sounddevice

https://pypi.org/project/sounddevice/0.5.3/

• add explicit_sample_format parameter to WasapiSettings

https://pypi.org/project/sounddevice/0.5.2/

• Better error if frames/channels are non-integers

https://pypi.org/project/sounddevice/0.5.1/

• Windows wheel: bundle both non-ASIO and ASIO DLLs, the latter can be chosen by defining the SD_ENABLE_ASIO environment variable

https://pypi.org/project/sounddevice/0.5.0/

• Remove ASIO support from bundled DLLs (DLLs with ASIO can be manually selected)

https://pypi.org/project/sounddevice/0.4.7/

• support paWinWasapiAutoConvert with auto_convert flag in WasapiSettings
• Avoid exception in PortAudioError.__str__()

https://pypi.org/project/sounddevice/0.4.6/

• Redirect stderr with os.dup2() instead of CFFI calls

https://pypi.org/project/sounddevice/0.4.5/

• Add index field to device dict
• Require Python >= 3.7
• Add PaWasapi_IsLoopback() to cdef (high-level interface not yet available)

https://pypi.org/project/sounddevice/0.4.4/

• Exact device string matches can now include the host API name

https://pypi.org/project/sounddevice/0.4.3/

• Fix dimension check in Stream.write()
• Provide "universal" (x86_64 and arm64) .dylib for macOS

https://pypi.org/project/sounddevice/0.4.2/

• Update PortAudio binaries to version 19.7.0
• Wheel names are now shorter

https://pypi.org/project/sounddevice/0.4.1/

• CallbackFlags attributes are now writable

https://pypi.org/project/sounddevice/0.4.0/

• Drop support for Python 2.x
• Fix memory issues in play(), rec() and playrec()
• Example application play_stream.py

https://pypi.org/project/sounddevice/0.3.15/

• This will be the last release supporting Python 2.x!

https://pypi.org/project/sounddevice/0.3.14/

• Examples play_sine.py and rec_gui.py
• Redirect stderr only during initialization

https://pypi.org/project/sounddevice/0.3.13/

• Examples asyncio_coroutines.py and asyncio_generators.py

https://pypi.org/project/sounddevice/0.3.12/

• Support for the dylib from Anaconda

https://pypi.org/project/sounddevice/0.3.11

• Support for the DLL from conda-forge

https://pypi.python.org/pypi/sounddevice/0.3.10

• Change the way how the PortAudio library is located

https://pypi.python.org/pypi/sounddevice/0.3.9

• Add sounddevice.Stream.closed
• Switch CFFI usage to "out-of-line ABI" mode

https://pypi.python.org/pypi/sounddevice/0.3.8

• Add more ignore_errors arguments
• Add sounddevice.PortAudioError.args
• Add sounddevice.CoreAudioSettings

https://pypi.python.org/pypi/sounddevice/0.3.7

• Add sounddevice.get_stream()
• Support for CData function pointers as callbacks

https://pypi.python.org/pypi/sounddevice/0.3.6

• Example application play_long_file.py

https://pypi.python.org/pypi/sounddevice/0.3.5

• Add extra_settings option for host-API-specific stream settings
• Add sounddevice.AsioSettings and sounddevice.WasapiSettings

https://pypi.python.org/pypi/sounddevice/0.3.4

• Example application rec_unlimited.py

https://pypi.python.org/pypi/sounddevice/0.3.3

• Add loop argument to sounddevice.play()

https://pypi.python.org/pypi/sounddevice/0.3.2

• mapping=[1] works now on all host APIs
• Example application plot_input.py showing the live microphone signal(s)
• Device substrings are now allowed in sounddevice.query_devices()

https://pypi.python.org/pypi/sounddevice/0.3.1

• Add sounddevice.check_input_settings() and sounddevice.check_output_settings()
• Send PortAudio output to /dev/null (on Linux and OSX)

https://pypi.python.org/pypi/sounddevice/0.3.0

• Remove sounddevice.print_devices(), sounddevice.query_devices() can be used instead, since it now returns a sounddevice.DeviceList object.

https://pypi.python.org/pypi/sounddevice/0.2.2

• Devices can now be selected by substrings of device name and host API name

https://pypi.python.org/pypi/sounddevice/0.2.1

• Example applications wire.py (based on PortAudio's patest_wire.c) and spectrogram.py (based on code by Mauris Van Hauwe)

https://pypi.python.org/pypi/sounddevice/0.2.0

• support for wheels including a dylib for Mac OS X and DLLs for Windows. The code for creating the wheels is largely taken from PySoundFile.
• remove logging (this seemed too intrusive)
• return callback status from sounddevice.wait() and add the new function sounddevice.get_status()
• sounddevice.playrec(): Rename the arguments input_channels and input_dtype to channels and dtype, respectively

https://pypi.python.org/pypi/sounddevice/0.1.0

Initial release.

Some ideas are taken from PySoundCard. Thanks to Bastian Bechtold for many fruitful discussions during the development of several features which python-sounddevice inherited from there.