itsmattkc
e6c47b14f4
This mostly builds on the keyframing we already set up for video, but the audio renderer will now appropriately updated keyframe inputs per sample in accordance with keyframe values.
66 lines
2.2 KiB
C++
66 lines
2.2 KiB
C++
#include "audioworker.h"
|
|
|
|
AudioWorker::AudioWorker(QObject *parent) :
|
|
AudioRenderWorker(parent)
|
|
{
|
|
}
|
|
|
|
void AudioWorker::FrameToValue(StreamPtr stream, FramePtr frame, NodeValueTable *table)
|
|
{
|
|
Q_UNUSED(stream)
|
|
|
|
table->Push(NodeParam::kSamples, frame->ToByteArray());
|
|
}
|
|
|
|
void AudioWorker::RunNodeAccelerated(const Node *node, const TimeRange &range, const NodeValueDatabase &input_params_in, NodeValueTable *output_params)
|
|
{
|
|
// Check if node processes samples
|
|
if (!node->ProcessesSamplesFrom()) {
|
|
return;
|
|
}
|
|
|
|
// Copy database so we can make some temporary modifications to it
|
|
NodeValueDatabase input_params = input_params_in;
|
|
|
|
// Try to find the sample buffer in the table
|
|
QVariant samples_var = input_params[node->ProcessesSamplesFrom()].Get(NodeParam::kSamples);
|
|
|
|
// If there isn't one, there's nothing to do
|
|
if (samples_var.isNull()) {
|
|
return;
|
|
}
|
|
|
|
QByteArray input_buffer = samples_var.toByteArray();
|
|
QByteArray output_buffer(input_buffer.size(), 0);
|
|
|
|
int sample_count = input_buffer.size() / audio_params().bytes_per_sample_per_channel();
|
|
|
|
// FIXME: Hardcoded float sample format
|
|
for (int i=0;i<sample_count;i++) {
|
|
// Calculate the exact rational time at this sample
|
|
int sample_out_of_channel = i / audio_params().channel_count();
|
|
double sample_to_second = static_cast<double>(sample_out_of_channel) / static_cast<double>(audio_params().sample_rate());
|
|
|
|
rational this_sample_time = rational::fromDouble(range.in().toDouble() + sample_to_second);
|
|
|
|
// Update all non-sample and non-footage inputs
|
|
foreach (NodeParam* param, node->parameters()) {
|
|
if (param->type() == NodeParam::kInput
|
|
&& param != node->ProcessesSamplesFrom()) {
|
|
NodeInput* input = static_cast<NodeInput*>(param);
|
|
|
|
input_params.Insert(input, ProcessInput(input,
|
|
TimeRange(this_sample_time, this_sample_time)));
|
|
}
|
|
}
|
|
|
|
node->ProcessSamples(&input_params,
|
|
audio_params(),
|
|
reinterpret_cast<const float*>(input_buffer.constData()),
|
|
reinterpret_cast<float*>(output_buffer.data()),
|
|
i);
|
|
}
|
|
|
|
output_params->Push(NodeParam::kSamples, output_buffer);
|
|
}
|