Differences

This shows you the differences between two versions of the page.

--- inputting_20real-time_20audio [2024/04/19 14:58] – draft richardrussell
+++ inputting_20real-time_20audio [2024/04/19 17:43] (current) – richardrussell
@@ Line 45: / Line 45: @@
 You could of course present the choices and accept the selection in different ways, for example a dialogue box containing a list box plus OK and Cancel buttons.  Or you might prefer it to be determined by a configuration file rather than a user selection made each time the program is run.  That's entirely up to you; adapt the above code accordingly.
+==== Choosing the buffer size ====
+The first step is to decide how large the audio buffer should be. To some extent this is an arbitrary decision, but it will depend on things like //latency// (how much time elapses between the audio arriving and it being processed by your program) and the amount of work needed to process the received audio data.\\ \\  It is vitally important that your program can process the audio data quickly enough, otherwise data will be lost, with undesirable results. If there is any variability in the rate at which you can process the data (for example it depends on disk or network accesses) then you may need to use a larger buffer to 'iron out' the fluctuation. In the example below the length of the buffer is 1024 samples; at 44100 Hz stereo that implies a latency of at least 24 milliseconds.
+<code bb4w>
+        SamplesPerBuffer% = 1024
+</code>
 ==== Selecting the audio format ====
-The first step is to decide the audio //format// you will use: the principal choices being of sampling rate (the main ones being 11025 Hz, 22050 Hz and 44100 Hz) and number of channels (mono, 1 channel, or stereo, 2 channels). The higher the sampling rate the higher the audio frequency that can be received, but the more work your software needs to do. Normally you should choose the lowest sampling rate suitable for your application, remembering that it needs to be at least **double** the highest audio frequency in which you are interested (according to the [[http://en.wikipedia.org/wiki/Nyquist_rate|Nyquist criterion]]).
+The next step is to decide the audio //format// you will use: the principal choices being of sampling rate (the main ones being 11025 Hz, 22050 Hz and 44100 Hz) and number of channels (mono, 1 channel, or stereo, 2 channels). The higher the sampling rate the higher the audio frequency that can be received, but the more work your software needs to do. Normally you should choose the lowest sampling rate suitable for your application, remembering that it needs to be at least **double** the highest audio frequency in which you are interested (according to the [[http://en.wikipedia.org/wiki/Nyquist_rate|Nyquist criterion]]).
 You set up the required audio format and open the audio capture device as follows:
@@ Line 59: / Line 67: @@
       want.format.l& = &10 : REM AUDIO_S16LSB
       want.channels& = 2
-      want.samples% = Window%
+      want.samples% = SamplesPerBuffer%
+</code>
+==== Opening the audio device and creating the buffer ====
+Now the audio capture device can be opened and the buffer created:
+<code bb4w>
       SYS "SDL_OpenAudioDevice", deviceName$(index%), 1, want{}, have{}, 9, @memhdc% TO Device%
       IF Device% = 0 ERROR 100, "Couldn't open audio device " + deviceName$(index%)
-      SamplingRate% = have.freq%
+      BytesPerBuffer% = have.size%
+      WordsPerBuffer% = BytesPerBuffer% DIV 4
+      DIM Buffer%(WordsPerBuffer% - 1)
 </code>
 This code allows for the possibility that the capture device doesn't support the sampling rate you have chosen and selects a different one.
-==== Creating and initialising the buffers ====
+==== Starting audio capture ====
-The next step is to decide how many audio buffers you need and how large they should be. To some extent this is an arbitrary decision, but it will depend on things like //latency// (how much time elapses between the audio arriving and it being processed by your program) and the amount of work needed to process the received audio data.\\ \\  Normally you should have at least three buffers: one inputting the sampled sound, one being processed by your program, and one spare (the buffers are reused cyclically). It is vitally important that your program can process the audio data quickly enough, otherwise data will be lost, with undesirable results. If there is any variability in the rate at which you can process the data (for example it depends on disk or network accesses) then you may need to use more and/or larger buffers to 'iron out' the fluctuation. Using more buffers is generally preferable to using larger buffers, to minimise any increase in latency.\\ \\  In the example below the number of buffers is three and the length of each buffer is 1024 samples; at 44100 Hz that implies a latency of at least 24 milliseconds. The code for creating and initialising the buffers is as follows:
+Once you have initialised the audio device using the above code, you can start the real-time capture as follows:
 <code bb4w>
-        nBuffers% = 3
+      SYS "SDL_PauseAudioDevice", Device%, 0, @memhdc%
-        SamplesPerBuffer% = 1024
+</code>
-        BytesPerBuffer% = SamplesPerBuffer% * Format.nBlockAlign.l&
-        DIM Buffers{(nBuffers%-1) a&(SamplesPerBuffer% - 1)}
+==== Inputting in real-time ====
+Once the above code has been executed you need to process the received audio buffers fast enough to keep up with the incoming data. The following code constantly cycles, filling the buffer and calling the **PROCprocessbuffer** routine:
+<code bb4w>
+        REPEAT
+          p%% = ^Buffer%(0)
+          R% = BytesPerBuffer%
+          PROCprocessbuffer(p%%, SamplesPerBuffer%)
+          REPEAT
+            SYS "SDL_DequeueAudio", Device%, p%%, R%, @memhdc% TO I%
+            p%% += I% : R% -= I%
+          UNTIL R% <= 0
+        UNTIL FALSE
+        END
 </code>
+In this example the audio processing continues indefinitely, but you can terminate the program prematurely if you wish. If you do, don't forget to execute **PROCcleanup** before exiting the program.
+==== Processing the audio data ====
+Obviously it's only possible to describe this aspect in general terms, because precisely what audio processing takes place will depend on what the program is designed to do. The code below simply calculates the RMS (Root Mean Square) value of the incoming audio:
+<code bb4w>
+        DEF PROCprocessbuffer(B%%, N%)
+        LOCAL I%, V%, sumsq
+        FOR I% = 0 TO N%*2-2 STEP 2
+          V% = B%%!I% AND &FFFF : IF V% >= &8000 V% -= 65536
+          sumsq += V%^2
+        NEXT
+        RMS = SQR(sumsq / N%)
+        ENDPROC
+</code>
+This code is appropriate for monaural input (one channel) where each audio sample consists of a signed 16-bit value in the range -32768 to +32767.
+==== Cleaning up ====
+When you stop the sound capture, or exit the program, you need to shut down the audio input in a controlled fashion:
+<code bb4w>
+      DEF PROCcleanup
+      Device% += 0
+      IF Device% THEN
+        SYS "SDL_PauseAudioDevice", Device%, 1, @memhdc%
+        SYS "SDL_CloseAudioDevice", Device%, @memhdc%
+        Device% = 0
+      ENDIF
+      ENDPROC
+</code>
+This code might form part of a larger routine, if there are other things that need to be shut down.