cava

bugfixes:

• fix source being ignored when using pipewire by @gardk in https://github.com/karlstav/cava/pull/557
• fix stdlib.h not being included after iniparser upgrade #560

improvements:

• add waveform visualization #484
• exit with error if pipewire loop creation fails #561

Full Changelog: https://github.com/karlstav/cava/compare/0.10.1...0.10.2

What's Changed

• JACK support by @bsdcode
• default to pulseaudio before sndio

Full Changelog: https://github.com/karlstav/cava/compare/0.10.0...0.10.1

• OSS input backend and general improved FreeBSD support
• pipewire default if supported
• noncurses is now default

bugfixes:

• hardcoded pipewire include flags #519
• default input wrongly set to alsa instead of pulseaudio/pipewire #518

new:

• glsl shader output
• windows operating system support
• gradient support in sdl output

breaking:

• some config options are now 0-100 instead of 0-1 due to windows ini not supporting floats

bugfixes:

• fix some calculations on x-axis label
• cavacore: channels where sorted wrong according to doc
• fix autosensing sometimes not working

improvements:

• improved response time by decreasing audio buffer size
• add support for alacritty synchronized updates
• brought EQ back by popular demand
• added support for hex colors and gradient in noncurses mode
• add configurable orientation of bars

bugfixes:

• fix buffer overflow in audio buffer when calling reset_buffer (used in fifo input)
• cavacore: fix segfault when setting 0Hz lower cut off
• framerate <= 1 msec wait was multiplied by 1000 instead of divided

improvements:

• cavacore: don't print to stdout or call exit, use error status flag in cava_plan instead
• cavacore: add extern C construct for c++ compilers
• add license in cavacore header file
• remove build dependency on xxd, use incbin instead

bugfixes:

• fix critical bug where default config value caused always maxed out bars #447

breaking changes:

• split out core functionality in separate library cavacore

this is built as a separate object file with it's own header file and can be used in other applications
see CAVACORE.md for details

• removed ignore config option
• merge gravity and integral config option to "noise reduction"

improvements:

• use double precision all the way, improves details in visualization.
• increase buffer sizes for large sample rates 96k++
• noise filters should now scale better with resolution and framerate
• embed default config file and populate on first launch instead of installing it
• Add new output format, noritake VFD bitmap

bugfixes:

• fixed some potential buffer overflows
• fix some bugs in sdl output

improvements:

• added sdl output support
• better smoothing on higher resolutions, better scaling

bugfixes:

• fixed buffer overflow in shmem input
• fixed bug where freebsd was not detected as supported os during autoconf

other:

• removed bundled iniparser

improvements:

• added faster autosens at startup

bugfixes:

• better installation of custom console font
• suppress error about vla parameter introduced in gcc11 or 10?

improvements:

• Added Hann windowing and proper ring buffer (this improves visualisation significantly!)
• added mutex lock on write to fftw input buffers and fftw execute (this removes som glitches)
• improved portaudio audio input
• #370 made timeout configurable
• ncurses mode is now default again
• #379 removed key controls from noncurses to fix strange beaviour
• added config option: x-axis scale (frequency) #289

bugfixes

• #375 shmem audio input
• #377 small memory leak
• #385 fix eq

improvements:

• added key controls for noncurses mode #364

bugfixes:

• fix drawing of background color for noncurses mode
• fix install on macos #365

improvements:

• set default gradient colors instead of exiting when not defined.

bugfixes:

• minor memory leak in ncurses
• fix warnings when building on macos
• updated macos building instructions

A lot of things went down since the last release. I sort of left out creating new releases until I realized things like the AUR package where using the old 0.6.1 release. Will try to create releases regularly from now if there are changes.

Braking changes

• Reworked noncurses output mode and made it defualt.

  The reworked noncurses mode uses a custom framebuffer technique that
  only draws changes per frame and only writes to the terminal once per
  frame. This results in far less cpu usage then ncurses. It is also
  free of the flickering that we see in ncurses mode #196

  Colors are no longer set by default so hopefully cava will change
  color automatically in noncurses mode when using tools such as
  pywall #340 #181

  Unfortunately html colors, gradient and key controls are not supported
  in noncurses.

Other improvements:

• separated bass, midtones and treble in to separate fftw operations

for a complete list of changes see:

https://github.com/karlstav/cava/compare/0.6.1...0.7.0

new features:

• sndio backend

bug fixes:

• Fix fftw_complex usage, drastically improving visualization
• better cleanup and exiting
• added a check to see if pulseaudio is running
• fixed bug preventing color in noncurses mode from working
• Fix build errors and warnings if not using ncurses

Generally improved fifo input and raw output

• Autosens of low values (dynamic range)
• Removed config parameter 'overshoot'
• Removed config parameter 'mode'
• New config parameter 'waves'
• Changed config parameter 'style' to 'channels' (was either 'mono' or 'stereo' anyway)
• Parameters 'integral' and 'gravity' are now in percentage

new features:

• gradient

bugfixes:

• memory leak
• increased config reload speed
• raw output not goint to /dev/stdout

new features:

• support for more colors in ncurses output.

bugfixes:

• fixed bug where sensing of low values was done even when no sound
• fixed bug in raw output bit format check
• adde a check for too many bars >200

new features:

• raw-output mode
• autosens for low values at start-up
• Reload config file on SIGUSR1
• config option for building with iniparser version 3.1