saturation.
Each tunerchannelsource was sharing a thread pool and as the processing
load increased, the thread pool was not expanding and the processing
tasks were backing up, causing the buffers to grow with unprocessed
data.
Changed so that each tuner channel source is handed a thread pool to use
for processing incoming sample data. Processing chain maintains the
thread pool and starts the channel source. This allows traffic channels
to reuse the thread pool across multiple iterations.
Resolves issue in tuner controller where controller was not releasing
sourced tuner channels, preventing re-tuning to other frequencies.
Resolves issue with channel timeout on MPT1327 monitored traffic
channels
calculation when additional channels are allocated.
Incorporates tuner specific definition of overall usable bandwidth, so
that tuners like airspy can specify 90% of 10 MHz (ie 9 MHz) is truly
usable and the fringe 10% is unusable rolloff.
Also includes tuner-specific defining of the width of the central spike,
or a value of 0 indicates that the entire range is usable (ie airspy).
when tuner channels are allocated by the airspy. New method determines
the min/max requested frequency range and sets tuner center to the
center of that range, while limiting the total usable bandwidth to 90%
of the bandwidth, to account for rolloff of 5% at the upper/lower edges.
When a new channel is requested, if the channel fits within the current
center frequency and usable bandwidth, the center tune frequency is not
adjusted, to minimize the number of retunes on a busy system.
Updated p25 decoder state to better track calls and call detects from
a trunked channel
Fixed issue with false detects on LDU2 encrypted call messages
Fixed Airspy channel allocation ... will allocate channels across the
entire bandwidth, minus 50 kHz on either end, and will not adjust the
center frequency if it's not needed.
Fixed traffic channel teardown -- traffic channels are immediately
removed now, without a fade delay.
AudioMetadata - updated flag is now reset upon copying the metadata
Note: Several of the decoders are currently non-functioning in this
commit. If you want the latest functioning codebase, please checkout
the commit prior to this one. All decoders will eventually be updated
with this new architecture.
Currently broken:
- Logging and recording
- Aux decoders (Fleetsync/MDC/LoJack/Tait)
- LTR/Passport decoders
Changes in this commit:
- Module architecture for decoders and components with central
provisioning by the DecoderFactory
- Transition from Complex sample stream processing to float buffer
processing for improved memory management
- Redesigned filter structures - to support buffers and streams
- Resolved several memory leak issues
- Universal sharing of the thread pool manager
- Added centralized audio management with call/user prioritization and
new gui with mute/volume controls
- Embedded audio metadata for eventual streaming support
- Redesigned ProcessingChain to support independent re-use
- Added new TrafficChannelManager to handle traffic channel allocations
- Traffic Channel pool size and Call timeout available in P25 and
MPT1327 configs
- Redesigned channel and decoder gui components
Changed C4FM symbol filter to reset frequency tracking values when the tuner's frequency or PPM values change.
Moved all frequency control pieces into the source.tuner.frequency package.
Converted to use float primitives and to keep samples in buffer format until the sample rate is decimated to 48kHz.
Added new non-recursive, CIC decimation filter -- untested
Updated sound card/mixer input source management. Now testing for mono/stereo support for each card and limiting channel choices accordingly. Separated the mixersource into a ComplexMixer and and SimplexMixerSource. The ComplexMixer is used for sound card based tuners like the funcube. The SimplexMixerSource is used for demodulated sound card inputs.
Interfaces for sample handling - standardized on using 'Complex' to indicate I/Q samples and 'Simplex' to indicate demodulated float samples. The processing chain now accepts either a decimated (48k) list of complex samples from a tuner channel source, or a 48k simplex buffer (array of demodulated float samples).
Resolved issue with HackRF tuner not applying tuner configuration information.
