Memory leak on Win10?

[Haibane]

Hi!
I love this visualization a lot and I have had it run on ESP8266, Raspberry Pi and crouton. Somehow it seems to take up all 16Gb of RAM in approximately 8 minutes with the config attached below. I have very limited experience in coding and I suspect a memory leak. Any ideas?

I am using VoiceMeeter for ASIO bit-to-bit audio loopback and Equalizer APO for EQ and delay so I can sync the visualization with the audio I hear. CPU is 8 core @ 3.5Ghz. Graphics card is 1070ti. Display where the Colorchord window is running in is 1024x768@120Hz, attached to the graphics card.

# This is the configuration file for colorchord.
# Most values are already defaulted in the software.
# This file is constantly checked for new versions.
# \r, and ; are used as terminators, so you can put
# multiple entries on the same line.

#Whether to limit the control loop to ~60ish FPS.
cpu_autolimit = 0

#General GUI properties.
title = PA Test
set_screenx = 800
set_screeny = 800

#Sound properties.
buffer = 1024
play = 0
rec = 1
channels = 2
samplerate = 96000
wininput = -1

#Compiled version will default this.
#sound_source = ALSA
#-1 indicates left and right, 0 left, 1 right.

sample_channel = -1
sourcename = default
#alsa_output.pci-0000_00_1f.3.analog-stereo.monitor
#default
# alsa_output.pci-0000_00_1b.0.analog-stereo.monitor
#alsa_output.pci-0000_00_1f.3.analog-stereo.monitor << New laptop
#use pactl list | grep pci- | grep monitor

##################################
# General ColorChord properties. #
##################################

# How much to amplify the incoming signal.
amplify = 0.8

# What is the base note?  I.e. the lowest note. 
# Note that it won't have very much impact until an octave up though!
base_hz = 55

# This is only used when dealing with the slow decompose (now defunct)
# decompose_iterations = 1000
# default_sigma = 1.4000

# DFT properties for the DFT up top.
dft_iir = 0.6
dft_q = 20.0000
dft_speedup = 1000.0000
octaves = 5

# Should we use a progressive DFT?
# 0 = DFT Quick
# 1 = DFT Progressive
# 2 = DFT Progressive Integer	
# 3 = DFT Progressive Integer Skippy
# 4 = Integer, 32-Bit, Progressive, Skippy.
do_progressive_dft = 4


filter_iter = 2
filter_strength = .5

# How many bins per octave to use? Can not be changed
freqbins = 24


# For the final note information... How much to slack everything?
#How much to change brightness based on amplitude
note_attach_amp_iir = 0.3500
note_attach_amp_iir2 = 0.250
#How much to change color based on frequency
note_attach_freq_iir = 0.3000

#How many bins a note can jump from frame to frame to be considered a slide.
#this is used to prevent notes from popping in and out a lot.

#Bigger=less likely to have circles
note_combine_distance = 0.5000
#I cant tell what it does yet
note_jumpability = 200.000
#I cant tell what it does yet either
note_minimum_new_distribution_value = 0.0200
#Bigger=easier to dim to black when low volume
note_out_chop = 0.08000

#compress_coefficient = 4.0
#compress_exponent = .5


#=======================================================================
#Outputs


This is a vornoi thing: 
outdrivers = OutputVoronoi, DisplayArray
#Optimize fps over res!
lightx = 64
lighty = 64
fromsides = 1
shape_cutoff = 0.03
satamp = 5.000
amppow = 2.510
distpow = 1.500

[cnlohr]

What mechanism are you using to build your Windows build so I can mimic it? Is it the normal TCC in .bat in the Windows folder?

[Haibane]

I was using mingw-32. I am trying to recreate and verify the issue to see if it persists. Thank you.

[cnlohr]

Understood. Can you provide instructions how you are going about building mingw-32? Are you using Linux? WSL? Running it from Windows host?

[cnlohr]

I regret that none of us can repro this. If you are able to get more information please provide it and re-open. :(