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
|
From 4a0853f26e24617aa68b3f58a81a4d047f6b7212 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Robert-Andr=C3=A9=20Mauchin?= <zebob.m@gmail.com>
Date: Sun, 5 May 2024 09:38:02 +0200
Subject: [PATCH 1/2] Add compatibility with FFMPEG 7.0
channel_layout has been replaced with ch_layout
---
torchvision/csrc/io/decoder/audio_sampler.cpp | 18 +++++++++++++++
torchvision/csrc/io/decoder/audio_stream.cpp | 22 +++++++++++++++----
2 files changed, 36 insertions(+), 4 deletions(-)
diff --git a/torchvision/csrc/io/decoder/audio_sampler.cpp b/torchvision/csrc/io/decoder/audio_sampler.cpp
index e26d788d9c7..9f7811eda48 100644
--- a/torchvision/csrc/io/decoder/audio_sampler.cpp
+++ b/torchvision/csrc/io/decoder/audio_sampler.cpp
@@ -48,6 +48,23 @@ bool AudioSampler::init(const SamplerParameters& params) {
return false;
}
+#if LIBAVUTIL_VERSION_INT >= AV_VERSION_INT(57, 28, 100)
+ SwrContext *swrContext_ = NULL;
+ AVChannelLayout channel_out;
+ AVChannelLayout channel_in;
+ av_channel_layout_default(&channel_out, params.out.audio.channels);
+ av_channel_layout_default(&channel_in, params.in.audio.channels);
+ int ret = swr_alloc_set_opts2(
+ &swrContext_,
+ &channel_out,
+ (AVSampleFormat)params.out.audio.format,
+ params.out.audio.samples,
+ &channel_in,
+ (AVSampleFormat)params.in.audio.format,
+ params.in.audio.samples,
+ 0,
+ logCtx_);
+#else
swrContext_ = swr_alloc_set_opts(
nullptr,
av_get_default_channel_layout(params.out.audio.channels),
@@ -58,6 +75,7 @@ bool AudioSampler::init(const SamplerParameters& params) {
params.in.audio.samples,
0,
logCtx_);
+#endif
if (swrContext_ == nullptr) {
LOG(ERROR) << "Cannot allocate SwrContext";
return false;
diff --git a/torchvision/csrc/io/decoder/audio_stream.cpp b/torchvision/csrc/io/decoder/audio_stream.cpp
index 0f6c57e5588..82f23b70ab7 100644
--- a/torchvision/csrc/io/decoder/audio_stream.cpp
+++ b/torchvision/csrc/io/decoder/audio_stream.cpp
@@ -6,26 +6,34 @@
namespace ffmpeg {
namespace {
+static int get_nb_channels(const AVFrame* frame, const AVCodecContext* codec) {
+#if LIBAVUTIL_VERSION_INT >= AV_VERSION_INT(57, 28, 100)
+ return frame ? frame->ch_layout.nb_channels : codec->ch_layout.nb_channels;
+#else
+ return frame ? frame->channels : codec->channels;
+#endif
+}
+
bool operator==(const AudioFormat& x, const AVFrame& y) {
return x.samples == static_cast<size_t>(y.sample_rate) &&
- x.channels == static_cast<size_t>(y.channels) && x.format == y.format;
+ x.channels == static_cast<size_t>(get_nb_channels(&y, nullptr)) && x.format == y.format;
}
bool operator==(const AudioFormat& x, const AVCodecContext& y) {
return x.samples == static_cast<size_t>(y.sample_rate) &&
- x.channels == static_cast<size_t>(y.channels) && x.format == y.sample_fmt;
+ x.channels == static_cast<size_t>(get_nb_channels(nullptr, &y)) && x.format == y.sample_fmt;
}
AudioFormat& toAudioFormat(AudioFormat& x, const AVFrame& y) {
x.samples = y.sample_rate;
- x.channels = y.channels;
+ x.channels = get_nb_channels(&y, nullptr);
x.format = y.format;
return x;
}
AudioFormat& toAudioFormat(AudioFormat& x, const AVCodecContext& y) {
x.samples = y.sample_rate;
- x.channels = y.channels;
+ x.channels = get_nb_channels(nullptr, &y);
x.format = y.sample_fmt;
return x;
}
@@ -54,9 +62,15 @@ int AudioStream::initFormat() {
if (format_.format.audio.samples == 0) {
format_.format.audio.samples = codecCtx_->sample_rate;
}
+#if LIBAVUTIL_VERSION_INT >= AV_VERSION_INT(57, 28, 100)
+ if (format_.format.audio.channels == 0) {
+ format_.format.audio.channels = codecCtx_->ch_layout.nb_channels;
+ }
+#else
if (format_.format.audio.channels == 0) {
format_.format.audio.channels = codecCtx_->channels;
}
+#endif
if (format_.format.audio.format == AV_SAMPLE_FMT_NONE) {
format_.format.audio.format = codecCtx_->sample_fmt;
}
From 9047266c48173c23a68203441390a1ca186e0ee5 Mon Sep 17 00:00:00 2001
From: Nicolas Hug <contact@nicolas-hug.com>
Date: Thu, 6 Jun 2024 14:37:22 +0100
Subject: [PATCH 2/2] Lint
---
torchvision/csrc/io/decoder/audio_sampler.cpp | 2 +-
torchvision/csrc/io/decoder/audio_stream.cpp | 10 ++++++----
2 files changed, 7 insertions(+), 5 deletions(-)
diff --git a/torchvision/csrc/io/decoder/audio_sampler.cpp b/torchvision/csrc/io/decoder/audio_sampler.cpp
index 9f7811eda48..d46b93ddc69 100644
--- a/torchvision/csrc/io/decoder/audio_sampler.cpp
+++ b/torchvision/csrc/io/decoder/audio_sampler.cpp
@@ -49,7 +49,7 @@ bool AudioSampler::init(const SamplerParameters& params) {
}
#if LIBAVUTIL_VERSION_INT >= AV_VERSION_INT(57, 28, 100)
- SwrContext *swrContext_ = NULL;
+ SwrContext* swrContext_ = NULL;
AVChannelLayout channel_out;
AVChannelLayout channel_in;
av_channel_layout_default(&channel_out, params.out.audio.channels);
diff --git a/torchvision/csrc/io/decoder/audio_stream.cpp b/torchvision/csrc/io/decoder/audio_stream.cpp
index 82f23b70ab7..9d7354e02f5 100644
--- a/torchvision/csrc/io/decoder/audio_stream.cpp
+++ b/torchvision/csrc/io/decoder/audio_stream.cpp
@@ -8,20 +8,22 @@ namespace ffmpeg {
namespace {
static int get_nb_channels(const AVFrame* frame, const AVCodecContext* codec) {
#if LIBAVUTIL_VERSION_INT >= AV_VERSION_INT(57, 28, 100)
- return frame ? frame->ch_layout.nb_channels : codec->ch_layout.nb_channels;
+ return frame ? frame->ch_layout.nb_channels : codec->ch_layout.nb_channels;
#else
- return frame ? frame->channels : codec->channels;
+ return frame ? frame->channels : codec->channels;
#endif
}
bool operator==(const AudioFormat& x, const AVFrame& y) {
return x.samples == static_cast<size_t>(y.sample_rate) &&
- x.channels == static_cast<size_t>(get_nb_channels(&y, nullptr)) && x.format == y.format;
+ x.channels == static_cast<size_t>(get_nb_channels(&y, nullptr)) &&
+ x.format == y.format;
}
bool operator==(const AudioFormat& x, const AVCodecContext& y) {
return x.samples == static_cast<size_t>(y.sample_rate) &&
- x.channels == static_cast<size_t>(get_nb_channels(nullptr, &y)) && x.format == y.sample_fmt;
+ x.channels == static_cast<size_t>(get_nb_channels(nullptr, &y)) &&
+ x.format == y.sample_fmt;
}
AudioFormat& toAudioFormat(AudioFormat& x, const AVFrame& y) {
|