smokephil/ardour
1
0
Fork 0
mirror of https://github.com/Ardour/ardour.git synced 2025-12-07 15:25:01 +01:00
Code Issues Projects Releases Wiki Activity

Fix a typo in HiAndLowPass.lua

Browse source 
Instead of mixing two consecutive biquad orders with a linear xfade, the
code used the same weight for both signals to mix, due to a copy-paste
blunder. Brown-bag fix this now. Also change a little the threshold for
using 32-samples chunks to cater for steepness jumping from 0 to 4 (this
doesn't change anything for commonly used rates and is just for
robustness in never-happening cases anyway).
This commit is contained in:
Julien "_FrnchFrgg_" RIVAUD 2016-07-19 16:38:26 +02:00
parent 55bd8ff4d8
commit 2e933a6414
1 changed files with 6 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

12
scripts/HiAndLowPass.lua
View file

@ -70,11 +70,11 @@ function dsp_init (rate)
-- For performance reasons, we want to ensure that two consecutive values
-- of the interpolated "steepness" are less that 1 apart. By choosing the
-- interpolation chunk size to be 64 in most cases, but 32 if the rate is
-- strictly less than 16kHz (there's only 8kHz in standard rates), we can
-- strictly less than 22kHz (there's only 8kHz in standard rates), we can
-- ensure that steepness interpolation will never change the parameter by
-- more than ~0.82.
-- more than ~0.86.
lpf_chunk = 64
if rate < 16000 then lpf_chunk = 32 end
if rate < 22000 then lpf_chunk = 32 end
-- We apply a discrete version of the standard RC low-pass, with a cutoff
-- frequency of 15Hz. For more information about the underlying math, see
-- https://en.wikipedia.org/wiki/Low-pass_filter#Discrete-time_realization
@ -198,11 +198,11 @@ function dsp_run (ins, outs, n_samples)
if xfade > 0 then
ARDOUR.DSP.apply_gain_to_buffer (outs[c]:offset (off), siz, 1 - xfade)
hp[c][ho+1]:run (mem:to_float (off), siz)
ARDOUR.DSP.mix_buffers_with_gain (outs[c]:offset (off), mem:to_float (off), siz, 1 - xfade)
ARDOUR.DSP.mix_buffers_with_gain (outs[c]:offset (off), mem:to_float (off), siz, xfade)
-- also run the next biquad because it needs to have the correct state
-- in case it start affecting the next chunck of output. Higher order
-- ones are guaranteed not to be needed for the next run because the
-- interpolated order won't increase more than 0.8 in one step thanks
-- interpolated order won't increase more than 0.86 in one step thanks
-- to the choice of the value of |lpf|.
if ho + 2 <= 4 then hp[c][ho+2]:run (mem:to_float (off), siz) end
elseif ho + 1 <= 4 then
@ -227,7 +227,7 @@ function dsp_run (ins, outs, n_samples)
if xfade > 0 then
ARDOUR.DSP.apply_gain_to_buffer (outs[c]:offset (off), siz, 1 - xfade)
lp[c][lo+1]:run (mem:to_float (off), siz)
ARDOUR.DSP.mix_buffers_with_gain (outs[c]:offset (off), mem:to_float (off), siz, 1 - xfade)
ARDOUR.DSP.mix_buffers_with_gain (outs[c]:offset (off), mem:to_float (off), siz, xfade)
-- also run the next biquad in case it start affecting the next
-- chunck of output.
if lo + 2 <= 4 then lp[c][lo+2]:run (mem:to_float (off), siz) end