-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #51 from matejsemancik/feature/video-export-refactor
New video export
- Loading branch information
Showing
7 changed files
with
306 additions
and
199 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
95 changes: 95 additions & 0 deletions
95
core/src/main/kotlin/dev/matsem/astral/core/tools/videoexport/FFTSerializer.kt
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,95 @@ | ||
| package dev.matsem.astral.core.tools.videoexport | ||
|
|
||
| import ddf.minim.AudioSample | ||
| import ddf.minim.Minim | ||
| import ddf.minim.analysis.BeatDetect | ||
| import ddf.minim.analysis.FFT | ||
| import dev.matsem.astral.core.tools.audio.AudioProcessor | ||
| import processing.core.PApplet | ||
| import java.io.PrintWriter | ||
|
|
||
| /** | ||
| * This class generates FFT and BeatDetect analysis of music file using Minim library and then serializes this info | ||
| * to text file for later use by [VideoExporter] to mock the audio samples into exported movie. | ||
| */ | ||
| class FFTSerializer(private val parent: PApplet, private val minim: Minim) { | ||
|
|
||
| companion object { | ||
| /** | ||
| * Text separator used for FFT sample (de)serialization. | ||
| */ | ||
| const val SEP = "|" | ||
| } | ||
|
|
||
| /** | ||
| * Serializes FFT and BeatDetect analysis of provided [audioFilePath] audio file into text file at the same location | ||
| * as the [audioFilePath]. | ||
| */ | ||
| fun serialize(audioFilePath: String) { | ||
| val output: PrintWriter = parent.createWriter(parent.dataPath("$audioFilePath.txt")) | ||
| val track = minim.loadSample(audioFilePath, 2048) | ||
|
|
||
| val fftSize = 1024 | ||
| val sampleRate = track.sampleRate() | ||
|
|
||
| val beatDetect = BeatDetect(fftSize, sampleRate) | ||
|
|
||
| val fftSamplesL = FloatArray(fftSize) | ||
| val fftSamplesR = FloatArray(fftSize) | ||
|
|
||
| val samplesL = track.getChannel(AudioSample.LEFT) | ||
| val samplesR = track.getChannel(AudioSample.RIGHT) | ||
|
|
||
| val fftL = FFT(fftSize, sampleRate) | ||
| val fftR = FFT(fftSize, sampleRate) | ||
|
|
||
| fftL.logAverages(AudioProcessor.FFT_BANDWIDTH, AudioProcessor.FFT_OCTAVES) | ||
| fftR.logAverages(AudioProcessor.FFT_BANDWIDTH, AudioProcessor.FFT_OCTAVES) | ||
|
|
||
| val totalChunks = samplesL.size / fftSize + 1 | ||
| val fftSlices = fftL.avgSize() | ||
|
|
||
| for (ci in 0 until totalChunks) { | ||
| val chunkStartIndex = ci * fftSize | ||
| val chunkSize = PApplet.min(samplesL.size - chunkStartIndex, fftSize) | ||
|
|
||
| System.arraycopy(samplesL, chunkStartIndex, fftSamplesL, 0, chunkSize) | ||
| System.arraycopy(samplesR, chunkStartIndex, fftSamplesR, 0, chunkSize) | ||
| if (chunkSize < fftSize) { | ||
| java.util.Arrays.fill(fftSamplesL, chunkSize, fftSamplesL.size - 1, 0.0f) | ||
| java.util.Arrays.fill(fftSamplesR, chunkSize, fftSamplesR.size - 1, 0.0f) | ||
| } | ||
|
|
||
| fftL.forward(fftSamplesL) | ||
| fftR.forward(fftSamplesR) | ||
| beatDetect.detect(fftSamplesL) | ||
|
|
||
| // The format of the saved txt file. | ||
| // The file contains many rows. Each row looks like this: | ||
| // T|B|L|R|L|R|L|R|... etc | ||
| // where T is the time in seconds and B is BeatDetect data | ||
| // Then we alternate left and right channel FFT values | ||
| // The first L and R values in each row are low frequencies (bass) | ||
| // and they go towards high frequency as we advance towards | ||
| // the end of the line. | ||
| val msg = StringBuilder(PApplet.nf(chunkStartIndex / sampleRate, 0, 3).replace(',', '.')) | ||
| val beat = when { | ||
| beatDetect.isKick -> 1 | ||
| beatDetect.isSnare -> 2 | ||
| beatDetect.isHat -> 3 | ||
| else -> 0 | ||
| } | ||
|
|
||
| msg.append(SEP + beat.toString()) | ||
| for (i in 0 until fftSlices) { | ||
| msg.append(SEP + PApplet.nf(fftL.getAvg(i), 0, 4).replace(',', '.')) | ||
| msg.append(SEP + PApplet.nf(fftR.getAvg(i), 0, 4).replace(',', '.')) | ||
| } | ||
| output.println(msg.toString()) | ||
| } | ||
|
|
||
| track.close() | ||
| output.flush() | ||
| output.close() | ||
| } | ||
| } |
Oops, something went wrong.