not yet complete.
• Added logging and statistics for output thread overruns and mixer
  input/output overruns (thx @charlie-foxtrot).
• The program can now be built on MacOS.
• Miscellaneous bug fixes and code cleanups.

Version 3.2.1

• Fixed a compile error when using libshout older than 2.4.0

Version 3.2.0

• Added split_per_transmission output file option which allows creating
  a new file for every transmission on the channel (thx @charlie-foxtrot).
• Added include_freq output file option, which causes the channel frequency
  to be appended to the file name (thx @charlie-foxtrot).
• Added support for notch filters for eliminating narrowband interference,
  like CTCSS tones (thx @charlie-foxtrot).
• Added bandwidth channel option which causes the channelized I/Q signal
  to be lowpass-filtered before demodulation. This might help in situations
  where neighboring channels are closely spaced and interfere with the channel
  of interest. It also reduces the bandwidth of the resulting audio signal,
  and thus eliminates the high-frequency noise (thx @charlie-foxtrot).
• Added support for multithreaded demodulation. Each device can now have its
  own demodulation thread. This allows spreading the demodulation work across
  multiple CPU cores. Enable with multiple_demod_threads global option
  (thx @charlie-foxtrot).
• Added support for highpass/lowpass MP3 filters for mixers (thx @charlie-foxtrot)
• Added support for frequency usage statistics (thx @charlie-foxtrot).
• Workaround for Fitipower tuner problem of not honoring the first gain
  setting when the device is first used (thx @eshaz).
• Finalize the MP3 file properly before opening a new one (thx @jratke).
• Close the RTL device properly on program exit (thx @jratke).
• Updated the SoapySDR input driver to reflect changes in SoapySDR library API.
• Minor cleanups.

Version 3.1.0

• SoapySDR: added support for complex float 32-bit samples
• SoapySDR: allow using AGC if the device supports it. Gain setting for
  soapy devices is now optional - if it's not specified, the program will
  try to enable AGC.
• Use lowpass/highpass filters provided by LAME library to improve audio
  quality of MP3 streams. Filter cutoff frequencies may be configured per
  output, using highpass and lowpass config options. Credit: clydebarrow.
• Added log_scan_activity global config option. When set to true, a
  log message is written whenever a squelch opens on a scanned channel,
  effectively producing a channel activity log. Credit: clam-i-am.
• Improved squelch behaviour in some corner cases.
• Fix for incorrect naming of pulseaudio context. Name set in the config
  was not used as it should. Credit: Darryl Pogue.
• Don't fail when the configured gain value is negative. Some SDRs support
  this (eg. FC0012-based dongles).
• Fix a bug which in some cases could prevent the icecast output from
  reconnecting with the Icecast server after the connection has failed.