File: AudioFeedbackLatencyTest.m

package info (click to toggle)
psychtoolbox-3 3.0.19.14.dfsg1-1
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid, trixie
  • size: 86,796 kB
  • sloc: ansic: 176,245; cpp: 20,103; objc: 5,393; sh: 2,753; python: 1,397; php: 384; makefile: 193; java: 113
file content (291 lines) | stat: -rw-r--r-- 10,500 bytes parent folder | download | duplicates (2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
function AudioFeedbackLatencyTest(roundtrip, trigger, deviceid, nrtrials, freq, freqout, fullduplex, runmode)
% AudioFeedbackLatencyTest([roundtrip=0][, trigger=0.1][, deviceid=auto][, nrtrials=10][, freq=auto][, freqout=auto][, fullduplex=0][, runmode=1])
%
% Tries to test sound onset accuracy of PsychPortAudio without need for
% external measurement equipment: Sound signals are played back via
% PsychPortAudio at well defined points in time, using low-latency mode. At
% the same time, sound is captured via PsychPortAudio's capture facilities.
% The idea is that the microphone or line-in connector should pick up and
% capture the sound signals emitted through line-out (via a line-out ->
% line-in feedback cable) or emitted through the speakers. We measure and
% compare timing of emitted vs. captured sound spikes.
%
% Results on MacbookPro, Windows, Linux, suggest that the method works, not with
% 100% accuracy, so its still better to use external measurement equipment to test!
%
% EARLY BETA CODE: USE ONLY WITH GREAT CAUTION AND SUSPICION!
%
% Optional parameters:
%
% 'roundtrip' If set to 0 then this measures scheduling accuracy of sound
% onset, as measured by sound capture -- should be no worse than ~ 1 msec
% on a well working system, and input detection latency, ie., how long does
% it take from physical sound onset to detection of sound onset by the
% script. This would be a useful measure of how fast a "Voicekey" could
% respond to voice onset in a best case scenario.
%
% If set to 1 then this measures time from issuing the PsychPortAudio('Start')
% command to start playback until actual start of playback (by driver self-
% assessment, and by measuring via audio capture), and also as "Roundtrip"
% how long it would take to detect the onset by the script.
%
% 'trigger' = Trigger level for detection of sound onset in captured sound.
%
% 'deviceid' = Index of audio in/out device, if one device is used. If omitted,
% the default audio device is chosen. You can also specify a vector of two device
% indices, to specify separate input and output devices deviceid = [input, output].
% With different devices, full-duplex mode is obviously not supported.
%
% 'ntrials' = Number of measurement trials to perform.
%
% 'freq' = Samplerate of capture device.
%
% 'freqout' = Samplerate of playback device.
%
% 'fullduplex' = Use soundcard in full-duplex mode.
%
% 'runmode' = Runmode for PsychPortAudio to choose.
%
% Obviously this test function can only be used in a very silent room!
%

% History:
% 06/30/2007 Written (MK)
% 06/17/2020 Rewritten (MK)
% 10/08/2021 Improved: Auto-detect and handle different number of channels, frequency.
%            Allow different devices for input and output. (MK)

% Running on PTB-3? Abort otherwise.
AssertOpenGL;

if nargin < 1 || isempty(roundtrip)
    roundtrip = 0;
end

if nargin < 2 || isempty(trigger)
    trigger = 0.1;
end

if nargin < 3 || isempty(deviceid)
    indeviceid = [];
    outdeviceid = [];
else
    if length(deviceid) < 2
        indeviceid = deviceid;
        outdeviceid = deviceid;
    else
        indeviceid = deviceid(1);
        outdeviceid = deviceid(2);
    end
end

if nargin < 4 || isempty(nrtrials)
    nrtrials = 10;
end

if nargin < 5 || isempty(freq)
    freq = [];
end

if nargin < 6 || isempty(freqout)
    freqout = [];
end

if nargin < 7 || isempty(fullduplex)
   fullduplex = 0;
end

if nargin < 8 || isempty(runmode)
   runmode = 1;
end

if ~isempty(indeviceid) && (indeviceid ~= outdeviceid) && fullduplex
    fullduplex = 0;
    fprintf('Warning: Different input and output audio device, so force-disabling requested full-duplex.\n');
end

fprintf('Using runmode %i.\n', runmode);

% Wait for release of all keys on keyboard:
KbReleaseWait;

% Perform basic initialization of the sound driver and request low-latency
% preinit:
InitializePsychSound(1);

PsychPortAudio('Verbosity', 6);

if fullduplex
   % Open the default audio device indeviceid, with mode 3 (== Full-Duplex),
   % and a required latencyclass of 2 == agressive low-latency mode, as well as
   % a frequency of freq Hz and auto sound channel for capture.
   % This returns a handle to the audio device:
   pahandlerec = PsychPortAudio('Open', indeviceid, 3, 2, freq);
   pahandleout = pahandlerec;
else
   % Open the default audio device indeviceid, with mode 2 (== Only audio capture),
   % and a required latencyclass of 2 == agressive low-latency mode, as well as
   % a frequency of freq Hz and auto sound channel for capture.
   % This returns a handle to the audio device:
   pahandlerec = PsychPortAudio('Open', indeviceid, 2, 2, freq);

   % Open 2nd audio device for playback of our test signal with same settings
   % otherwise:
   pahandleout = PsychPortAudio('Open', outdeviceid, 1, 2, freqout);
end

PsychPortAudio('RunMode', pahandlerec, runmode);
PsychPortAudio('RunMode', pahandleout, runmode);

% Find number of output channels and playback frequency:
status = PsychPortAudio('GetStatus', pahandleout);
freqout = status.SampleRate;
props = PsychPortAudio('GetDevices', [], status.OutDeviceIndex);
outChannels = min(2, props.NrOutputChannels);

% Find number of input channels and capture frequency:
status = PsychPortAudio('GetStatus', pahandlerec);
freq = status.SampleRate;
props = PsychPortAudio('GetDevices', [], status.InDeviceIndex);
inChannels = min(2, props.NrInputChannels);

% Build 1khZ, 90% peak amplitude beep tone of 0.1 secs duration, suitable
% for playback at 'freqout' Hz:
testsound = 0.9 * MakeBeep(1000, 0.1, freqout);
testsound = repmat(testsound, outChannels, 1);

% Initialize sound output buffer with it:

PsychPortAudio('FillBuffer', pahandleout, testsound);

% Measurement loop, runs nrtrials trials:
for i = 0:nrtrials
    % Preallocate an internal audio recording  buffer with a capacity of 10 seconds:
    % We do this in the trial-loop instead of just once. This way, the
    % buffer gets reset to initial conditions.
    PsychPortAudio('GetAudioData', pahandlerec, 10);

    % Start audio capture immediately and wait for the capture to start.
    % Return estimated timestamp of when the first sample hit the
    % microphone/sound input jack. We set the number of 'repetitions' to zero,
    % i.e. record until recording is manually stopped.
    if ~fullduplex
       recstart = PsychPortAudio('Start', pahandlerec, 0, 0, 1);
    end

    % Start scheduled playback of test sound in one second from now, one
    % repetition, wait for start, retrieve true estimated onset timestamp t1:
    t1 = GetSecs + 1;

    if roundtrip
        t1 = WaitSecs('UntilTime', t1);
        t2 = PsychPortAudio('Start', pahandleout, 1, 0, 1);
    else
        t1 = PsychPortAudio('Start', pahandleout, 1, t1, 1);
    end
    failed = 0;

    % Audiotrigger code: Fetch audio data and check against threshold:
    level = 0;

    % Repeat as long as below trigger-threshold:
    while level < trigger
        % Fetch current audiodata:
        [audiodata, offset, ~, recstart]= PsychPortAudio('GetAudioData', pahandlerec);

        % Compute maximum signal amplitude in this chunk of data:
        if ~isempty(audiodata)
            level = max(max(abs(audiodata)));
        else
            level = 0;
        end

        % Below trigger-threshold?
        if level < trigger
            % fprintf('Level %f < Trigger %f\n', level, trigger);
            % Wait for a millisecond before next scan:
            WaitSecs('YieldSecs', 0.001);
        end
    end

    if i > 0
        % Determine roundtrip or input latency:
        rtl(i) = GetSecs - t1; %#ok<AGROW>
    end

    % Ok, last fetched chunk was above threshold!
    % Find exact location of first above threshold sample.
    idx = find(max(abs(audiodata)) >= trigger, 1);

    if i > 0
        % Compute "real" latency, taking real starting time of recording and
        % offset of the triggersample in the buffer into account:
        dt(i) = (recstart + ((offset + idx - 1) / freq)) - t1; %#ok<AGROW>

        if dt(i) < -0.1
            % Invalid measurement! Abort whole procedure...
            fprintf('Trial %i: INVALID MEASUREMENT %f secs detected (Value %i -> %i := %f). Aborting. Please raise the trigger threshold and retry.\n', i, dt(i), offset + idx, idx, audiodata(1,idx));
            failed = 1;
            break;
        end
    end

    % Initialize our recordedaudio vector with captured data starting from
    % triggersample:
    recordedaudio = audiodata(:, idx:end);

    % Stop capture after a tiny bit more time:
    WaitSecs(0.1);
    if ~fullduplex
       PsychPortAudio('Stop', pahandlerec);
    end

    % Stop playback:
    outstatus = PsychPortAudio('GetStatus', pahandleout); %#ok<NASGU>
    PsychPortAudio('Stop', pahandleout);

    % Perform a last fetch operation to get all remaining data from the capture engine:
    recordedaudio = [recordedaudio PsychPortAudio('GetAudioData', pahandlerec)]; %#ok<AGROW,NASGU>
    recstatus = PsychPortAudio('GetStatus', pahandlerec); %#ok<NASGU>

    % Plot it, just for the fun of it:
    nrsamples = size(audiodata(:,idx:end), 2);
    if inChannels >= 2
        plot(1:nrsamples, audiodata(1,idx:end), 'r', 1:nrsamples, audiodata(2,idx:end), 'b', 1:nrsamples, repmat(trigger, 1, nrsamples), '-', 1:nrsamples, repmat(-trigger, 1, nrsamples), '-');
    else
        plot(1:nrsamples, audiodata(1,idx:end), 'r', 1:nrsamples, repmat(trigger, 1, nrsamples), '-', 1:nrsamples, repmat(-trigger, 1, nrsamples), '-');
    end
    drawnow;

    % Print the stats:
    if i > 0
        if roundtrip
            fprintf('%i : Estimated roundtrip latency %f ms : Measured output start to sound onset latency %f ms. Sound output delay %f ms.\n', i, 1000 * rtl(i), 1000 * dt(i), 1000 * (t2 - t1));
        else
            fprintf('%i : Estimated input latency %f ms : Diff real - scheduled %f ms. \n', i, 1000 * rtl(i), 1000 * dt(i));
        end
    end

    % Pause for a second, then next trial...
    WaitSecs(1);
end

% Close the audio devices:
PsychPortAudio('Close');

if failed
    return;
end

rtl = 1000 * rtl;
dt = 1000 * dt;
if roundtrip
    fprintf('Mean roundtrip latency             %f ms, stddev %f ms.\n', mean(rtl), std(rtl));
    fprintf('Mean startup to sound output delay %f ms, stddev %f ms.\n', mean(dt), std(dt));
else
    fprintf('Mean input latency     %f ms, stddev %f ms.\n', mean(rtl), std(rtl));
    fprintf('Mean scheduling offset %f ms, stddev %f ms.\n', mean(dt), std(dt));
end

% Done.
fprintf('Demo finished, bye!\n');