-
Notifications
You must be signed in to change notification settings - Fork 74k
/
decode_wav_op_test.cc
180 lines (150 loc) · 6.83 KB
/
decode_wav_op_test.cc
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
/* Copyright 2017 The TensorFlow Authors. All Rights Reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
==============================================================================*/
#define EIGEN_USE_THREADS
#include <functional>
#include <memory>
#include <vector>
#include "tensorflow/cc/client/client_session.h"
#include "tensorflow/cc/ops/audio_ops.h"
#include "tensorflow/cc/ops/const_op.h"
#include "tensorflow/cc/ops/math_ops.h"
#include "tensorflow/core/framework/shape_inference_testutil.h"
#include "tensorflow/core/framework/tensor_testutil.h"
#include "tensorflow/core/framework/types.h"
#include "tensorflow/core/framework/types.pb.h"
#include "tensorflow/core/kernels/ops_util.h"
#include "tensorflow/core/lib/core/status_test_util.h"
#include "tensorflow/core/platform/test.h"
namespace tensorflow {
namespace ops {
namespace {
TEST(DecodeWavOpTest, DecodeWavTest) {
Scope root = Scope::NewRootScope();
std::vector<uint8> wav_data = {
'R', 'I', 'F', 'F', 44, 0, 0, 0, // size of whole file - 8
'W', 'A', 'V', 'E', 'f', 'm', 't', ' ', 16, 0, 0,
0, // size of fmt block - 8: 24 - 8
1, 0, // format: PCM (1)
1, 0, // channels: 1
0x13, 0x37, 0, 0, // sample rate: 14099
0x26, 0x6e, 0, 0, // byte rate: 2 * 14099
2, 0, // block align: NumChannels * BytesPerSample
16, 0, // bits per sample: 2 * 8
'd', 'a', 't', 'a', 8, 0, 0, 0, // size of payload: 8
0, 0, // first sample: 0
0xff, 0x3f, // second sample: 16383
0xff, 0x7f, // third sample: 32767 (saturated)
0x00, 0x80, // fourth sample: -32768 (saturated)
};
Tensor content_tensor =
test::AsScalar<tstring>(string(wav_data.begin(), wav_data.end()));
Output content_op =
Const(root.WithOpName("content_op"), Input::Initializer(content_tensor));
DecodeWav decode_wav_op =
DecodeWav(root.WithOpName("decode_wav_op"), content_op);
TF_ASSERT_OK(root.status());
ClientSession session(root);
std::vector<Tensor> outputs;
TF_EXPECT_OK(session.Run(ClientSession::FeedType(),
{decode_wav_op.audio, decode_wav_op.sample_rate},
&outputs));
const Tensor& audio = outputs[0];
const int sample_rate = outputs[1].flat<int32>()(0);
EXPECT_EQ(2, audio.dims());
EXPECT_EQ(1, audio.dim_size(1));
EXPECT_EQ(4, audio.dim_size(0));
EXPECT_NEAR(0.0f, audio.flat<float>()(0), 1e-4f);
EXPECT_NEAR(0.5f, audio.flat<float>()(1), 1e-4f);
EXPECT_NEAR(1.0f, audio.flat<float>()(2), 1e-4f);
EXPECT_NEAR(-1.0f, audio.flat<float>()(3), 1e-4f);
EXPECT_EQ(14099, sample_rate);
}
TEST(DecodeWavOpTest, DecodeWav_ShapeFn) {
ShapeInferenceTestOp op("DecodeWav");
INFER_ERROR("Shape must be rank 0 but is rank 1", op, "[1]");
// audio shape is unknown when desired_{samples,channels} are default.
TF_ASSERT_OK(NodeDefBuilder("test", "DecodeWav")
.Input({"a", 0, DT_STRING})
.Finalize(&op.node_def));
INFER_OK(op, "[]", "[?,?];[]");
TF_ASSERT_OK(NodeDefBuilder("test", "DecodeWav")
.Input({"a", 0, DT_STRING})
.Attr("desired_samples", 42)
.Finalize(&op.node_def));
INFER_OK(op, "[]", "[42,?];[]");
// Negative sample value is rejected.
TF_ASSERT_OK(NodeDefBuilder("test", "DecodeWav")
.Input({"a", 0, DT_STRING})
.Attr("desired_samples", -2)
.Finalize(&op.node_def));
INFER_ERROR("samples must be non-negative, got -2", op, "[]");
TF_ASSERT_OK(NodeDefBuilder("test", "DecodeWav")
.Input({"a", 0, DT_STRING})
.Attr("desired_channels", 2)
.Finalize(&op.node_def));
INFER_OK(op, "[]", "[?,2];[]");
// Negative channel value is rejected.
TF_ASSERT_OK(NodeDefBuilder("test", "DecodeWav")
.Input({"a", 0, DT_STRING})
.Attr("desired_channels", -2)
.Finalize(&op.node_def));
INFER_ERROR("channels must be non-negative, got -2", op, "[]");
}
TEST(DecodeWavOpTest, DecodeWavWithJunkChunkTest) {
Scope root = Scope::NewRootScope();
std::vector<uint8> wav_data = {
'R', 'I', 'F', 'F', 76, 0, 0, 0, // size of whole file - 8
'W', 'A', 'V', 'E', 'J', 'U', 'N', 'K', // JUNK tag
28, 0, 0, 0, // size of JUNK chunk
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // JUNK chunk
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 'f', 'm', 't', ' ',
16, 0, 0, 0, // size of fmt
// block - 8: 24 -
// 8
1, 0, // format: PCM (1)
1, 0, // channels: 1
0x13, 0x37, 0, 0, // sample rate: 14099
0x26, 0x6e, 0, 0, // byte rate: 2 * 14099
2, 0, // block align: NumChannels * BytesPerSample
16, 0, // bits per sample: 2 * 8
'd', 'a', 't', 'a', 8, 0, 0, 0, // size of payload: 8
0, 0, // first sample: 0
0xff, 0x3f, // second sample: 16383
0xff, 0x7f, // third sample: 32767 (saturated)
0x00, 0x80, // fourth sample: -32768 (saturated)
};
Tensor content_tensor =
test::AsScalar<tstring>(string(wav_data.begin(), wav_data.end()));
Output content_op =
Const(root.WithOpName("content_op"), Input::Initializer(content_tensor));
DecodeWav decode_wav_op =
DecodeWav(root.WithOpName("decode_wav_op"), content_op);
TF_ASSERT_OK(root.status());
ClientSession session(root);
std::vector<Tensor> outputs;
TF_EXPECT_OK(session.Run(ClientSession::FeedType(),
{decode_wav_op.audio, decode_wav_op.sample_rate},
&outputs));
const Tensor& audio = outputs[0];
const int sample_rate = outputs[1].flat<int32>()(0);
EXPECT_EQ(2, audio.dims());
EXPECT_EQ(1, audio.dim_size(1));
EXPECT_EQ(4, audio.dim_size(0));
EXPECT_NEAR(0.0f, audio.flat<float>()(0), 1e-4f);
EXPECT_NEAR(0.5f, audio.flat<float>()(1), 1e-4f);
EXPECT_NEAR(1.0f, audio.flat<float>()(2), 1e-4f);
EXPECT_NEAR(-1.0f, audio.flat<float>()(3), 1e-4f);
EXPECT_EQ(14099, sample_rate);
}
} // namespace
} // namespace ops
} // namespace tensorflow