--- /dev/null
+Format: http://www.debian.org/doc/packaging-manuals/copyright-format/1.0/
+Upstream-Name: codec2
+Source: <url://example.com>
+
+Files: *
+Copyright: <years> <put author's name and email here>
+ <years> <likewise for another author>
+License: <special license>
+ <Put the license of the package here indented by 1 space>
+ <This follows the format of Description: lines in control file>
+ .
+ <Including paragraphs>
+
+# If you want to use GPL v2 or later for the /debian/* files use
+# the following clauses, or change it to suit. Delete these two lines
+Files: debian/*
+Copyright: 2015 unknown <stuartl@unknown>
+License: GPL-2+
+ This package is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 of the License, or
+ (at your option) any later version.
+ .
+ This package is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+ .
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>
+ .
+ On Debian systems, the complete text of the GNU General
+ Public License version 2 can be found in "/usr/share/common-licenses/GPL-2".
+
+# Please also look if there are files or directories which have a
+# different copyright/license attached and list them here.
+# Please avoid to pick license terms that are more restrictive than the
+# packaged work, as it may make Debian's contributions unacceptable upstream.
% [ ] return data for plotting, like slope m
% [ ] quantise m
+% dealing with UV, BG noise. Prob is flat spectra. When we fit a low
+% freq masking model it's formant shaped rather than flat. So we get
+% these gaps in spectra that come and go - waterfall noises. In
+% particular at low frequencies. Good news is they don't need to be
+% quantised too finely. This model has the disadvantage of not having
+% variable bandwidths.
+% when do waterfall noises appear?
+% idea: we could add the ability to have wider bands
+% Add some sort of slope or floor
+% increase spacing of samples? Like min spacing in bark dimension
+
+% can we fit a different shape?
+
function [decmaskdB local_maxima_sort] = make_decmask(maskdB, AmdB, Wo, L, mask_sample_freqs_kHz)
% band pass filter: limit search to 250 to 3800 Hz
% Ahh, takes me back to when I was a slip of a speech coder, playing
% with my first CELP codec!
-function [decmaskdB dec_samples] = make_decmask_abys(maskdB, AmdB, Wo, L, mask_sample_freqs_kHz)
+function [decmaskdB dec_samples error_log candidate_log target_log] = make_decmask_abys(maskdB, AmdB, Wo, L, mask_sample_freqs_kHz)
% band pass filter: limit search to 250 to 3800 Hz
- m_st = max(1,floor((pi*250/4000)/Wo));
- m_en = floor((pi*3800/4000)/Wo);
+ %m_st = max(1,floor((pi*250/4000)/Wo));
+ %m_en = floor((pi*3800/4000)/Wo);
+ m_st = 1;
+ m_en = L;
target = maskdB;
decmaskdB = zeros(1,L);
dec_samples = [];
-
+ error_log = [];
+ candidate_log = [];
for sample=1:4
+ target_log = target;
+
% find best position for sample to minimise distance to target
min_mse = 1E32;
for m=m_st:m_en
single_mask_m = schroeder(m*Wo*4/pi, mask_sample_freqs_kHz) + AmdB(m);
candidate = max(decmaskdB, single_mask_m);
- candidate = max(zeros(1,L), candidate);
+ candidate = max(min(maskdB), candidate);
error = target - candidate;
- mse = sum(error .^ 2);
+ error_log = [error_log; error];
+ candidate_log = [candidate_log; candidate];
+ mse = sum(abs(error)); % MSE in log domain
%printf("m: %d f: %f error: %f\n", m, m*Wo*4/pi, mse);
if (mse < min_mse)
min_mse = mse;
decmaskdB = min_candidate;
dec_samples = [dec_samples; AmdB(min_mse_m) min_mse_m];
+ %printf("sample: %d min_mse_m: %d\n", sample, min_mse_m);
+ end
+
+ % simulate quantisation of amplitudes by adding some noise
+
+ if 0
+ masker_amps_dB = dec_samples(:,1);
+ masker_amps_dB += 3*(1 - 2*rand(4,1));
+ masker_freqs_kHz = dec_samples(:,2)*Wo*4/pi;
+ decmaskdB = determine_mask(masker_amps_dB, masker_freqs_kHz, mask_sample_freqs_kHz);
+ end
+
+ % quantisation of amplitudes. Determine and subtract mean. Quantise difference
+ % from mean to 4 levels (2 bits), at 6dB/level:
+ %
+ % Level
+ % 0 -9
+ % 1 -3
+ % 2 +3
+ % 3 +9
+
+ if 0
+ masker_amps_dB = dec_samples(:,1);
+ masker_freqs_kHz = dec_samples(:,2)*Wo*4/pi;
+
+ energy_dB = mean(masker_amps_dB);
+ masker_amps_dB -= energy_dB;
+ for i=1:4
+ masker_amps_dB(i) = quantise([-9 -3 3 9], masker_amps_dB(i));
+ end
+ masker_amps_dB += energy_dB;
+ decmaskdB = determine_mask(masker_amps_dB, masker_freqs_kHz, mask_sample_freqs_kHz);
end
+ % fit straight line to amplitudes (sounds bad, still a bug somewhere)
+
+ if 0
+ [gradient intercept] = linreg(dec_samples(:,2), dec_samples(:,1), 4);
+ masker_amps_dB = dec_samples(:,2)*gradient + intercept;
+ masker_freqs_kHz = dec_samples(:,2)*Wo*4/pi;
+ decmaskdB = determine_mask(masker_amps_dB, masker_freqs_kHz, mask_sample_freqs_kHz);
+ end
+endfunction
+
+
+% quantise input sample to nearest value in table
+
+function quant_out = quantise(levels, quant_in)
+ best_se = 1E32;
+ for i=1:length(levels)
+ se = (levels(i) - quant_in)^2;
+ if se < best_se
+ quant_out = levels(i);
+ best_se = se;
+ end
+ end
endfunction
% beneath 1.5kHz wider to match the width of F1 and
% "fill in" the spectra better for UV sounds.
- if freq_tone_kHz <= 0.5
- y = 0.5;
+ x1 = 0.5; x2 = 1.5;
+ y1 = 0.5; y2 = 1;
+ grad = (y2 - y1)/(x2 - x1);
+ y_int = y1 - grad*x1;
+
+ if freq_tone_kHz <= x1
+ y = y1;
end
if (freq_tone_kHz > 0.5) && (freq_tone_kHz < 1.5)
- y = 0.5*freq_tone_kHz + 0.25;
+ y = grad*freq_tone_kHz + y_int;
end
- if freq_tone_kHz >= 1.5
- y = 1;
+ if freq_tone_kHz >= x2
+ y = y2;
end
dz = y*(bark(freq_tone_kHz*1000) - bark(f_Hz));
else
endfunction
+% -12dB/octave mask to model speech articulation
+
+function maskdB = resonator(freq_tone_kHz, mask_sample_freqs_kHz)
+ maskdB = zeros(1, length(mask_sample_freqs_kHz));
+ for m=1:length(mask_sample_freqs_kHz)
+ maskdB(m) = -12*abs(log2(freq_tone_kHz/mask_sample_freqs_kHz(m)));
+ printf("m: %d ft: %f fm: %f ft/fm: %f maskdB: %f\n", m, freq_tone_kHz, mask_sample_freqs_kHz(m), freq_tone_kHz/mask_sample_freqs_kHz(m), maskdB(m));
+ end
+endfunction
+
% plot some masking curves, used for working on masking filter changes
function plot_masking
figure(1)
mask_sample_freqs_kHz = 0.1:0.1:(Fs/1000)/2;
- maskdB = schroeder(0.5, mask_sample_freqs_kHz, 0);
- plot(mask_sample_freqs_kHz, maskdB);
+ maskdB_cb = schroeder(0.5, mask_sample_freqs_kHz, 1);
+ plot(mask_sample_freqs_kHz, maskdB_cb);
hold on;
- maskdB = schroeder(0.5, mask_sample_freqs_kHz, 1);
- plot(mask_sample_freqs_kHz, maskdB,'g');
-
- for f=1:0.5:3
- maskdB = schroeder(f, mask_sample_freqs_kHz, 0);
- plot(mask_sample_freqs_kHz, maskdB);
- maskdB = schroeder(f, mask_sample_freqs_kHz, 1);
- plot(mask_sample_freqs_kHz, maskdB,'g');
+ maskdB_res = resonator(0.5, mask_sample_freqs_kHz);
+ plot(mask_sample_freqs_kHz, maskdB_res,'g');
+
+ for f=0.5:0.5:3
+ maskdB_cb = schroeder(f, mask_sample_freqs_kHz, 1);
+ plot(mask_sample_freqs_kHz, maskdB_cb);
+ maskdB_res = resonator(f, mask_sample_freqs_kHz);
+ plot(mask_sample_freqs_kHz, maskdB_res,'g');
end
hold off;
axis([0.1 4 -30 0])
+ grid
figure(2)
- plot(mask_sample_freqs_kHz, bark(mask_sample_freqs_kHz*1000))
- hold on;
- plot(mask_sample_freqs_kHz, modified_bark(mask_sample_freqs_kHz*1000),'g')
- hold off;
+ clf;
+ w = pi/4; beta = 0.9;
+ X = freqz(1,[1 -2*beta*cos(w) beta*beta],4000);
+ plot(10*log10(abs(X)))
grid
endfunction
projects / freetel-svn-tracking.git / commitdiff