karlstav / cava Goto Github PK

pi@raspberrypi:~/Desktop/cava $ cava
cava: error while loading shared libraries: libiniparser.so.4: cannot open shared object file: No such file or directory

It's not running under Raspberry Pi 2B. I have installed Raspbian GNU/Linux 8.0 (jessie).
There was no errors when I was compiling, I compiled it two times for being sure.

I'm looking into the feasibility of using CAVA output to drive a visualizer for Volumio (https://volumio.org). I was envisioning displaying the spectrum analyzer output on the client side, in their browser.

Any chance CAVA can output the FFT results to fifo? We can then push it through a websocket, and let the client draw the spectrum.

Thanks!

Although it says:

This program is not intended for scientific use.

But I begin to wonder the output. It all started with #9.

I don't have any tools that can generate test audio, so I generated on my own (see below), an 1kHz, s16le, 2ch, raw PCM audio, 1000hz-s16le-2c-5sec.raw, using wave:

% echo $((44100*5)) 1 1000 | wave -r 44100 > 1000hz-f32le-1c-5sec.raw
% # aplay -f FLOAT_LE -r 44100 -c 1 -q 1000hz-f32le-1c-5sec.raw
% ffmpeg -f f32le -acodec pcm_f32le -ac 1 -i 1000hz-f32le-1c-5sec.raw -f s16le -acodec pcm_s16le -filter_complex "[0:a][0:a]amerge=inputs=2[aout]" -map "[aout]" -ac 2 1000hz-s16le-2c-5sec.raw

I commented out the parts in db80571 that I think they would be affecting: L568-570 and L595-606. I may miss some, but from the output of ALSA input, I get rid of them enough.

FIFO

Here is the output of ./cava -i fifo -p 1000hz-s16le-2c-5sec.raw:

ALSA

And output of ./cava -i alsa -d pulse with playback by aplay -f S16_LE -r 44100 -c 2 -q 1000hz-s16le-2c-5sec.raw in background:

Questions

Both peaks did indicate the 1kHz, when I changed the frequency, they shifted, so I believe that does show, but there should not have anything but a peak in FIFO.

• Is FIFO processing correctly?

  Both outputs of inputs should look more or less the same, right? Did I feed in right format of data?

  Below 1k, the leftmost three peaks, there is something interfering, they look related, but I can't tell what it is.

  (Forgot almost everything I learned in signal processing. Sorry, professors!)

• What are the frequency bands, exactly in the output we see?

  The locations of the peaks don't look right at all

  If the output is equally dividing 50-12kHz, 1kHz should be about one 12th of window width from the left. But both are close to the center.

  From what I can see, in -i alsa, the rightmost frequency/bar in output is somewhere around 9kHz.

  Is it log scale? 1k almost in middle, 10k almost at right, that sounds about right, doesn't it?

started working on module based source code using file structure suggested by @CelestialWalrus in #34 . See separate branch "modules".

Only been able to push ncurses drawing out in separate files for now. Made alsa and fifo files, but haven't gotten around to move the code yet.

It looks like it's plausible, but I have never done anything similar before. So any hints/tips much appreciated, thanks!

Well, I was thinking about porting cava to Windows & OS X, and I think a switch to PortAudio would be great, because it solves some of the issues.

Also: I think you can use PortAudio on Android, so that'd be nice.

PortAudio is a free, cross-platform, open-source, audio I/O library. It lets you write simple audio programs in 'C' or C++ that will compile and run on many platforms including Windows, Macintosh OS X, and Unix (OSS/ALSA).

The error "not permitted"

ALSA lib /var/tmp/portage/media-libs/alsa-lib-1.0.29/work/alsa-lib-1.0.29/src/pcm/pcm_hw.c:1693:(_snd_pcm_hw_open) Invalid value for card
error opening stream: Operation not permitted

is a bit misleading, since adding sudo will not fix the problem. It would be cool if cava could check for the existence of /sys/module/ (sysfs mounted) and then for /sys/module/snd_aloop/ (module loaded). That should allow producing a more helpful error message. What do you think?

I'd suggest to Show time, date or even both (maybe selectable over arguments) in the Top-Right Corner.

Issue was first raised in #36 by @CelestialWalrus, see illustration here.

Might be the same as what @rolf1304 describers as " flickering cursors at the bars' tops" in #36 .

I'm not seeing the same issue in Terminator on my computer, suggesting that it might be related to something else then the terminal software. @CelestialWalrus can you provide info about font and terminal settings? @rolf1304 can you look at the gif and confirm if it is the same issue you are seeing?

Building fails on Gentoo Linux. ./autogen.sh and ./configure ran fine though.

https://bpaste.net/show/08086febace8

Maybe you want to add the build requirements for fedora to the readme?

For fedora you need:

• alsa-lib-devel
• ncurses-devel
• fftw3-devel

I also wrote a simple specs file for automated rpm building: https://gist.github.com/sedrubal/617d40deb7452a25d004#file-cava-spec

And thanks for this program ;)

Until now the only way to change the available colors is to change the default terminal colors, is it possible to support more colors out of the box in the future?

Not necessarily an issue with the program... but I'm having trouble installing it on Ubuntu 14.04.3 LTS. I really like everything I've seen about the program. With regards to installation, my issues lie when I get to the sections: "Getting Started" and "Installing Manually". I believe I understand the sections after that... I just have no idea how to install it (And somehow allow the program to use FFTW?). Like I said, I'm very interested in using CAVA, I just need some guidance. Thanks!

I am not capable of using your tool. Everytime I run it nothing happens. I can use "cava -v" and if I press CTRL-C it will stops but it shows no bars at all. By the way I'm running it on a Raspberry Pi.

I'd like to see the top of bars with ~80% height, e.g. in red (like many other visualizers do, like this

is that possible?

https://aur.archlinux.org/packages/cava/

If it gets 10 votes it'll be (probably) moved to community repo.

Using Funtoo. I'm happy to list any relevant versions of different libs.

It seems like cava is not repainting some parts of the window after window resizes, properly. This is is what I see:

Looking at other visualizer like projectM, I think pulseaudio is able to grab audio directly from sink. This would eliminate the need for the complicated 'alsa loopback'+'pulseaudio combined sink' setup.

Now I know a lot of people don't like pulseaudio. I use it, mostly because it came with my disrto and it works. The alsa part will of course be kept as an option.

I think this will also require a configure script that checks for the presence of alsa or pulseaudio and creates the makefile accordingly.

So is it worth it making cava able to use pulseaudio directly? Do people use pulseaudio?

We should tag stable versions (i.e. for the purposes of packaging).

I've made an AUR package for git version: https://aur.archlinux.org/packages/cava-git and it'd be good to make a stable package.

Using urxvt with the smallest width allowed, the program crashes and gives off this error:
floating point exception (core dump).

http://gfycat.com/WelltodoInfatuatedGypsymoth -- note that the music is paused.

I use mpd's fifo: cava -i fifo -p /tmp/mpd.fifo -b 60

I noticed a few things and I'd like to contribute, however, there is no license and I don't like to touch code without written open source license.

@karlstav would you license C.A.V.A? If yes, @anko and @CelestialWalrus do you agree on the license?

Here is a list I might do:

• no consistent coding style, and issues like \t and space mixed in indentations and trailing whitespace, etc.

  This one has to be done first, or more inconsistence would be made.

• makefile, use canned recipe and turn on -Wall and -Wextra, etc.

• warnings like the following and things popping out after turning on warnings:

  cava.c:586:40: warning: universal character names are only valid in C++ and C99 [enabled by default]
                                   printf("\u2588");
  

  It's just warning, although cava works fine, but I think this should be dealt with properly.

  I think this need @karlstav to decide the direction, either -std=c99 or print out UTF8-encoded? I am not sure since I have never dealt with the Unicode, not really a C coder. (And #4 related or helped?)

  If taking C99 route, usleep might need to be replaced with nanosleep or there would be another warning.

Continuation of #34.

Current progress:

• Compiles under Cygwin
• Doesn't compile under MSYS/MinGW (pthread missing)
• ALSA will probably never work
• FIFO will have to be replaced with named pipes on Windows
• PulseAudio input works, but I couldn't get PulseAudio to get my Stereo Mix

Tested under Windows 10.

# make -j5
[..]
gcc: error: ./iniparser/libiniparser.a: No such file or directory
[..]

The issue is missing dependencies. Target cava actually depends on presence of iniparser/libiniparser.a but the dependency is not coded into the Makefile.

Cava doesn't currently set the window title so titlebars, etc show the terminal emulator's default title.

Here's a screen, notice how the terminal running cava still says "urxvt" while the other terminal running ncmpcpp has it's own title.

The title doesn't have to be complex, just "cava" would be fine!

Thanks!

Just another thought I had. I frequently listen to music that varies in volume (I can't ReplayGain online streaming services), and I have autosens turned on so that the high peaks get toned down in cava. However, this means for quieter songs, cava is just a bubbling brook at the bottom of the terminal window.

It would be cool to set a lower threshold (or just default to something reasonable, like 80%) for autosens to adjust the sensitivity UP to for the peaks to fall into, to keep it entertaining. I'd imagine the timeout for lows could/should be longer than for highs, but again I'm probably asking for something much more complicated than I realize. :-)

Now that we've got all sorts of pull requests going on, it might be good to at least think about unit-testing. We've had regressions and they're always hard to catch manually, but could have been prevented with automated tests that check for known past issues and correctness.

I don't know exactly how the tests would look, but @livibetter's manual test of piping generated sound into cava's inputs (#15) might make a good starting point.

In its present state, cava is tough to write tests for, as everything is in a single file that does many things. Separating the model ("how high should bars be") from its presentation ("how do we print this") would be a good first step.

Further thoughts, ideas?

so, i installed everything, did everything right, but when i try to run it it says segmentation fault

ncurses6 was added to the testing repos on arch and cava fails to build because it can't find ncurses5.

For ncurses6 users, I've found two temporary solutions:

• Symlink the ncurses6 binary

  sudo ln -s /bin/ncursesw6-config /bin/ncursesw5-config
  

• Or Change "ncursesw5-config" on line 18 of the makefile to "ncursesw6-config"

Rebuild after either symlinking or editing the makefile and it should build and run fine.

[forget those, forgot to update cava; coming soon]

This is the error:

root@moode:~/cava# make debug=1
cd iniparser && make
make[1]: Entering directory '/root/cava/iniparser'
CC      src/iniparser.o
CC      src/dictionary.o
AR      libiniparser.a
a - src/iniparser.o
a - src/dictionary.o
RANLIB  libiniparser.a
LINK    libiniparser.so
make[1]: Leaving directory '/root/cava/iniparser'
gcc -std=c99 -Wall -Wextra -DPACKAGE=\"cava\" -DVERSION=\"0.3.5-8-g74e74e4\" -D_POSIX_SOURCE -D _POSIX_C_SOURCE=200809L -Iiniparser/src -DDEBUG   cava.c iniparser/libiniparser.a  iniparser/libiniparser.a -lasound -lm -lfftw3 -lpthread -I/usr/include/ncursesw -lncursesw -ltinfo -o cava
cava.c: In function ‘main’:
cava.c:804:28: error: ‘peak’ undeclared (first use in this function)
        lcf[o], hcf[o], peak[o], f[o]);
                ^
cava.c:804:28: note: each undeclared identifier is reported only once for each function it appears in
cava.c:380:6: warning: variable ‘flastd’ set but not used [-Wunused-but-set-variable]
  int flastd[200];
      ^
<builtin>: recipe for target 'cava' failed
make: *** [cava] Error 1

Hi!

Some Linux distros do have packages for iniparser, e.g. Gentoo does. It would be cool to be able to build cava against the iniparser installed system-wide rather than the bundled one. For a list of reasons why that would be helpful, let me point to this page:
https://wiki.gentoo.org/wiki/Why_not_bundle_dependencies

In case a related patch does not seem worth your own time at the moment, please let me know if you would accept a patch in that direction, in general.

Best, Sebastian

I can't figure out how to do this and would like to make it 220-370 instead of 20-12000 Hz.

EDIT: Massive PEBCAK here, it's been a long day. Turns out reading the top comment and realizing that ; is used as a comment character as well as # would help...

It seems the config file parsing is causing a crash, at least for me. Whenever I launch without parameters (ie, it's reading the config), or whenever I refresh the config inside of cava, it freezes.

For the former (launching w/out parameters), it just sits with the bars all at their lowest value, for the latter (refreshing config while cava is open and working), it blanks the screen and stops working.

I just re-tried, and it no longer freezes after refreshing in cava, it just fails to load the config.

I'm on build cava 0.3.4-10-ge02e383-dirty on Linux 4.2.0-gentoo-r1 x86_64.

I tried identifying the license in LICENSE. It's a permissive license, but seems to match none of ISC, BSD, zlib, .. exactly. Did you make a custom license or did I just miss the one? If it's non-custom, please mention the name somewhere. If it's custom, please consider moving to something non-custom.

Thanks and best, Sebastian

gettting an error
error opening stream: Operation not permitted
cava: pcm_params.c:2288: snd_pcm_hw_refine: Assertion `pcm && params' failed.
Aborted (core dumped)
upon trying to start the program.

Pre-e2a1265:

Post-e2a1265:

The bass is now recessed.

It happens with at least 10 tracks on my playlist. (probably more, only checked with ten)

It seems less responsive for me.

At the moment, cava renders each frame by printing term_width * term_height characters to fill the terminal.

This could be sped up using ncurses to only erase and re-print what changed.

Example:

Going from bars like

#   #
##  #
### #
#####

to

#    
##  #
#####
#####

would just mean doing this (d for delete, i for insert)

    d

   i 

with two cursor jumps, instead of re-printing the whole lot.

Here's a screenshot of what I am seeing. If I run cava in the tty, the bars are displayed correctly, so this makes me think that it is a font issue, but I have tried different fonts with nothing changing.

@dylanaraps @CelestialWalrus or who ever might be reading this, I started working on the config script in a separate branch. It's not really my area of expertise, yet... and I'm kinda stuck. I guess I will figure it out eventually, but any help would be much appreciated. Thanks!

When trying to compile cava without having a path to the config-file, cp fails with the following message:

cp: cannot create regular file ‘/home/phijor/.config/cava/config’: No such file or directory

Steps to reproduce:

1. rm -r ~/.config/cava (that's where the config should up for me, because $XDG_CONFIG_HOME isn't set on my machine).
2. make.
3. See it fail.

I will try to fix the issue, shouldn't be to difficult. I think it's the best if $CONFIGDIR actually contains the directory the config will reside in, not the absolute path to the file itself. That way one can check if it actually exists before trying to put something in it.

I'm loving cava so far, especially now that 0.4.0 has added direct PulseAudio support!

One thing I thought might be cool is if cava can reload the configuration when sent a signal, for example SIGUSR1, like httpd and various other programs which use USR1 as a command to reload the config.

The prime example of this is if I change the pulseaudio output/source, I want to automatically set the appropriate variable in cava's config and tell cava to re-read it, via a script. This would also apply for any other config changes.

Perhaps better, cava could watch the config file for changes and automatically reload it, but I think that may be a bit more involved than necessary.

I'm not sure if this is as simple as adding an if statement to sig_handler, or not, as I've rather little C experience.

Having just built cava and rebooted (I needed to anyway for a kernel upgrade), so the pulseaudio server would restart, I now try to run it and receive the following error:

ALSA lib pcm_hw.c:1693:(_snd_pcm_hw_open) Invalid value for card
error opening stream: Operation not permitted

I am not in the audio as systemd handles session permissions sanely. When I try to run cava as root (mostly just for giggles), it segfaults with a coredump (from which I pulled the following backtrace):

#0  0x00007f502cfa21d4 in fclose@@GLIBC_2.2.5 () from /usr/lib/libc.so.6
#1  0x0000000000403cad in ?? ()
#2  0x000000000040449c in ?? ()
#3  0x00007f502cf5a790 in __libc_start_main () from /usr/lib/libc.so.6
#4  0x00000000004022c9 in ?? ()

The symbols were obviously stripped during the creation of the package for installation (under ArchLinux).

The config file will in INI format:

#CONFIG FILE
paramter1=value1
;parameter2=value2

Suggested parameters:

• input method
• alsa device
• fifo path
• output method
• color/background color
• frame rate
• Sensitivity
• number of bars
• cutoff frequencies
  • High frequency cutoff
  • Low frequency cutoff
• smoothing
  • "integral" smoothing
  • "moster cat" smoothing
  • gravity
• equalizer
  • base
  • mid tone
  • treble

Please make suggestions for other parameters.

Other considerations:

• Should all parameters be overridable via command line arguments?
• Should all parameters be adjustable real time via keyboard shortcuts?(nice to have)
• Should there be a possibility to save adjusted parameters real time back in to the config file?(nice to have)

@CelestialWalrus has volunteered to do the initial work. Feel free to modify or reject any of my suggestions.