aboutsummaryrefslogtreecommitdiff
path: root/samples/SpeechRecognition/include/AudioCapture.hpp
diff options
context:
space:
mode:
authorÉanna Ó Catháin <eanna.ocathain@arm.com>2021-04-07 14:35:25 +0100
committerNarumol Prangnawarat <narumol.prangnawarat@arm.com>2021-05-07 18:12:36 +0100
commit95b930c940c47e2f9a783cf17c87449cab4633c5 (patch)
tree3fbc7bc62f245fea6476e84ec9bc80b08f7a7a59 /samples/SpeechRecognition/include/AudioCapture.hpp
parentabee0016b5b5c4129bc8d30f440449d835a6404a (diff)
downloadarmnn-95b930c940c47e2f9a783cf17c87449cab4633c5.tar.gz
MLECO-1252 ASR sample application using the public ArmNN C++ API.
Change-Id: I98cd505b8772a8c8fa88308121bc94135bb45068 Signed-off-by: Éanna Ó Catháin <eanna.ocathain@arm.com>
Diffstat (limited to 'samples/SpeechRecognition/include/AudioCapture.hpp')
-rw-r--r--samples/SpeechRecognition/include/AudioCapture.hpp62
1 files changed, 62 insertions, 0 deletions
diff --git a/samples/SpeechRecognition/include/AudioCapture.hpp b/samples/SpeechRecognition/include/AudioCapture.hpp
new file mode 100644
index 0000000000..90c2eccacf
--- /dev/null
+++ b/samples/SpeechRecognition/include/AudioCapture.hpp
@@ -0,0 +1,62 @@
+//
+// Copyright © 2020 Arm Ltd and Contributors. All rights reserved.
+// SPDX-License-Identifier: MIT
+//
+
+#pragma once
+
+#include <string>
+#include <iostream>
+
+#include <math.h>
+
+#include <vector>
+
+#include <exception>
+
+#include "SlidingWindow.hpp"
+
+namespace asr
+{
+
+/**
+* @brief Class used to capture the audio data loaded from file, and to provide a method of
+ * extracting correctly positioned and appropriately sized audio windows
+*
+*/
+ class AudioCapture
+ {
+ public:
+
+ SlidingWindow<const float> m_window;
+ int lastReadIdx= 0;
+
+ /**
+ * @brief Default constructor
+ */
+ AudioCapture()
+ {};
+
+ /**
+ * @brief Function to load the audio data captured from the
+ * input file to memory.
+ */
+ std::vector<float> LoadAudioFile(std::string filePath);
+
+ /**
+ * @brief Function to initialize the sliding window. This will set its position in memory, its
+ * window size and its stride.
+ */
+ void InitSlidingWindow(float* data, size_t dataSize, int minSamples, size_t stride);
+
+ /**
+ * Checks whether there is another block of audio in memory to read
+ */
+ bool HasNext();
+
+ /**
+ * Retrieves the next block of audio if its available
+ */
+ std::vector<float> Next();
+ };
+} // namespace asr \ No newline at end of file