Files
spacetris/src/states/VideoState.cpp
2025-12-25 19:41:19 +01:00

391 lines
12 KiB
C++

// VideoState.cpp
#include "VideoState.h"
#include "../video/VideoPlayer.h"
#include "../audio/Audio.h"
#include "../audio/AudioManager.h"
#include "../core/state/StateManager.h"
#include <SDL3/SDL.h>
#include <algorithm>
#include <cmath>
#include <cstring>
#include <cstdint>
extern "C" {
#include <libavformat/avformat.h>
#include <libavcodec/avcodec.h>
#include <libavutil/avutil.h>
#include <libavutil/channel_layout.h>
#include <libswresample/swresample.h>
}
VideoState::VideoState(StateContext& ctx)
: State(ctx)
, m_player(std::make_unique<VideoPlayer>())
{
}
VideoState::~VideoState() {
onExit();
}
bool VideoState::begin(SDL_Renderer* renderer, const std::string& path) {
m_path = path;
if (!m_player) {
m_player = std::make_unique<VideoPlayer>();
}
if (!m_player->open(m_path, renderer)) {
SDL_LogWarn(SDL_LOG_CATEGORY_APPLICATION, "[VideoState] Failed to open intro video: %s", m_path.c_str());
return false;
}
if (!m_player->decodeFirstFrame()) {
SDL_LogWarn(SDL_LOG_CATEGORY_APPLICATION, "[VideoState] Failed to decode first frame: %s", m_path.c_str());
// Still allow entering; we will likely render black.
}
return true;
}
void VideoState::onEnter() {
m_phase = Phase::FadeInFirstFrame;
m_phaseClockMs = 0.0;
m_blackOverlayAlpha = 1.0f;
m_audioDecoded.store(false);
m_audioDecodeFailed.store(false);
m_audioStarted = false;
m_audioPcm.clear();
m_audioRate = 44100;
m_audioChannels = 2;
// Decode audio in the background during fade-in.
m_audioThread = std::make_unique<std::jthread>([this](std::stop_token st) {
(void)st;
std::vector<int16_t> pcm;
int rate = 44100;
int channels = 2;
const bool ok = decodeAudioPcm16Stereo44100(m_path, pcm, rate, channels);
if (!ok) {
m_audioDecodeFailed.store(true);
m_audioDecoded.store(true, std::memory_order_release);
return;
}
// Transfer results.
m_audioRate = rate;
m_audioChannels = channels;
m_audioPcm = std::move(pcm);
m_audioDecoded.store(true, std::memory_order_release);
});
}
void VideoState::onExit() {
stopAudio();
if (m_audioThread) {
// Request stop and join.
m_audioThread.reset();
}
}
void VideoState::handleEvent(const SDL_Event& e) {
(void)e;
}
void VideoState::startAudioIfReady() {
if (m_audioStarted) return;
if (!m_audioDecoded.load(std::memory_order_acquire)) return;
if (m_audioDecodeFailed.load()) return;
if (m_audioPcm.empty()) return;
// Use the existing audio output path (same device as music/SFX).
if (auto sys = AudioManager::get()) sys->playSfx(m_audioPcm, m_audioChannels, m_audioRate, 1.0f);
m_audioStarted = true;
}
void VideoState::stopAudio() {
// We currently feed intro audio as an SFX buffer into the mixer.
// It will naturally end; no explicit stop is required.
}
void VideoState::update(double frameMs) {
switch (m_phase) {
case Phase::FadeInFirstFrame: {
m_phaseClockMs += frameMs;
const float t = (FADE_IN_MS > 0.0) ? float(std::clamp(m_phaseClockMs / FADE_IN_MS, 0.0, 1.0)) : 1.0f;
m_blackOverlayAlpha = 1.0f - t;
if (t >= 1.0f) {
m_phase = Phase::Playing;
m_phaseClockMs = 0.0;
if (m_player) {
m_player->start();
}
startAudioIfReady();
}
break;
}
case Phase::Playing: {
startAudioIfReady();
if (m_player) {
m_player->update(frameMs);
if (m_player->isFinished()) {
m_phase = Phase::FadeOutToBlack;
m_phaseClockMs = 0.0;
m_blackOverlayAlpha = 0.0f;
}
} else {
m_phase = Phase::FadeOutToBlack;
m_phaseClockMs = 0.0;
m_blackOverlayAlpha = 0.0f;
}
break;
}
case Phase::FadeOutToBlack: {
m_phaseClockMs += frameMs;
const float t = (FADE_OUT_MS > 0.0) ? float(std::clamp(m_phaseClockMs / FADE_OUT_MS, 0.0, 1.0)) : 1.0f;
m_blackOverlayAlpha = t;
if (t >= 1.0f) {
// Switch to MAIN (Menu) with a fade-in from black.
if (ctx.startupFadeAlpha) {
*ctx.startupFadeAlpha = 1.0f;
}
if (ctx.startupFadeActive) {
*ctx.startupFadeActive = true;
}
if (ctx.stateManager) {
ctx.stateManager->setState(AppState::Menu);
}
m_phase = Phase::Done;
}
break;
}
case Phase::Done:
default:
break;
}
}
void VideoState::render(SDL_Renderer* renderer, float logicalScale, SDL_Rect logicalVP) {
(void)logicalScale;
(void)logicalVP;
if (!renderer) return;
int winW = 0, winH = 0;
SDL_GetRenderOutputSize(renderer, &winW, &winH);
// Draw video fullscreen if available.
if (m_player && m_player->isTextureReady()) {
SDL_SetRenderViewport(renderer, nullptr);
SDL_SetRenderScale(renderer, 1.0f, 1.0f);
m_player->render(renderer, winW, winH);
} else {
SDL_SetRenderDrawColor(renderer, 0, 0, 0, 255);
SDL_FRect r{0.f, 0.f, (float)winW, (float)winH};
SDL_RenderFillRect(renderer, &r);
}
// Apply fade overlay (black).
if (m_blackOverlayAlpha > 0.0f) {
const Uint8 a = (Uint8)std::clamp((int)std::lround(m_blackOverlayAlpha * 255.0f), 0, 255);
SDL_SetRenderDrawBlendMode(renderer, SDL_BLENDMODE_BLEND);
SDL_SetRenderDrawColor(renderer, 0, 0, 0, a);
SDL_FRect full{0.f, 0.f, (float)winW, (float)winH};
SDL_RenderFillRect(renderer, &full);
}
}
bool VideoState::decodeAudioPcm16Stereo44100(
const std::string& path,
std::vector<int16_t>& outPcm,
int& outRate,
int& outChannels
) {
outPcm.clear();
outRate = 44100;
outChannels = 2;
AVFormatContext* fmt = nullptr;
if (avformat_open_input(&fmt, path.c_str(), nullptr, nullptr) != 0) {
return false;
}
if (avformat_find_stream_info(fmt, nullptr) < 0) {
avformat_close_input(&fmt);
return false;
}
int audioStream = -1;
for (unsigned i = 0; i < fmt->nb_streams; ++i) {
if (fmt->streams[i]->codecpar->codec_type == AVMEDIA_TYPE_AUDIO) {
audioStream = (int)i;
break;
}
}
if (audioStream < 0) {
avformat_close_input(&fmt);
return false;
}
AVCodecParameters* codecpar = fmt->streams[audioStream]->codecpar;
const AVCodec* codec = avcodec_find_decoder(codecpar->codec_id);
if (!codec) {
avformat_close_input(&fmt);
return false;
}
AVCodecContext* dec = avcodec_alloc_context3(codec);
if (!dec) {
avformat_close_input(&fmt);
return false;
}
if (avcodec_parameters_to_context(dec, codecpar) < 0) {
avcodec_free_context(&dec);
avformat_close_input(&fmt);
return false;
}
if (avcodec_open2(dec, codec, nullptr) < 0) {
avcodec_free_context(&dec);
avformat_close_input(&fmt);
return false;
}
AVChannelLayout outLayout{};
av_channel_layout_default(&outLayout, 2);
AVChannelLayout inLayout{};
if (av_channel_layout_copy(&inLayout, &dec->ch_layout) < 0 || inLayout.nb_channels <= 0) {
av_channel_layout_uninit(&inLayout);
av_channel_layout_default(&inLayout, 2);
}
SwrContext* swr = nullptr;
if (swr_alloc_set_opts2(
&swr,
&outLayout,
AV_SAMPLE_FMT_S16,
44100,
&inLayout,
dec->sample_fmt,
dec->sample_rate,
0,
nullptr
) < 0) {
av_channel_layout_uninit(&inLayout);
av_channel_layout_uninit(&outLayout);
avcodec_free_context(&dec);
avformat_close_input(&fmt);
return false;
}
if (swr_init(swr) < 0) {
swr_free(&swr);
av_channel_layout_uninit(&inLayout);
av_channel_layout_uninit(&outLayout);
avcodec_free_context(&dec);
avformat_close_input(&fmt);
return false;
}
AVPacket* pkt = av_packet_alloc();
AVFrame* frame = av_frame_alloc();
if (!pkt || !frame) {
if (pkt) av_packet_free(&pkt);
if (frame) av_frame_free(&frame);
swr_free(&swr);
av_channel_layout_uninit(&inLayout);
av_channel_layout_uninit(&outLayout);
avcodec_free_context(&dec);
avformat_close_input(&fmt);
return false;
}
const int outRateConst = 44100;
const int outCh = 2;
while (av_read_frame(fmt, pkt) >= 0) {
if (pkt->stream_index != audioStream) {
av_packet_unref(pkt);
continue;
}
if (avcodec_send_packet(dec, pkt) < 0) {
av_packet_unref(pkt);
continue;
}
av_packet_unref(pkt);
while (true) {
const int rr = avcodec_receive_frame(dec, frame);
if (rr == AVERROR(EAGAIN) || rr == AVERROR_EOF) {
break;
}
if (rr < 0) {
break;
}
const int64_t delay = swr_get_delay(swr, dec->sample_rate);
const int dstNbSamples = (int)av_rescale_rnd(delay + frame->nb_samples, outRateConst, dec->sample_rate, AV_ROUND_UP);
std::vector<uint8_t> outBytes;
outBytes.resize((size_t)dstNbSamples * (size_t)outCh * sizeof(int16_t));
uint8_t* outData[1] = { outBytes.data() };
const uint8_t** inData = (const uint8_t**)frame->data;
const int converted = swr_convert(swr, outData, dstNbSamples, inData, frame->nb_samples);
if (converted > 0) {
const size_t samplesOut = (size_t)converted * (size_t)outCh;
const int16_t* asS16 = (const int16_t*)outBytes.data();
const size_t oldSize = outPcm.size();
outPcm.resize(oldSize + samplesOut);
std::memcpy(outPcm.data() + oldSize, asS16, samplesOut * sizeof(int16_t));
}
av_frame_unref(frame);
}
}
// Flush decoder
avcodec_send_packet(dec, nullptr);
while (avcodec_receive_frame(dec, frame) >= 0) {
const int64_t delay = swr_get_delay(swr, dec->sample_rate);
const int dstNbSamples = (int)av_rescale_rnd(delay + frame->nb_samples, outRateConst, dec->sample_rate, AV_ROUND_UP);
std::vector<uint8_t> outBytes;
outBytes.resize((size_t)dstNbSamples * (size_t)outCh * sizeof(int16_t));
uint8_t* outData[1] = { outBytes.data() };
const uint8_t** inData = (const uint8_t**)frame->data;
const int converted = swr_convert(swr, outData, dstNbSamples, inData, frame->nb_samples);
if (converted > 0) {
const size_t samplesOut = (size_t)converted * (size_t)outCh;
const int16_t* asS16 = (const int16_t*)outBytes.data();
const size_t oldSize = outPcm.size();
outPcm.resize(oldSize + samplesOut);
std::memcpy(outPcm.data() + oldSize, asS16, samplesOut * sizeof(int16_t));
}
av_frame_unref(frame);
}
av_frame_free(&frame);
av_packet_free(&pkt);
swr_free(&swr);
av_channel_layout_uninit(&inLayout);
av_channel_layout_uninit(&outLayout);
avcodec_free_context(&dec);
avformat_close_input(&fmt);
outRate = outRateConst;
outChannels = outCh;
return !outPcm.empty();
}