summaryrefslogtreecommitdiff
path: root/src/audio/audio_decoder.cpp
blob: 310f57401eb298ceb4467941e43c33ca417e5109 (plain)
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
/*
 * Copyright 2023 jacqueline <me@jacqueline.id.au>
 *
 * SPDX-License-Identifier: GPL-3.0-only
 */

#include "audio_decoder.hpp"

#include <string.h>

#include <algorithm>
#include <cstddef>
#include <cstdint>
#include <memory>
#include <variant>

#include "codec.hpp"
#include "freertos/FreeRTOS.h"

#include "esp_heap_caps.h"
#include "esp_log.h"
#include "freertos/message_buffer.h"
#include "freertos/portmacro.h"

#include "audio_element.hpp"
#include "chunk.hpp"
#include "fatfs_audio_input.hpp"
#include "stream_info.hpp"

namespace audio {

static const char* kTag = "DEC";

AudioDecoder::AudioDecoder()
    : IAudioElement(),
      current_codec_(),
      current_input_format_(),
      current_output_format_(),
      has_samples_to_send_(false),
      has_input_remaining_(false) {}

AudioDecoder::~AudioDecoder() {}

auto AudioDecoder::ProcessStreamInfo(const StreamInfo& info) -> bool {
  if (!std::holds_alternative<StreamInfo::Encoded>(info.format)) {
    return false;
  }
  ESP_LOGI(kTag, "got new stream");
  const auto& encoded = std::get<StreamInfo::Encoded>(info.format);

  // Reuse the existing codec if we can. This will help with gapless playback,
  // since we can potentially just continue to decode as we were before,
  // without any setup overhead.
  // TODO(jacqueline): Reconsider this. It makes a lot of things harder to smash
  // streams together at this layer.
  /*
  if (current_codec_ != nullptr && current_input_format_) {
    auto cur_encoding = std::get<StreamInfo::Encoded>(*current_input_format_);
    if (cur_encoding.type == encoded.type) {
      ESP_LOGI(kTag, "reusing existing decoder");
      current_input_format_ = info.format;
      return true;
    }
  }
  */
  current_input_format_ = info.format;

  ESP_LOGI(kTag, "creating new decoder");
  auto result = codecs::CreateCodecForType(encoded.type);
  if (result.has_value()) {
    current_codec_.reset(result.value());
  } else {
    ESP_LOGE(kTag, "no codec for this file");
    return false;
  }

  return true;
}

auto AudioDecoder::NeedsToProcess() const -> bool {
  return has_samples_to_send_ || has_input_remaining_;
}

auto AudioDecoder::Process(const std::vector<InputStream>& inputs,
                           OutputStream* output) -> void {
  auto input = inputs.begin();
  const StreamInfo& info = input->info();

  // Check the input stream's format has changed (or, by extension, if this is
  // the first stream).
  if (!current_input_format_ || *current_input_format_ != info.format) {
    ESP_LOGI(kTag, "beginning new stream");
    has_samples_to_send_ = false;
    ProcessStreamInfo(info);
    auto res = current_codec_->BeginStream(input->data());
    input->consume(res.first);
    if (res.second.has_error()) {
      // TODO(jacqueline): Handle errors.
      return;
    }

    // The stream started successfully. Record what format the samples are in.
    codecs::ICodec::OutputFormat format = res.second.value();
    current_output_format_ = StreamInfo::Pcm{
        .channels = format.num_channels,
        .bits_per_sample = format.bits_per_sample,
        .sample_rate = format.sample_rate_hz,
    };

    if (info.seek_to_seconds) {
      seek_to_sample_ = *info.seek_to_seconds * format.sample_rate_hz;
    } else {
      seek_to_sample_.reset();
    }
  }

  while (seek_to_sample_) {
    ESP_LOGI(kTag, "seeking forwards...");
    auto res = current_codec_->SeekStream(input->data(), *seek_to_sample_);
    input->consume(res.first);
    if (res.second.has_error()) {
      auto err = res.second.error();
      if (err == codecs::ICodec::Error::kOutOfInput) {
        return;
      } else {
        // TODO(jacqueline): Handle errors.
        seek_to_sample_.reset();
      }
    } else {
      seek_to_sample_.reset();
    }
  }

  has_input_remaining_ = true;
  while (true) {
    // TODO(jacqueline): Pass through seek info here?
    if (!output->prepare(*current_output_format_)) {
      ESP_LOGI(kTag, "waiting for buffer to become free");
      break;
    }

    auto res = current_codec_->ContinueStream(input->data(), output->data());
    input->consume(res.first);
    if (res.second.has_error()) {
      if (res.second.error() == codecs::ICodec::Error::kOutOfInput) {
        ESP_LOGW(kTag, "out of input");
        ESP_LOGW(kTag, "(%u bytes left)", input->data().size_bytes());
        has_input_remaining_ = false;
        // We can't be halfway through sending samples if the codec is asking
        // for more input.
        has_samples_to_send_ = false;
        input->mark_incomplete();
      } else {
        // TODO(jacqueline): Handle errors.
        ESP_LOGE(kTag, "codec return fatal error");
      }
      return;
    }

    ESP_LOGI(kTag, "enc read: %u", res.first);

    codecs::ICodec::OutputInfo out_info = res.second.value();
    output->add(out_info.bytes_written);
    has_samples_to_send_ = !out_info.is_finished_writing;

    ESP_LOGI(kTag, "enc wrote: %u", out_info.bytes_written);
    if (out_info.is_finished_writing) {
      ESP_LOGI(kTag, "(write finished)");
    }

    if (has_samples_to_send_) {
      // We weren't able to fit all the generated samples into the output
      // buffer. Stop trying; we'll finish up during the next pass.
      break;
    }
  }
}

}  // namespace audio