Audio: MFCC: Update example run script run_mfcc.sh

This patch contains several updates:

- The run is with valgrind is added to catch memory leaks.
- The  script applied duplicate "-i" and "-o" arguments. They are
  removed from "OPT" variables.
- The sof-testbench4 can't override the channels count in topology
  similarly as the IPC3 testbench could. Since the current topology
  is for stereo 16 kHz the input data and command line must be for
  such too.
- To be able to compare MFCC output for successive runs, the "-R"
  option is added to run of sox audio convert utility to prevent e.g.
  randomization of dither.
- The script converts input to s24 and s32 formats and runs them
  for easier check for correct operation with supported formats. The
  conversion is done from the s16 version to be able to compare
  the output audio features those should be the same if internal
  processing is 16 bit.
- A run with Mel configured MFCC is added for s16/24/32 formats.
- A script to decode and visualize Mel spectrogram data is added
  as decode_mel.m.

Signed-off-by: Seppo Ingalsuo <seppo.ingalsuo@linux.intel.com>

Author: singalsu

src/audio/mfcc/tune/README.txt

@@ -16,7 +16,7 @@ The output file is hard-coded to mfcc.raw.
 
 The output can be plotted and retrieved with Matlab or Octave command:
 
-[ceps, t, n] = decode_ceps('mfcc.raw', 13);
+[ceps, t, n] = decode_ceps('mfcc_s16.raw', 13);
 
 In the above it's known from configuration script that MFCC was set up to
 output 13 cepstral coefficients from each FFT -> Mel -> DCT -> Cepstral
@@ -27,3 +27,9 @@ e.g. other sound files found in computer.
 
 ./run_mfcc.sh /usr/share/sounds/gnome/default/alerts/bark.ogg
 ./run_mfcc.sh /usr/share/sounds/gnome/default/alerts/sonar.ogg
+
+The script runs the same input sample with s16/24/32 formats for
+cepstral coefficients data output and Mel frequency spectrogram
+output. The 80 bands Mel output can be visualized with command:
+
+[ceps, t, n] = decode_mel('mel_s16.raw', 80);

src/audio/mfcc/tune/decode_mel.m

@@ -0,0 +1,101 @@
+% [mel, t, n] = decode_mel(fn, num_mel, num_channels)
+%
+% Input
+%   fn - File with MFCC data in .raw or .wav format
+%   num_mel - number of Mel coefficients per frame
+%   num_channels - needed for .raw format, omit for .wav
+%
+% Outputs
+%   mel - Mel coefficients
+%   t - time vector for plotting
+%   n - mel 1..num_mel vector for plotting
+
+% SPDX-License-Identifier: BSD-3-Clause
+% Copyright(c) 2026 Intel Corporation.
+
+function [mel, t, n] = decode_mel(fn, num_mel, num_channels)
+
+if nargin < 3
+	num_channels = 1;
+end
+
+% MFCC stream
+fs = 16e3;
+qformat = 7;
+magic = [25443 28006]; % ASCII 'mfcc' as int16
+
+% Load output data
+[data, num_channels] = get_file(fn, num_channels);
+
+idx1 = find(data == magic(1));
+idx = [];
+for i = 1:length(idx1)
+	if data(idx1(i) + 1) == magic(2)
+		idx = [idx idx1(i)];
+	end
+end
+
+if isempty(idx)
+	error('No magic value markers found from stream');
+end
+
+period_mel = idx(2)-idx(1);
+num_frames = length(idx);
+
+% Last frame can be incomplete due to span over multiple periods
+last = idx(end) + num_mel - 1;
+if (last > length(data))
+    num_frames = num_frames - 1;
+end
+
+t_mel = period_mel / num_channels / fs;
+t = (0:num_frames -1) * t_mel;
+n = 1:num_mel;
+
+mel = zeros(num_mel, num_frames);
+for i = 1:num_frames
+	i1 = idx(i) + 2;
+	i2 = i1 + num_mel - 1;
+	mel(:,i) = data(i1:i2) / 2^qformat;
+end
+
+figure;
+imagesc(t, n, mel);
+axis xy;
+colormap(jet);
+colorbar;
+tstr = sprintf('SOF MFCC Mel coefficients (%s)', fn);
+title(tstr, 'Interpreter', 'None');
+xlabel('Time (s)');
+ylabel('Mel coef #');
+
+end
+
+function [data, num_channels] = get_file(fn, num_channels)
+
+[~, ~, ext] = fileparts(fn);
+
+switch lower(ext)
+	case '.raw'
+		fh = fopen(fn, 'r');
+		data = fread(fh, 'int16');
+		fclose(fh);
+	case '.wav'
+		tmp = audioread(fn, 'native');
+		t = whos('tmp');
+		if ~strcmp(t.class, 'int16')
+			error('Only 16-bit wav file format is supported');
+		end
+		s = size(tmp);
+		num_channels = s(2);
+		if num_channels > 1
+			data = int16(zeros(prod(s), 1));
+			for i = 1:num_channels
+				data(i:num_channels:end) = tmp(:, i);
+			end
+		end
+	otherwise
+		error('Unknown audio format');
+end
+
+end

src/audio/mfcc/tune/run_mfcc.sh

@@ -4,19 +4,53 @@
 
 set -e
 
-RAW_INPUT=in.raw
-RAW_OUTPUT=mfcc.raw
+RAW_INPUT_S16=in_s16.raw
+RAW_INPUT_S24=in_s24.raw
+RAW_INPUT_S32=in_s32.raw
+RAW_OUTPUT_S16=mfcc_s16.raw
+RAW_OUTPUT_S24=mfcc_s24.raw
+RAW_OUTPUT_S32=mfcc_s32.raw
 
+VALGRIND="valgrind --leak-check=full"
 TESTBENCH=$SOF_WORKSPACE/sof/tools/testbench/build_testbench/install/bin/sof-testbench4
-TOPOLOGY=$SOF_WORKSPACE/sof/tools/build_tools/topology/topology2/development/sof-hda-benchmark-mfcc16.tplg
-OPT="-r 16000 -c 2 -b S16_LE -p 3,4 -t $TOPOLOGY -i $RAW_INPUT -o $RAW_OUTPUT"
+TOPOLOGY_S16=$SOF_WORKSPACE/sof/tools/build_tools/topology/topology2/development/sof-hda-benchmark-mfcc16.tplg
+TOPOLOGY_S24=$SOF_WORKSPACE/sof/tools/build_tools/topology/topology2/development/sof-hda-benchmark-mfcc24.tplg
+TOPOLOGY_S32=$SOF_WORKSPACE/sof/tools/build_tools/topology/topology2/development/sof-hda-benchmark-mfcc32.tplg
+OPT_S16="-r 16000 -c 2 -b S16_LE -p 3,4 -t $TOPOLOGY_S16"
+OPT_S24="-r 16000 -c 2 -b S24_LE -p 3,4 -t $TOPOLOGY_S24"
+OPT_S32="-r 16000 -c 2 -b S32_LE -p 3,4 -t $TOPOLOGY_S32"
 
-# Convert input audio file raw 16 kHz 1 channel 16 bit
-sox --encoding signed-integer "$1" -L -r 16000 -c 1 -b 16 "$RAW_INPUT"
+# Convert input audio file raw 16 kHz 2 channel 16 bit
+sox -R --encoding signed-integer "$1" -L -r 16000 -c 2 -b 16 "$RAW_INPUT_S16"
+sox -R --no-dither --encoding signed-integer -L -r 16000 -c 2 -b 16 "$RAW_INPUT_S16" -b 32 "$RAW_INPUT_S32"
+sox -R --no-dither --encoding signed-integer -L -r 16000 -c 2 -b 16 "$RAW_INPUT_S16" -b 32 "$RAW_INPUT_S24" vol 0.003906250000
 
 # Run testbench
-$TESTBENCH $OPT -i "$RAW_INPUT" -o "$RAW_OUTPUT"
+$VALGRIND $TESTBENCH $OPT_S16 -i "$RAW_INPUT_S16" -o "$RAW_OUTPUT_S16"
+$VALGRIND $TESTBENCH $OPT_S24 -i "$RAW_INPUT_S24" -o "$RAW_OUTPUT_S24"
+$VALGRIND $TESTBENCH $OPT_S32 -i "$RAW_INPUT_S32" -o "$RAW_OUTPUT_S32"
 
-echo -----------------------------------------------
-echo   The MFCC data was output to file $RAW_OUTPUT
-echo -----------------------------------------------
+echo ----------------------------------------------------------------------------------
+echo The MFCC data was output to file $RAW_OUTPUT_S16, $RAW_OUTPUT_S24, $RAW_OUTPUT_S32
+echo ----------------------------------------------------------------------------------
+
+RAW_OUTPUT_S16=mel_s16.raw
+RAW_OUTPUT_S24=mel_s24.raw
+RAW_OUTPUT_S32=mel_s32.raw
+
+TESTBENCH=$SOF_WORKSPACE/sof/tools/testbench/build_testbench/install/bin/sof-testbench4
+TOPOLOGY_S16=$SOF_WORKSPACE/sof/tools/build_tools/topology/topology2/development/sof-hda-benchmark-mfccmel16.tplg
+TOPOLOGY_S24=$SOF_WORKSPACE/sof/tools/build_tools/topology/topology2/development/sof-hda-benchmark-mfccmel24.tplg
+TOPOLOGY_S32=$SOF_WORKSPACE/sof/tools/build_tools/topology/topology2/development/sof-hda-benchmark-mfccmel32.tplg
+OPT_S16="-r 16000 -c 2 -b S16_LE -p 3,4 -t $TOPOLOGY_S16"
+OPT_S24="-r 16000 -c 2 -b S24_LE -p 3,4 -t $TOPOLOGY_S24"
+OPT_S32="-r 16000 -c 2 -b S32_LE -p 3,4 -t $TOPOLOGY_S32"
+
+# Run testbench
+$VALGRIND $TESTBENCH $OPT_S16 -i "$RAW_INPUT_S16" -o "$RAW_OUTPUT_S16"
+$VALGRIND $TESTBENCH $OPT_S24 -i "$RAW_INPUT_S24" -o "$RAW_OUTPUT_S24"
+$VALGRIND $TESTBENCH $OPT_S32 -i "$RAW_INPUT_S32" -o "$RAW_OUTPUT_S32"
+
+echo ----------------------------------------------------------------------------------
+echo The MFCC Mel data was output to file $RAW_OUTPUT_S16, $RAW_OUTPUT_S24, $RAW_OUTPUT_S32
+echo ----------------------------------------------------------------------------------